4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
27 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
28 type is internally recognised so the workaround is no longer needed.
30 Functionality is still retained as it is, but will only work with
31 EVP_PKEYs with a legacy internal key.
35 * Changed all "STACK" functions to be macros instead of inline functions. Macro
36 parameters are still checked for type safety at compile time via helper
41 * Remove the RAND_DRBG API
43 The RAND_DRBG API did not fit well into the new provider concept as
44 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
45 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
46 and some of its API calls are rather low-level. This holds in particular
47 for the callback mechanism (RAND_DRBG_set_callbacks()).
49 Adding a compatibility layer to continue supporting the RAND_DRBG API as
50 a legacy API for a regular deprecation period turned out to come at the
51 price of complicating the new provider API unnecessarily. Since the
52 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
55 *Paul Dale and Matthias St. Pierre*
57 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
58 as well as actual hostnames.
62 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
63 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
64 conversely, silently ignore DTLS protocol version bounds when configuring
65 TLS-based contexts. The commands can be repeated to set bounds of both
66 types. The same applies with the corresponding "min_protocol" and
67 "max_protocol" command-line switches, in case some application uses both TLS
70 SSL_CTX instances that are created for a fixed protocol version (e.g.
71 TLSv1_server_method()) also silently ignore version bounds. Previously
72 attempts to apply bounds to these protocol versions would result in an
73 error. Now only the "version-flexible" SSL_CTX instances are subject to
74 limits in configuration files in command-line options.
78 * Deprecated the `ENGINE` API. Engines should be replaced with providers
83 * Reworked the recorded ERR codes to make better space for system errors.
84 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
85 given code is a system error (true) or an OpenSSL error (false).
89 * Reworked the test perl framework to better allow parallel testing.
91 *Nicola Tuveri and David von Oheimb*
93 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
94 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
98 * 'Configure' has been changed to figure out the configuration target if
99 none is given on the command line. Consequently, the 'config' script is
100 now only a mere wrapper. All documentation is changed to only mention
103 *Rich Salz and Richard Levitte*
105 * Added a library context that applications as well as other
106 libraries can use to form a separate context within which libcrypto
107 operations are performed.
109 There are two ways this can be used:
111 - Directly, by passing a library context to functions that take
112 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
114 - Indirectly, by creating a new library context and then assigning
115 it as the new default, with `OPENSSL_CTX_set0_default`.
117 All public OpenSSL functions that take an `OPENSSL_CTX` pointer,
118 apart from the functions directly related to `OPENSSL_CTX`, accept
119 NULL to indicate that the default library context should be used.
121 Library code that changes the default library context using
122 `OPENSSL_CTX_set0_default` should take care to restore it with a
123 second call before returning to the caller.
127 * Handshake now fails if Extended Master Secret extension is dropped
132 * Dropped interactive mode from the 'openssl' program. From now on,
133 the `openssl` command without arguments is equivalent to `openssl
138 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
139 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
140 While the old function names have been retained for backward compatibility
141 they should not be used in new developments
142 because their return values are confusing: Unlike other `_cmp()` functions
143 they do not return 0 in case their arguments are equal.
147 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
148 EC_GROUP_get_field_type().
152 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
153 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
154 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
155 Applications should rely on the library automatically assigning a suitable
156 EC_METHOD internally upon EC_GROUP construction.
160 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
161 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
162 assigned internally without application intervention.
163 Users of EC_GROUP_new() should switch to a different suitable constructor.
167 * Add CAdES-BES signature verification support, mostly derived
168 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
170 *Filipe Raimundo da Silva*
172 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
176 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
177 functions are not widely used and now OpenSSL automatically perform this
178 conversion when needed.
182 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
183 EC_KEY_precompute_mult(). These functions are not widely used and
184 applications should instead switch to named curves which OpenSSL has
185 hardcoded lookup tables for.
189 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
190 should instead use the L<EC_POINT_mul(3)> function.
194 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
195 that are not applicable to the new provider model. Applications should
196 instead use EVP_default_properties_is_fips_enabled() and
197 EVP_default_properties_enable_fips().
201 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
202 is set, an unexpected EOF is ignored, it pretends a close notify was received
203 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
207 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
208 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
209 used and applications should instead use the
210 L<EC_POINT_set_affine_coordinates(3)> and
211 L<EC_POINT_get_affine_coordinates(3)> functions.
215 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
216 arrays to be more easily constructed via a series of utility functions.
217 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
218 the various push functions and finally convert to a passable OSSL_PARAM
219 array using OSSL_PARAM_BLD_to_param().
223 * The security strength of SHA1 and MD5 based signatures in TLS has been
224 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
225 working at the default security level of 1 and instead requires security
226 level 0. The security level can be changed either using the cipher string
227 with @SECLEVEL, or calling SSL_CTX_set_security_level().
231 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
232 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
233 internal keys, if they correspond to one of those built in types.
237 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
238 contain a provider side internal key.
242 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
243 They are old functions that we don't use, and that you could disable with
244 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
248 * Project text documents not yet having a proper file name extension
249 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
250 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
254 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
255 have been converted to Markdown with the goal to produce documents
256 which not only look pretty when viewed online in the browser, but
257 remain well readable inside a plain text editor.
259 To achieve this goal, a 'minimalistic' Markdown style has been applied
260 which avoids formatting elements that interfere too much with the
261 reading flow in the text file. For example, it
263 * avoids [ATX headings][] and uses [setext headings][] instead
264 (which works for `<h1>` and `<h2>` headings only).
265 * avoids [inline links][] and uses [reference links][] instead.
266 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
268 [ATX headings]: https://github.github.com/gfm/#atx-headings
269 [setext headings]: https://github.github.com/gfm/#setext-headings
270 [inline links]: https://github.github.com/gfm/#inline-link
271 [reference links]: https://github.github.com/gfm/#reference-link
272 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
273 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
275 *Matthias St. Pierre*
277 * The test suite is changed to preserve results of each test recipe.
278 A new directory test-runs/ with subdirectories named like the
279 test recipes are created in the build tree for this purpose.
283 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
284 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
285 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
287 *David von Oheimb, Martin Peylo*
289 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
290 The legacy OCSP-focused and only partly documented API is retained.
291 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
295 * Added `util/check-format.pl`, a tool for checking adherence to the
296 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
297 The checks performed are incomplete and yield some false positives.
298 Still the tool should be useful for detecting most typical glitches.
302 * BIO_do_connect and BIO_do_handshake have been extended:
303 If domain name resolution yields multiple IP addresses all of them are tried
304 after connect() failures.
308 * All of the low level RSA functions have been deprecated including:
310 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
311 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
312 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
313 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
314 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
315 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
316 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
317 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
318 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
319 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
320 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
321 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
322 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
323 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
324 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
325 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
326 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
327 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
328 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
329 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
330 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
331 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
332 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
333 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
334 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
335 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
336 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
337 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
339 Use of these low level functions has been informally discouraged for a long
340 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
341 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
342 L<EVP_PKEY_decrypt(3)>.
346 * X509 certificates signed using SHA1 are no longer allowed at security
348 In TLS/SSL the default security level is 1. It can be set either
349 using the cipher string with @SECLEVEL, or calling
350 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
351 a call to SSL_CTX_use_certificate() will fail if the security level is not
353 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
354 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
359 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
360 modified to use PKEY APIs. These commands are now in maintenance mode
361 and no new features will be added to them.
365 * The command line utility rsautl has been deprecated.
366 Instead use the pkeyutl program.
370 * The command line utilities genrsa and rsa have been modified to use PKEY
371 APIs. They now write PKCS#8 keys by default. These commands are now in
372 maintenance mode and no new features will be added to them.
376 * All of the low level DH functions have been deprecated including:
378 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
379 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
380 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
381 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
382 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
383 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
384 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
385 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
386 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
387 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
388 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
389 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
390 DH_meth_get_generate_params and DH_meth_set_generate_params.
392 Use of these low level functions has been informally discouraged for a long
393 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
394 and L<EVP_PKEY_derive(3)>.
398 * All of the low level DSA functions have been deprecated including:
400 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
401 DSA_get_default_method, DSA_set_method, DSA_get_method,
402 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
403 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
404 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
405 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
406 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
407 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
408 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
409 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
410 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
411 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
412 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
415 Use of these low level functions has been informally discouraged for a long
416 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
417 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
421 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
422 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
423 This means that applications don't have to look at the curve NID and
424 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
425 However, they still can, that EVP_PKEY_set_alias_type() call acts as
426 a no-op when the EVP_PKEY is already of the given type.
428 Parameter and key generation is also reworked to make it possible
429 to generate EVP_PKEY_SM2 parameters and keys without having to go
430 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
431 However, code that does the latter will still work as before.
435 * Deprecated low level ECDH and ECDSA functions. These include:
437 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
438 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
441 Use of these low level functions has been informally discouraged for a long
442 time. Instead applications should use the EVP_PKEY_derive(3),
443 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
447 * Deprecated the EC_KEY_METHOD functions. These include:
449 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
450 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
451 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
452 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
453 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
454 EC_KEY_METHOD_get_verify.
456 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
460 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
461 and EVP_PKEY_decrypt() instead.
462 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
463 and EVP_PKEY_encrypt() instead.
467 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
468 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
469 a new formulation to include all the things it can be used for,
470 as well as words of caution.
474 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
475 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
479 * All of the low level HMAC functions have been deprecated including:
481 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
482 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
485 Use of these low level functions has been informally discouraged for a long
486 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
487 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
488 and L<EVP_MAC_final(3)>.
492 * Over two thousand fixes were made to the documentation, including:
493 - Common options (such as -rand/-writerand, TLS version control, etc)
494 were refactored and point to newly-enhanced descriptions in openssl.pod.
495 - Added style conformance for all options (with help from Richard Levitte),
496 documented all reported missing options, added a CI build to check
497 that all options are documented and that no unimplemented options
499 - Documented some internals, such as all use of environment variables.
500 - Addressed all internal broken L<> references.
504 * All of the low level CMAC functions have been deprecated including:
506 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
507 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
509 Use of these low level functions has been informally discouraged for a long
510 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
511 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
512 and L<EVP_MAC_final(3)>.
516 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
517 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
520 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
521 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
522 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
523 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
524 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
525 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
526 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
527 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
528 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
529 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
531 Use of these low level functions has been informally discouraged
532 for a long time. Applications should use the EVP_DigestInit_ex(3),
533 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
537 * Corrected the documentation of the return values from the `EVP_DigestSign*`
538 set of functions. The documentation mentioned negative values for some
539 errors, but this was never the case, so the mention of negative values
542 Code that followed the documentation and thereby check with something
543 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
547 * All of the low level cipher functions have been deprecated including:
549 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
550 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
551 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
552 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
553 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
554 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
555 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
556 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
557 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
558 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
559 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
560 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
561 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
562 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
563 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
564 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
565 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
566 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
567 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
568 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
569 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
570 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
571 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
572 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
573 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
574 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
575 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
576 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
577 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
579 Use of these low level functions has been informally discouraged for
580 a long time. Applications should use the high level EVP APIs, e.g.
581 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
582 equivalently named decrypt functions instead.
584 *Matt Caswell and Paul Dale*
586 * Removed include/openssl/opensslconf.h.in and replaced it with
587 include/openssl/configuration.h.in, which differs in not including
588 <openssl/macros.h>. A short header include/openssl/opensslconf.h
589 was added to include both.
591 This allows internal hacks where one might need to modify the set
592 of configured macros, for example this if deprecated symbols are
593 still supposed to be available internally:
595 #include <openssl/configuration.h>
597 #undef OPENSSL_NO_DEPRECATED
598 #define OPENSSL_SUPPRESS_DEPRECATED
600 #include <openssl/macros.h>
602 This should not be used by applications that use the exported
603 symbols, as that will lead to linking errors.
607 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
608 used in exponentiation with 512-bit moduli. No EC algorithms are
609 affected. Analysis suggests that attacks against 2-prime RSA1024,
610 3-prime RSA1536, and DSA1024 as a result of this defect would be very
611 difficult to perform and are not believed likely. Attacks against DH512
612 are considered just feasible. However, for an attack the target would
613 have to re-use the DH512 private key, which is not recommended anyway.
614 Also applications directly using the low level API BN_mod_exp may be
615 affected if they use BN_FLG_CONSTTIME.
620 * Most memory-debug features have been deprecated, and the functionality
621 replaced with no-ops.
625 * Added documentation for the STACK API.
629 * Introduced a new method type and API, OSSL_ENCODER, to
630 represent generic encoders. An implementation is expected to
631 be able to encode an object associated with a given name (such
632 as an algorithm name for an asymmetric key) into forms given by
633 implementation properties.
635 Encoders are primarily used from inside libcrypto, through
636 calls to functions like EVP_PKEY_print_private(),
637 PEM_write_bio_PrivateKey() and similar.
639 Encoders are specified in such a way that they can be made to
640 directly handle the provider side portion of an object, if this
641 provider side part comes from the same provider as the encoder
642 itself, but can also be made to handle objects in parametrized
643 form (as an OSSL_PARAM array of data). This allows a provider to
644 offer generic encoders as a service for any other provider.
648 * Added a .pragma directive to the syntax of configuration files, to
649 allow varying behavior in a supported and predictable manner.
650 Currently added pragma:
654 This allows dollar signs to be a keyword character unless it's
655 followed by a opening brace or parenthesis. This is useful for
656 platforms where dollar signs are commonly used in names, such as
657 volume names and system directory names on VMS.
661 * Added functionality to create an EVP_PKEY from user data. This
662 is effectively the same as creating a RSA, DH or DSA object and
663 then assigning them to an EVP_PKEY, but directly using algorithm
664 agnostic EVP functions. A benefit is that this should be future
665 proof for public key algorithms to come.
669 * Change the interpretation of the '--api' configuration option to
670 mean that this is a desired API compatibility level with no
671 further meaning. The previous interpretation, that this would
672 also mean to remove all deprecated symbols up to and including
673 the given version, no requires that 'no-deprecated' is also used
674 in the configuration.
676 When building applications, the desired API compatibility level
677 can be set with the OPENSSL_API_COMPAT macro like before. For
678 API compatibility version below 3.0, the old style numerical
679 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
680 For version 3.0 and on, the value is expected to be the decimal
681 value calculated from the major and minor version like this:
683 MAJOR * 10000 + MINOR * 100
687 -DOPENSSL_API_COMPAT=30000 For 3.0
688 -DOPENSSL_API_COMPAT=30200 For 3.2
690 To hide declarations that are deprecated up to and including the
691 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
692 given when building the application as well.
696 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
697 access to certificate and CRL stores via URIs and OSSL_STORE
700 This adds the following functions:
702 - X509_LOOKUP_store()
703 - X509_STORE_load_file()
704 - X509_STORE_load_path()
705 - X509_STORE_load_store()
706 - SSL_add_store_cert_subjects_to_stack()
707 - SSL_CTX_set_default_verify_store()
708 - SSL_CTX_load_verify_file()
709 - SSL_CTX_load_verify_dir()
710 - SSL_CTX_load_verify_store()
714 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
715 The presence of this system service is determined at run-time.
719 * Added functionality to create an EVP_PKEY context based on data
720 for methods from providers. This takes an algorithm name and a
721 property query string and simply stores them, with the intent
722 that any operation that uses this context will use those strings
723 to fetch the needed methods implicitly, thereby making the port
724 of application written for pre-3.0 OpenSSL easier.
728 * The undocumented function NCONF_WIN32() has been deprecated; for
729 conversion details see the HISTORY section of doc/man5/config.pod
733 * Introduced the new functions EVP_DigestSignInit_ex() and
734 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
735 EVP_DigestVerifyUpdate() have been converted to functions. See the man
736 pages for further details.
740 * Over two thousand fixes were made to the documentation, including:
741 adding missing command flags, better style conformance, documentation
744 *Rich Salz, Richard Levitte*
746 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
747 X25519, X448, Ed25519 and Ed448.
751 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
756 * Deprecated the public definition of ERR_STATE as well as the function
757 ERR_get_state(). This is done in preparation of making ERR_STATE an
762 * Added ERR functionality to give callers access to the stored function
763 names that have replaced the older function code based functions.
765 New functions are ERR_get_error_func(), ERR_peek_error_func(),
766 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
767 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
768 and ERR_peek_last_error_all().
770 These functions have become deprecated: ERR_get_error_line_data(),
771 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
772 ERR_func_error_string().
776 * Extended testing to be verbose for failing tests only. The make variables
777 VERBOSE_FAILURE or VF can be used to enable this:
779 $ make VF=1 test # Unix
780 $ mms /macro=(VF=1) test ! OpenVMS
781 $ nmake VF=1 test # Windows
785 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
786 used even when parsing explicit parameters, when loading a encoded key
787 or calling `EC_GROUP_new_from_ecpkparameters()`/
788 `EC_GROUP_new_from_ecparameters()`.
789 This prevents bypass of security hardening and performance gains,
790 especially for curves with specialized EC_METHODs.
791 By default, if a key encoded with explicit parameters is loaded and later
792 encoded, the output is still encoded with explicit parameters, even if
793 internally a "named" EC_GROUP is used for computation.
797 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
798 this change, EC_GROUP_set_generator would accept order and/or cofactor as
799 NULL. After this change, only the cofactor parameter can be NULL. It also
800 does some minimal sanity checks on the passed order.
805 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
806 An attack is simple, if the first CMS_recipientInfo is valid but the
807 second CMS_recipientInfo is chosen ciphertext. If the second
808 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
809 encryption key will be replaced by garbage, and the message cannot be
810 decoded, but if the RSA decryption fails, the correct encryption key is
811 used and the recipient will not notice the attack.
812 As a work around for this potential attack the length of the decrypted
813 key must be equal to the cipher default key length, in case the
814 certifiate is not given and all recipientInfo are tried out.
815 The old behaviour can be re-enabled in the CMS code by setting the
816 CMS_DEBUG_DECRYPT flag.
820 * Early start up entropy quality from the DEVRANDOM seed source has been
821 improved for older Linux systems. The RAND subsystem will wait for
822 /dev/random to be producing output before seeding from /dev/urandom.
823 The seeded state is stored for future library initialisations using
824 a system global shared memory segment. The shared memory identifier
825 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
826 the desired value. The default identifier is 114.
830 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
831 when primes for RSA keys are computed.
832 Since we previously always generated primes == 2 (mod 3) for RSA keys,
833 the 2-prime and 3-prime RSA modules were easy to distinguish, since
834 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
835 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
836 This avoids possible fingerprinting of newly generated RSA modules.
840 * Correct the extended master secret constant on EBCDIC systems. Without this
841 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
842 negotiate EMS will fail. Unfortunately this also means that TLS connections
843 between EBCDIC systems with this fix, and EBCDIC systems without this
844 fix will fail if they negotiate EMS.
848 * Changed the library initialisation so that the config file is now loaded
849 by default. This was already the case for libssl. It now occurs for both
850 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
851 OPENSSL_init_crypto() to suppress automatic loading of a config file.
855 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
856 where the former acts as a replacement for ERR_put_error(), and the
857 latter replaces the combination ERR_put_error()+ERR_add_error_data().
858 ERR_raise_data() adds more flexibility by taking a format string and
859 an arbitrary number of arguments following it, to be processed with
864 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
865 to check if a named provider is loaded and available. When called, it
866 will also activate all fallback providers if such are still present.
870 * Enforce a minimum DH modulus size of 512 bits.
874 * Changed DH parameters to generate the order q subgroup instead of 2q.
875 Previously generated DH parameters are still accepted by DH_check
876 but DH_generate_key works around that by clearing bit 0 of the
877 private key for those. This avoids leaking bit 0 of the private key.
881 * Significantly reduce secure memory usage by the randomness pools.
885 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
890 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
891 algorithms. An implementation of a key exchange algorithm can be obtained
892 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
893 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
894 the older EVP_PKEY_derive_init() function. See the man pages for the new
895 functions for further details.
899 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
903 * Removed the function names from error messages and deprecated the
906 * Removed NextStep support and the macro OPENSSL_UNISTD
910 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
911 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
912 Also removed "export var as function" capability; we do not export
913 variables, only functions.
917 * RC5_32_set_key has been changed to return an int type, with 0 indicating
918 an error and 1 indicating success. In previous versions of OpenSSL this
919 was a void type. If a key was set longer than the maximum possible this
924 * Support SM2 signing and verification schemes with X509 certificate.
928 * Use SHA256 as the default digest for TS query in the ts app.
932 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
933 This checks that the salt length is at least 128 bits, the derived key
934 length is at least 112 bits, and that the iteration count is at least 1000.
935 For backwards compatibility these checks are disabled by default in the
936 default provider, but are enabled by default in the fips provider.
937 To enable or disable these checks use the control
938 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
942 * Default cipher lists/suites are now available via a function, the
943 #defines are deprecated.
947 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
948 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
949 for Windows Store apps easier. Also, the "no-uplink" option has been added.
953 * Join the directories crypto/x509 and crypto/x509v3
957 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
958 This changes the size when using the genpkey app when no size is given. It
959 fixes an omission in earlier changes that changed all RSA, DSA and DH
960 generation apps to use 2048 bits by default.
964 * Added command 'openssl kdf' that uses the EVP_KDF API.
968 * Added command 'openssl mac' that uses the EVP_MAC API.
972 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
973 as default directories. Also added the command 'openssl info'
974 for scripting purposes.
978 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
979 deprecated. These undocumented functions were never integrated into the EVP
980 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
981 Bi-directional IGE mode. These modes were never formally standardised and
982 usage of these functions is believed to be very small. In particular
983 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
984 is ever used. The security implications are believed to be minimal, but
985 this issue was never fixed for backwards compatibility reasons. New code
986 should not use these modes.
990 * Add prediction resistance to the DRBG reseeding process.
994 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
995 mandated by IEEE Std 1619-2018.
999 * Added newline escaping functionality to a filename when using openssl dgst.
1000 This output format is to replicate the output format found in the `*sum`
1001 checksum programs. This aims to preserve backward compatibility.
1003 *Matt Eaton, Richard Levitte, and Paul Dale*
1005 * Removed the heartbeat message in DTLS feature, as it has very
1006 little usage and doesn't seem to fulfill a valuable purpose.
1007 The configuration option is now deprecated.
1011 * Changed the output of 'openssl {digestname} < file' to display the
1012 digest name in its output.
1016 * Added a new generic trace API which provides support for enabling
1017 instrumentation through trace output. This feature is mainly intended
1018 as an aid for developers and is disabled by default. To utilize it,
1019 OpenSSL needs to be configured with the `enable-trace` option.
1021 If the tracing API is enabled, the application can activate trace output
1022 by registering BIOs as trace channels for a number of tracing and debugging
1025 The 'openssl' application has been expanded to enable any of the types
1026 available via environment variables defined by the user, and serves as
1027 one possible example on how to use this functionality.
1029 *Richard Levitte & Matthias St. Pierre*
1031 * Added build tests for C++. These are generated files that only do one
1032 thing, to include one public OpenSSL head file each. This tests that
1033 the public header files can be usefully included in a C++ application.
1035 This test isn't enabled by default. It can be enabled with the option
1036 'enable-buildtest-c++'.
1040 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1044 * Add KMAC to EVP_MAC.
1048 * Added property based algorithm implementation selection framework to
1053 * Added SCA hardening for modular field inversion in EC_GROUP through
1054 a new dedicated field_inv() pointer in EC_METHOD.
1055 This also addresses a leakage affecting conversions from projective
1056 to affine coordinates.
1058 *Billy Bob Brumley, Nicola Tuveri*
1060 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1061 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1062 those algorithms that were already supported through the EVP_PKEY API
1063 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1064 and scrypt are now wrappers that call EVP_KDF.
1068 * Build devcrypto engine as a dynamic engine.
1070 *Eneas U de Queiroz*
1072 * Add keyed BLAKE2 to EVP_MAC.
1076 * Fix a bug in the computation of the endpoint-pair shared secret used
1077 by DTLS over SCTP. This breaks interoperability with older versions
1078 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1079 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1080 interoperability with such broken implementations. However, enabling
1081 this switch breaks interoperability with correct implementations.
1083 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1084 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1088 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1092 * Change the license to the Apache License v2.0.
1096 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1098 - Major releases (indicated by incrementing the MAJOR release number)
1099 may introduce incompatible API/ABI changes.
1100 - Minor releases (indicated by incrementing the MINOR release number)
1101 may introduce new features but retain API/ABI compatibility.
1102 - Patch releases (indicated by incrementing the PATCH number)
1103 are intended for bug fixes and other improvements of existing
1104 features only (like improving performance or adding documentation)
1105 and retain API/ABI compatibility.
1109 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1113 * Remove the 'dist' target and add a tarball building script. The
1114 'dist' target has fallen out of use, and it shouldn't be
1115 necessary to configure just to create a source distribution.
1119 * Recreate the OS390-Unix config target. It no longer relies on a
1120 special script like it did for OpenSSL pre-1.1.0.
1124 * Instead of having the source directories listed in Configure, add
1125 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1130 * Add GMAC to EVP_MAC.
1134 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1138 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1139 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1140 to facilitate the continued use of MACs through raw private keys in
1141 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1145 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1146 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1150 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1151 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1152 are retained for backwards compatibility.
1156 * AES-XTS mode now enforces that its two keys are different to mitigate
1157 the attacked described in "Efficient Instantiations of Tweakable
1158 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1159 Details of this attack can be obtained from:
1160 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1164 * Rename the object files, i.e. give them other names than in previous
1165 versions. Their names now include the name of the final product, as
1166 well as its type mnemonic (bin, lib, shlib).
1170 * Added new option for 'openssl list', '-objects', which will display the
1171 list of built in objects, i.e. OIDs with names.
1175 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1176 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1181 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1182 improves application performance by removing data copies and providing
1183 applications with zero-copy system calls such as sendfile and splice.
1187 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1188 option is set, openssl cleanses (zeroize) plaintext bytes from
1189 internal buffers after delivering them to the application. Note,
1190 the application is still responsible for cleansing other copies
1191 (e.g.: data received by SSL_read(3)).
1195 * `PKCS12_parse` now maintains the order of the parsed certificates
1196 when outputting them via `*ca` (rather than reversing it).
1203 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1205 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1207 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1208 while reading in libssl then we would report an error back to the
1209 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1210 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1211 therefore give a hint as to what went wrong.
1215 * Check that ed25519 and ed448 are allowed by the security level. Previously
1216 signature algorithms not using an MD were not being checked that they were
1217 allowed by the security level.
1221 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1222 was not quite right. The behaviour was not consistent between resumption
1223 and normal handshakes, and also not quite consistent with historical
1224 behaviour. The behaviour in various scenarios has been clarified and
1225 it has been updated to make it match historical behaviour as closely as
1230 * *[VMS only]* The header files that the VMS compilers include automatically,
1231 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1232 that the C++ compiler doesn't understand. This is a shortcoming in the
1233 compiler, but can be worked around with `__cplusplus` guards.
1235 C++ applications that use OpenSSL libraries must be compiled using the
1236 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1237 functions. Otherwise, only functions with symbols of less than 31
1238 characters can be used, as the linker will not be able to successfully
1239 resolve symbols with longer names.
1243 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1244 The presence of this system service is determined at run-time.
1248 * Added newline escaping functionality to a filename when using openssl dgst.
1249 This output format is to replicate the output format found in the `*sum`
1250 checksum programs. This aims to preserve backward compatibility.
1252 *Matt Eaton, Richard Levitte, and Paul Dale*
1254 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1259 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1261 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1262 number generator (RNG). This was intended to include protection in the
1263 event of a fork() system call in order to ensure that the parent and child
1264 processes did not share the same RNG state. However this protection was not
1265 being used in the default case.
1267 A partial mitigation for this issue is that the output from a high
1268 precision timer is mixed into the RNG state so the likelihood of a parent
1269 and child process sharing state is significantly reduced.
1271 If an application already calls OPENSSL_init_crypto() explicitly using
1272 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1275 *Matthias St. Pierre*
1277 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1278 used even when parsing explicit parameters, when loading a encoded key
1279 or calling `EC_GROUP_new_from_ecpkparameters()`/
1280 `EC_GROUP_new_from_ecparameters()`.
1281 This prevents bypass of security hardening and performance gains,
1282 especially for curves with specialized EC_METHODs.
1283 By default, if a key encoded with explicit parameters is loaded and later
1284 encoded, the output is still encoded with explicit parameters, even if
1285 internally a "named" EC_GROUP is used for computation.
1289 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1290 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1291 NULL. After this change, only the cofactor parameter can be NULL. It also
1292 does some minimal sanity checks on the passed order.
1297 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1298 An attack is simple, if the first CMS_recipientInfo is valid but the
1299 second CMS_recipientInfo is chosen ciphertext. If the second
1300 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1301 encryption key will be replaced by garbage, and the message cannot be
1302 decoded, but if the RSA decryption fails, the correct encryption key is
1303 used and the recipient will not notice the attack.
1304 As a work around for this potential attack the length of the decrypted
1305 key must be equal to the cipher default key length, in case the
1306 certifiate is not given and all recipientInfo are tried out.
1307 The old behaviour can be re-enabled in the CMS code by setting the
1308 CMS_DEBUG_DECRYPT flag.
1313 * Early start up entropy quality from the DEVRANDOM seed source has been
1314 improved for older Linux systems. The RAND subsystem will wait for
1315 /dev/random to be producing output before seeding from /dev/urandom.
1316 The seeded state is stored for future library initialisations using
1317 a system global shared memory segment. The shared memory identifier
1318 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1319 the desired value. The default identifier is 114.
1323 * Correct the extended master secret constant on EBCDIC systems. Without this
1324 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1325 negotiate EMS will fail. Unfortunately this also means that TLS connections
1326 between EBCDIC systems with this fix, and EBCDIC systems without this
1327 fix will fail if they negotiate EMS.
1331 * Use Windows installation paths in the mingw builds
1333 Mingw isn't a POSIX environment per se, which means that Windows
1334 paths should be used for installation.
1339 * Changed DH_check to accept parameters with order q and 2q subgroups.
1340 With order 2q subgroups the bit 0 of the private key is not secret
1341 but DH_generate_key works around that by clearing bit 0 of the
1342 private key for those. This avoids leaking bit 0 of the private key.
1346 * Significantly reduce secure memory usage by the randomness pools.
1350 * Revert the DEVRANDOM_WAIT feature for Linux systems
1352 The DEVRANDOM_WAIT feature added a select() call to wait for the
1353 /dev/random device to become readable before reading from the
1354 /dev/urandom device.
1356 It turned out that this change had negative side effects on
1357 performance which were not acceptable. After some discussion it
1358 was decided to revert this feature and leave it up to the OS
1359 resp. the platform maintainer to ensure a proper initialization
1360 during early boot time.
1362 *Matthias St. Pierre*
1364 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1366 * Add build tests for C++. These are generated files that only do one
1367 thing, to include one public OpenSSL head file each. This tests that
1368 the public header files can be usefully included in a C++ application.
1370 This test isn't enabled by default. It can be enabled with the option
1371 'enable-buildtest-c++'.
1375 * Enable SHA3 pre-hashing for ECDSA and DSA.
1379 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1380 This changes the size when using the genpkey app when no size is given. It
1381 fixes an omission in earlier changes that changed all RSA, DSA and DH
1382 generation apps to use 2048 bits by default.
1386 * Reorganize the manual pages to consistently have RETURN VALUES,
1387 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1388 util/fix-doc-nits accordingly.
1390 *Paul Yang, Joshua Lock*
1392 * Add the missing accessor EVP_PKEY_get0_engine()
1396 * Have apps like 's_client' and 's_server' output the signature scheme
1397 along with other cipher suite parameters when debugging.
1401 * Make OPENSSL_config() error agnostic again.
1405 * Do the error handling in RSA decryption constant time.
1409 * Prevent over long nonces in ChaCha20-Poly1305.
1411 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1412 for every encryption operation. RFC 7539 specifies that the nonce value
1413 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1414 and front pads the nonce with 0 bytes if it is less than 12
1415 bytes. However it also incorrectly allows a nonce to be set of up to 16
1416 bytes. In this case only the last 12 bytes are significant and any
1417 additional leading bytes are ignored.
1419 It is a requirement of using this cipher that nonce values are
1420 unique. Messages encrypted using a reused nonce value are susceptible to
1421 serious confidentiality and integrity attacks. If an application changes
1422 the default nonce length to be longer than 12 bytes and then makes a
1423 change to the leading bytes of the nonce expecting the new value to be a
1424 new unique nonce then such an application could inadvertently encrypt
1425 messages with a reused nonce.
1427 Additionally the ignored bytes in a long nonce are not covered by the
1428 integrity guarantee of this cipher. Any application that relies on the
1429 integrity of these ignored leading bytes of a long nonce may be further
1430 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1431 is safe because no such use sets such a long nonce value. However user
1432 applications that use this cipher directly and set a non-default nonce
1433 length to be longer than 12 bytes may be vulnerable.
1435 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1441 * Add DEVRANDOM_WAIT feature for Linux systems
1443 On older Linux systems where the getrandom() system call is not available,
1444 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1445 Contrary to getrandom(), the /dev/urandom device will not block during
1446 early boot when the kernel CSPRNG has not been seeded yet.
1448 To mitigate this known weakness, use select() to wait for /dev/random to
1449 become readable before reading from /dev/urandom.
1451 * Ensure that SM2 only uses SM3 as digest algorithm
1455 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1457 * Change the info callback signals for the start and end of a post-handshake
1458 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1459 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1460 confused by this and assume that a TLSv1.2 renegotiation has started. This
1461 can break KeyUpdate handling. Instead we no longer signal the start and end
1462 of a post handshake message exchange (although the messages themselves are
1463 still signalled). This could break some applications that were expecting
1464 the old signals. However without this KeyUpdate is not usable for many
1469 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1471 * Timing vulnerability in DSA signature generation
1473 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1474 timing side channel attack. An attacker could use variations in the signing
1475 algorithm to recover the private key.
1477 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1482 * Timing vulnerability in ECDSA signature generation
1484 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1485 timing side channel attack. An attacker could use variations in the signing
1486 algorithm to recover the private key.
1488 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1493 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1494 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1495 of two gigabytes and the error handling improved.
1497 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1498 categorized as a normal bug, not a security issue, because the DRBG reseeds
1499 automatically and is fully functional even without additional randomness
1500 provided by the application.
1502 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1504 * Add a new ClientHello callback. Provides a callback interface that gives
1505 the application the ability to adjust the nascent SSL object at the
1506 earliest stage of ClientHello processing, immediately after extensions have
1507 been collected but before they have been processed. In particular, this
1508 callback can adjust the supported TLS versions in response to the contents
1513 * Add SM2 base algorithm support.
1517 * s390x assembly pack: add (improved) hardware-support for the following
1518 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1519 aes-cfb/cfb8, aes-ecb.
1523 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1524 parameter is no longer accepted, as it leads to a corrupt table. NULL
1525 pem_str is reserved for alias entries only.
1529 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1530 step for prime curves. The new implementation is based on formulae from
1531 differential addition-and-doubling in homogeneous projective coordinates
1532 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1533 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1534 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1535 to work in projective coordinates.
1537 *Billy Bob Brumley, Nicola Tuveri*
1539 * Change generating and checking of primes so that the error rate of not
1540 being prime depends on the intended use based on the size of the input.
1541 For larger primes this will result in more rounds of Miller-Rabin.
1542 The maximal error rate for primes with more than 1080 bits is lowered
1545 *Kurt Roeckx, Annie Yousar*
1547 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1551 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1552 moving between systems, and to avoid confusion when a Windows build is
1553 done with mingw vs with MSVC. For POSIX installs, there's still a
1554 symlink or copy named 'tsget' to avoid that confusion as well.
1558 * Revert blinding in ECDSA sign and instead make problematic addition
1559 length-invariant. Switch even to fixed-length Montgomery multiplication.
1563 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1564 step for binary curves. The new implementation is based on formulae from
1565 differential addition-and-doubling in mixed Lopez-Dahab projective
1566 coordinates, modified to independently blind the operands.
1568 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1570 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1571 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1572 EC_METHODs to implement their own specialized "ladder step", to take
1573 advantage of more favorable coordinate systems or more efficient
1574 differential addition-and-doubling algorithms.
1576 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1578 * Modified the random device based seed sources to keep the relevant
1579 file descriptors open rather than reopening them on each access.
1580 This allows such sources to operate in a chroot() jail without
1581 the associated device nodes being available. This behaviour can be
1582 controlled using RAND_keep_random_devices_open().
1586 * Numerous side-channel attack mitigations have been applied. This may have
1587 performance impacts for some algorithms for the benefit of improved
1588 security. Specific changes are noted in this change log by their respective
1593 * AIX shared library support overhaul. Switch to AIX "natural" way of
1594 handling shared libraries, which means collecting shared objects of
1595 different versions and bitnesses in one common archive. This allows to
1596 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1597 doesn't affect the way 3rd party applications are linked, only how
1598 multi-version installation is managed.
1602 * Make ec_group_do_inverse_ord() more robust and available to other
1603 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1604 mitigations are applied to the fallback BN_mod_inverse().
1605 When using this function rather than BN_mod_inverse() directly, new
1606 EC cryptosystem implementations are then safer-by-default.
1610 * Add coordinate blinding for EC_POINT and implement projective
1611 coordinate blinding for generic prime curves as a countermeasure to
1612 chosen point SCA attacks.
1614 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1616 * Add blinding to ECDSA and DSA signatures to protect against side channel
1617 attacks discovered by Keegan Ryan (NCC Group).
1621 * Enforce checking in the pkeyutl command line app to ensure that the input
1622 length does not exceed the maximum supported digest length when performing
1623 a sign, verify or verifyrecover operation.
1627 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1628 I/O in combination with something like select() or poll() will hang. This
1629 can be turned off again using SSL_CTX_clear_mode().
1630 Many applications do not properly handle non-application data records, and
1631 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1632 around the problems in those applications, but can also break some.
1633 It's recommended to read the manpages about SSL_read(), SSL_write(),
1634 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1635 SSL_CTX_set_read_ahead() again.
1639 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1640 now allow empty (zero character) pass phrases.
1644 * Apply blinding to binary field modular inversion and remove patent
1645 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1649 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1650 binary and prime elliptic curves.
1654 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1655 constant time fixed point multiplication.
1659 * Revise elliptic curve scalar multiplication with timing attack
1660 defenses: ec_wNAF_mul redirects to a constant time implementation
1661 when computing fixed point and variable point multiplication (which
1662 in OpenSSL are mostly used with secret scalars in keygen, sign,
1663 ECDH derive operations).
1664 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1667 * Updated CONTRIBUTING
1671 * Updated DRBG / RAND to request nonce and additional low entropy
1672 randomness from the system.
1674 *Matthias St. Pierre*
1676 * Updated 'openssl rehash' to use OpenSSL consistent default.
1680 * Moved the load of the ssl_conf module to libcrypto, which helps
1681 loading engines that libssl uses before libssl is initialised.
1685 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1689 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1691 *Ingo Schwarze, Rich Salz*
1693 * Added output of accepting IP address and port for 'openssl s_server'
1697 * Added a new API for TLSv1.3 ciphersuites:
1698 SSL_CTX_set_ciphersuites()
1699 SSL_set_ciphersuites()
1703 * Memory allocation failures consistently add an error to the error
1708 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1709 in libcrypto when run as setuid/setgid.
1713 * Load any config file by default when libssl is used.
1717 * Added new public header file <openssl/rand_drbg.h> and documentation
1718 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1720 *Matthias St. Pierre*
1722 * QNX support removed (cannot find contributors to get their approval
1723 for the license change).
1727 * TLSv1.3 replay protection for early data has been implemented. See the
1728 SSL_read_early_data() man page for further details.
1732 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1733 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1734 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1735 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1736 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1737 configuration has been separated out. See the ciphers man page or the
1738 SSL_CTX_set_ciphersuites() man page for more information.
1742 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1743 in responder mode now supports the new "-multi" option, which
1744 spawns the specified number of child processes to handle OCSP
1745 requests. The "-timeout" option now also limits the OCSP
1746 responder's patience to wait to receive the full client request
1747 on a newly accepted connection. Child processes are respawned
1748 as needed, and the CA index file is automatically reloaded
1749 when changed. This makes it possible to run the "ocsp" responder
1750 as a long-running service, making the OpenSSL CA somewhat more
1751 feature-complete. In this mode, most diagnostic messages logged
1752 after entering the event loop are logged via syslog(3) rather than
1757 * Added support for X448 and Ed448. Heavily based on original work by
1762 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1763 objects loaded. This adds the functions OSSL_STORE_expect() and
1764 OSSL_STORE_find() as well as needed tools to construct searches and
1765 get the search data out of them.
1769 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1770 version of OpenSSL should review their configuration settings to ensure
1771 that they are still appropriate for TLSv1.3. For further information see:
1772 <https://wiki.openssl.org/index.php/TLS1.3>
1776 * Grand redesign of the OpenSSL random generator
1778 The default RAND method now utilizes an AES-CTR DRBG according to
1779 NIST standard SP 800-90Ar1. The new random generator is essentially
1780 a port of the default random generator from the OpenSSL FIPS 2.0
1781 object module. It is a hybrid deterministic random bit generator
1782 using an AES-CTR bit stream and which seeds and reseeds itself
1783 automatically using trusted system entropy sources.
1785 Some of its new features are:
1786 - Support for multiple DRBG instances with seed chaining.
1787 - The default RAND method makes use of a DRBG.
1788 - There is a public and private DRBG instance.
1789 - The DRBG instances are fork-safe.
1790 - Keep all global DRBG instances on the secure heap if it is enabled.
1791 - The public and private DRBG instance are per thread for lock free
1794 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1796 * Changed Configure so it only says what it does and doesn't dump
1797 so much data. Instead, ./configdata.pm should be used as a script
1798 to display all sorts of configuration data.
1802 * Added processing of "make variables" to Configure.
1806 * Added SHA512/224 and SHA512/256 algorithm support.
1810 * The last traces of Netware support, first removed in 1.1.0, have
1815 * Get rid of Makefile.shared, and in the process, make the processing
1816 of certain files (rc.obj, or the .def/.map/.opt files produced from
1817 the ordinal files) more visible and hopefully easier to trace and
1818 debug (or make silent).
1822 * Make it possible to have environment variable assignments as
1823 arguments to config / Configure.
1827 * Add multi-prime RSA (RFC 8017) support.
1831 * Add SM3 implemented according to GB/T 32905-2016
1832 *Jack Lloyd <jack.lloyd@ribose.com>,*
1833 *Ronald Tse <ronald.tse@ribose.com>,*
1834 *Erick Borsboom <erick.borsboom@ribose.com>*
1836 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1837 as documented in RFC6066.
1838 Based on a patch from Tomasz Moń
1840 *Filipe Raimundo da Silva*
1842 * Add SM4 implemented according to GB/T 32907-2016.
1843 *Jack Lloyd <jack.lloyd@ribose.com>,*
1844 *Ronald Tse <ronald.tse@ribose.com>,*
1845 *Erick Borsboom <erick.borsboom@ribose.com>*
1847 * Reimplement -newreq-nodes and ERR_error_string_n; the
1848 original author does not agree with the license change.
1852 * Add ARIA AEAD TLS support.
1856 * Some macro definitions to support VS6 have been removed. Visual
1857 Studio 6 has not worked since 1.1.0
1861 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1862 without clearing the errors.
1866 * Add "atfork" functions. If building on a system that without
1867 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1868 requirements. The RAND facility now uses/requires this.
1876 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1877 not possible to disable entirely. However, it's still possible to
1878 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1881 To disable, configure with 'no-ui-console'. 'no-ui' is still
1882 possible to use as an alias. Check at compile time with the
1883 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1884 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1888 * Add a STORE module, which implements a uniform and URI based reader of
1889 stores that can contain keys, certificates, CRLs and numerous other
1890 objects. The main API is loosely based on a few stdio functions,
1891 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1892 OSSL_STORE_error and OSSL_STORE_close.
1893 The implementation uses backends called "loaders" to implement arbitrary
1894 URI schemes. There is one built in "loader" for the 'file' scheme.
1898 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1899 then adjusted to work on FreeBSD 8.4 as well.
1900 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1901 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1905 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1906 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1907 error code calls like this:
1909 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1911 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1912 that can be encoded in C. For the foreseeable future, this will only
1915 *Richard Levitte and Tim Hudson*
1917 * Removed BSD cryptodev engine.
1921 * Add a build target 'build_all_generated', to build all generated files
1922 and only that. This can be used to prepare everything that requires
1923 things like perl for a system that lacks perl and then move everything
1924 to that system and do the rest of the build there.
1928 * In the UI interface, make it possible to duplicate the user data. This
1929 can be used by engines that need to retain the data for a longer time
1930 than just the call where this user data is passed.
1934 * Ignore the '-named_curve auto' value for compatibility of applications
1937 *Tomas Mraz <tmraz@fedoraproject.org>*
1939 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1940 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1941 alerts across multiple records (some of which could be empty). In practice
1942 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1943 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1944 support this at all. Supporting it adds significant complexity to the
1945 record layer, and its removal is unlikely to cause interoperability
1950 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1951 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1952 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1957 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1958 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1960 *Richard Levitte, Andy Polyakov*
1962 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1967 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1968 platform rather than 'mingw'.
1972 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1973 success if they are asked to add an object which already exists
1974 in the store. This change cascades to other functions which load
1975 certificates and CRLs.
1979 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1980 facilitate stack unwinding even from assembly subroutines.
1984 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1985 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1989 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1990 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1991 which is the minimum version we support.
1995 * Certificate time validation (X509_cmp_time) enforces stricter
1996 compliance with RFC 5280. Fractional seconds and timezone offsets
1997 are no longer allowed.
2001 * Add support for ARIA
2005 * s_client will now send the Server Name Indication (SNI) extension by
2006 default unless the new "-noservername" option is used. The server name is
2007 based on the host provided to the "-connect" option unless overridden by
2008 using "-servername".
2012 * Add support for SipHash
2016 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2017 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2018 prevent issues where no progress is being made and the peer continually
2019 sends unrecognised record types, using up resources processing them.
2023 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2024 using the algorithm defined in
2025 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2029 * Heartbeat support has been removed; the ABI is changed for now.
2031 *Richard Levitte, Rich Salz*
2033 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2037 * The RSA "null" method, which was partially supported to avoid patent
2038 issues, has been replaced to always returns NULL.
2045 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2047 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2048 used even when parsing explicit parameters, when loading a encoded key
2049 or calling `EC_GROUP_new_from_ecpkparameters()`/
2050 `EC_GROUP_new_from_ecparameters()`.
2051 This prevents bypass of security hardening and performance gains,
2052 especially for curves with specialized EC_METHODs.
2053 By default, if a key encoded with explicit parameters is loaded and later
2054 encoded, the output is still encoded with explicit parameters, even if
2055 internally a "named" EC_GROUP is used for computation.
2059 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2060 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2061 NULL. After this change, only the cofactor parameter can be NULL. It also
2062 does some minimal sanity checks on the passed order.
2067 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2068 An attack is simple, if the first CMS_recipientInfo is valid but the
2069 second CMS_recipientInfo is chosen ciphertext. If the second
2070 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2071 encryption key will be replaced by garbage, and the message cannot be
2072 decoded, but if the RSA decryption fails, the correct encryption key is
2073 used and the recipient will not notice the attack.
2074 As a work around for this potential attack the length of the decrypted
2075 key must be equal to the cipher default key length, in case the
2076 certifiate is not given and all recipientInfo are tried out.
2077 The old behaviour can be re-enabled in the CMS code by setting the
2078 CMS_DEBUG_DECRYPT flag.
2083 * Use Windows installation paths in the mingw builds
2085 Mingw isn't a POSIX environment per se, which means that Windows
2086 paths should be used for installation.
2091 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2093 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2094 This changes the size when using the genpkey app when no size is given. It
2095 fixes an omission in earlier changes that changed all RSA, DSA and DH
2096 generation apps to use 2048 bits by default.
2100 * Prevent over long nonces in ChaCha20-Poly1305.
2102 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2103 for every encryption operation. RFC 7539 specifies that the nonce value
2104 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2105 and front pads the nonce with 0 bytes if it is less than 12
2106 bytes. However it also incorrectly allows a nonce to be set of up to 16
2107 bytes. In this case only the last 12 bytes are significant and any
2108 additional leading bytes are ignored.
2110 It is a requirement of using this cipher that nonce values are
2111 unique. Messages encrypted using a reused nonce value are susceptible to
2112 serious confidentiality and integrity attacks. If an application changes
2113 the default nonce length to be longer than 12 bytes and then makes a
2114 change to the leading bytes of the nonce expecting the new value to be a
2115 new unique nonce then such an application could inadvertently encrypt
2116 messages with a reused nonce.
2118 Additionally the ignored bytes in a long nonce are not covered by the
2119 integrity guarantee of this cipher. Any application that relies on the
2120 integrity of these ignored leading bytes of a long nonce may be further
2121 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2122 is safe because no such use sets such a long nonce value. However user
2123 applications that use this cipher directly and set a non-default nonce
2124 length to be longer than 12 bytes may be vulnerable.
2126 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2132 * Added SCA hardening for modular field inversion in EC_GROUP through
2133 a new dedicated field_inv() pointer in EC_METHOD.
2134 This also addresses a leakage affecting conversions from projective
2135 to affine coordinates.
2137 *Billy Bob Brumley, Nicola Tuveri*
2139 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2140 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2144 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2148 * Remove the 'dist' target and add a tarball building script. The
2149 'dist' target has fallen out of use, and it shouldn't be
2150 necessary to configure just to create a source distribution.
2154 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2156 * Timing vulnerability in DSA signature generation
2158 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2159 timing side channel attack. An attacker could use variations in the signing
2160 algorithm to recover the private key.
2162 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2167 * Timing vulnerability in ECDSA signature generation
2169 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2170 timing side channel attack. An attacker could use variations in the signing
2171 algorithm to recover the private key.
2173 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2178 * Add coordinate blinding for EC_POINT and implement projective
2179 coordinate blinding for generic prime curves as a countermeasure to
2180 chosen point SCA attacks.
2182 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2184 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2186 * Client DoS due to large DH parameter
2188 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2189 malicious server can send a very large prime value to the client. This will
2190 cause the client to spend an unreasonably long period of time generating a
2191 key for this prime resulting in a hang until the client has finished. This
2192 could be exploited in a Denial Of Service attack.
2194 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2199 * Cache timing vulnerability in RSA Key Generation
2201 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2202 a cache timing side channel attack. An attacker with sufficient access to
2203 mount cache timing attacks during the RSA key generation process could
2204 recover the private key.
2206 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2207 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2212 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2213 parameter is no longer accepted, as it leads to a corrupt table. NULL
2214 pem_str is reserved for alias entries only.
2218 * Revert blinding in ECDSA sign and instead make problematic addition
2219 length-invariant. Switch even to fixed-length Montgomery multiplication.
2223 * Change generating and checking of primes so that the error rate of not
2224 being prime depends on the intended use based on the size of the input.
2225 For larger primes this will result in more rounds of Miller-Rabin.
2226 The maximal error rate for primes with more than 1080 bits is lowered
2229 *Kurt Roeckx, Annie Yousar*
2231 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2235 * Add blinding to ECDSA and DSA signatures to protect against side channel
2236 attacks discovered by Keegan Ryan (NCC Group).
2240 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2241 now allow empty (zero character) pass phrases.
2245 * Certificate time validation (X509_cmp_time) enforces stricter
2246 compliance with RFC 5280. Fractional seconds and timezone offsets
2247 are no longer allowed.
2251 * Fixed a text canonicalisation bug in CMS
2253 Where a CMS detached signature is used with text content the text goes
2254 through a canonicalisation process first prior to signing or verifying a
2255 signature. This process strips trailing space at the end of lines, converts
2256 line terminators to CRLF and removes additional trailing line terminators
2257 at the end of a file. A bug in the canonicalisation process meant that
2258 some characters, such as form-feed, were incorrectly treated as whitespace
2259 and removed. This is contrary to the specification (RFC5485). This fix
2260 could mean that detached text data signed with an earlier version of
2261 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2262 signed with a fixed OpenSSL may fail to verify with an earlier version of
2263 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2264 and use the "-binary" flag (for the "cms" command line application) or set
2265 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2269 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2271 * Constructed ASN.1 types with a recursive definition could exceed the stack
2273 Constructed ASN.1 types with a recursive definition (such as can be found
2274 in PKCS7) could eventually exceed the stack given malicious input with
2275 excessive recursion. This could result in a Denial Of Service attack. There
2276 are no such structures used within SSL/TLS that come from untrusted sources
2277 so this is considered safe.
2279 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2285 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2287 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2288 effectively reduced to only comparing the least significant bit of each
2289 byte. This allows an attacker to forge messages that would be considered as
2290 authenticated in an amount of tries lower than that guaranteed by the
2291 security claims of the scheme. The module can only be compiled by the
2292 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2294 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2300 * Add a build target 'build_all_generated', to build all generated files
2301 and only that. This can be used to prepare everything that requires
2302 things like perl for a system that lacks perl and then move everything
2303 to that system and do the rest of the build there.
2307 * Backport SSL_OP_NO_RENGOTIATION
2309 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2310 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2311 changes this is no longer possible in 1.1.0. Therefore the new
2312 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2313 1.1.0 to provide equivalent functionality.
2315 Note that if an application built against 1.1.0h headers (or above) is run
2316 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2317 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2321 * Removed the OS390-Unix config target. It relied on a script that doesn't
2326 * rsaz_1024_mul_avx2 overflow bug on x86_64
2328 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2329 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2330 Analysis suggests that attacks against RSA and DSA as a result of this
2331 defect would be very difficult to perform and are not believed likely.
2332 Attacks against DH1024 are considered just feasible, because most of the
2333 work necessary to deduce information about a private key may be performed
2334 offline. The amount of resources required for such an attack would be
2335 significant. However, for an attack on TLS to be meaningful, the server
2336 would have to share the DH1024 private key among multiple clients, which is
2337 no longer an option since CVE-2016-0701.
2339 This only affects processors that support the AVX2 but not ADX extensions
2340 like Intel Haswell (4th generation).
2342 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2343 was originally found via the OSS-Fuzz project.
2348 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2350 * bn_sqrx8x_internal carry bug on x86_64
2352 There is a carry propagating bug in the x86_64 Montgomery squaring
2353 procedure. No EC algorithms are affected. Analysis suggests that attacks
2354 against RSA and DSA as a result of this defect would be very difficult to
2355 perform and are not believed likely. Attacks against DH are considered just
2356 feasible (although very difficult) because most of the work necessary to
2357 deduce information about a private key may be performed offline. The amount
2358 of resources required for such an attack would be very significant and
2359 likely only accessible to a limited number of attackers. An attacker would
2360 additionally need online access to an unpatched system using the target
2361 private key in a scenario with persistent DH parameters and a private
2362 key that is shared between multiple clients.
2364 This only affects processors that support the BMI1, BMI2 and ADX extensions
2365 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2367 This issue was reported to OpenSSL by the OSS-Fuzz project.
2372 * Malformed X.509 IPAddressFamily could cause OOB read
2374 If an X.509 certificate has a malformed IPAddressFamily extension,
2375 OpenSSL could do a one-byte buffer overread. The most likely result
2376 would be an erroneous display of the certificate in text format.
2378 This issue was reported to OpenSSL by the OSS-Fuzz project.
2383 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2385 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2386 platform rather than 'mingw'.
2390 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2391 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2392 which is the minimum version we support.
2396 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2398 * Encrypt-Then-Mac renegotiation crash
2400 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2401 negotiated where it was not in the original handshake (or vice-versa) then
2402 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2403 and servers are affected.
2405 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2410 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2412 * Truncated packet could crash via OOB read
2414 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2415 cipher is being used, then a truncated packet can cause that host to
2416 perform an out-of-bounds read, usually resulting in a crash.
2418 This issue was reported to OpenSSL by Robert Święcki of Google.
2423 * Bad (EC)DHE parameters cause a client crash
2425 If a malicious server supplies bad parameters for a DHE or ECDHE key
2426 exchange then this can result in the client attempting to dereference a
2427 NULL pointer leading to a client crash. This could be exploited in a Denial
2430 This issue was reported to OpenSSL by Guido Vranken.
2435 * BN_mod_exp may produce incorrect results on x86_64
2437 There is a carry propagating bug in the x86_64 Montgomery squaring
2438 procedure. No EC algorithms are affected. Analysis suggests that attacks
2439 against RSA and DSA as a result of this defect would be very difficult to
2440 perform and are not believed likely. Attacks against DH are considered just
2441 feasible (although very difficult) because most of the work necessary to
2442 deduce information about a private key may be performed offline. The amount
2443 of resources required for such an attack would be very significant and
2444 likely only accessible to a limited number of attackers. An attacker would
2445 additionally need online access to an unpatched system using the target
2446 private key in a scenario with persistent DH parameters and a private
2447 key that is shared between multiple clients. For example this can occur by
2448 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2449 similar to CVE-2015-3193 but must be treated as a separate problem.
2451 This issue was reported to OpenSSL by the OSS-Fuzz project.
2456 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2458 * ChaCha20/Poly1305 heap-buffer-overflow
2460 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2461 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2462 crash. This issue is not considered to be exploitable beyond a DoS.
2464 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2469 * CMS Null dereference
2471 Applications parsing invalid CMS structures can crash with a NULL pointer
2472 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2473 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2474 structure callback if an attempt is made to free certain invalid encodings.
2475 Only CHOICE structures using a callback which do not handle NULL value are
2478 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2483 * Montgomery multiplication may produce incorrect results
2485 There is a carry propagating bug in the Broadwell-specific Montgomery
2486 multiplication procedure that handles input lengths divisible by, but
2487 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2488 and DH private keys are impossible. This is because the subroutine in
2489 question is not used in operations with the private key itself and an input
2490 of the attacker's direct choice. Otherwise the bug can manifest itself as
2491 transient authentication and key negotiation failures or reproducible
2492 erroneous outcome of public-key operations with specially crafted input.
2493 Among EC algorithms only Brainpool P-512 curves are affected and one
2494 presumably can attack ECDH key negotiation. Impact was not analyzed in
2495 detail, because pre-requisites for attack are considered unlikely. Namely
2496 multiple clients have to choose the curve in question and the server has to
2497 share the private key among them, neither of which is default behaviour.
2498 Even then only clients that chose the curve will be affected.
2500 This issue was publicly reported as transient failures and was not
2501 initially recognized as a security issue. Thanks to Richard Morgan for
2502 providing reproducible case.
2507 * Removed automatic addition of RPATH in shared libraries and executables,
2508 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2512 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2514 * Fix Use After Free for large message sizes
2516 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2517 message larger than approx 16k is received then the underlying buffer to
2518 store the incoming message is reallocated and moved. Unfortunately a
2519 dangling pointer to the old location is left which results in an attempt to
2520 write to the previously freed location. This is likely to result in a
2521 crash, however it could potentially lead to execution of arbitrary code.
2523 This issue only affects OpenSSL 1.1.0a.
2525 This issue was reported to OpenSSL by Robert Święcki.
2530 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2532 * OCSP Status Request extension unbounded memory growth
2534 A malicious client can send an excessively large OCSP Status Request
2535 extension. If that client continually requests renegotiation, sending a
2536 large OCSP Status Request extension each time, then there will be unbounded
2537 memory growth on the server. This will eventually lead to a Denial Of
2538 Service attack through memory exhaustion. Servers with a default
2539 configuration are vulnerable even if they do not support OCSP. Builds using
2540 the "no-ocsp" build time option are not affected.
2542 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2547 * SSL_peek() hang on empty record
2549 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2550 sends an empty record. This could be exploited by a malicious peer in a
2551 Denial Of Service attack.
2553 This issue was reported to OpenSSL by Alex Gaynor.
2558 * Excessive allocation of memory in tls_get_message_header() and
2559 dtls1_preprocess_fragment()
2561 A (D)TLS message includes 3 bytes for its length in the header for the
2562 message. This would allow for messages up to 16Mb in length. Messages of
2563 this length are excessive and OpenSSL includes a check to ensure that a
2564 peer is sending reasonably sized messages in order to avoid too much memory
2565 being consumed to service a connection. A flaw in the logic of version
2566 1.1.0 means that memory for the message is allocated too early, prior to
2567 the excessive message length check. Due to way memory is allocated in
2568 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2569 to service a connection. This could lead to a Denial of Service through
2570 memory exhaustion. However, the excessive message length check still takes
2571 place, and this would cause the connection to immediately fail. Assuming
2572 that the application calls SSL_free() on the failed connection in a timely
2573 manner then the 21Mb of allocated memory will then be immediately freed
2574 again. Therefore the excessive memory allocation will be transitory in
2575 nature. This then means that there is only a security impact if:
2577 1) The application does not call SSL_free() in a timely manner in the event
2578 that the connection fails
2580 2) The application is working in a constrained environment where there is
2581 very little free memory
2583 3) The attacker initiates multiple connection attempts such that there are
2584 multiple connections in a state where memory has been allocated for the
2585 connection; SSL_free() has not yet been called; and there is insufficient
2586 memory to service the multiple requests.
2588 Except in the instance of (1) above any Denial Of Service is likely to be
2589 transitory because as soon as the connection fails the memory is
2590 subsequently freed again in the SSL_free() call. However there is an
2591 increased risk during this period of application crashes due to the lack of
2592 memory - which would then mean a more serious Denial of Service.
2594 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2595 (CVE-2016-6307 and CVE-2016-6308)
2599 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2600 had to be removed. Primary reason is that vendor assembler can't
2601 assemble our modules with -KPIC flag. As result it, assembly
2602 support, was not even available as option. But its lack means
2603 lack of side-channel resistant code, which is incompatible with
2604 security by todays standards. Fortunately gcc is readily available
2605 prepackaged option, which we firmly point at...
2609 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2611 * Windows command-line tool supports UTF-8 opt-in option for arguments
2612 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2613 (to any value) allows Windows user to access PKCS#12 file generated
2614 with Windows CryptoAPI and protected with non-ASCII password, as well
2615 as files generated under UTF-8 locale on Linux also protected with
2620 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2621 have been disabled by default and removed from DEFAULT, just like RC4.
2622 See the RC4 item below to re-enable both.
2626 * The method for finding the storage location for the Windows RAND seed file
2627 has changed. First we check %RANDFILE%. If that is not set then we check
2628 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2629 all else fails we fall back to C:\.
2633 * The EVP_EncryptUpdate() function has had its return type changed from void
2634 to int. A return of 0 indicates and error while a return of 1 indicates
2639 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2640 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2641 off the constant time implementation for RSA, DSA and DH have been made
2642 no-ops and deprecated.
2646 * Windows RAND implementation was simplified to only get entropy by
2647 calling CryptGenRandom(). Various other RAND-related tickets
2650 *Joseph Wylie Yandle, Rich Salz*
2652 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2653 and `OPENSSL_LH_`, respectively. The old names are available
2654 with API compatibility. They new names are now completely documented.
2658 * Unify TYPE_up_ref(obj) methods signature.
2659 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2660 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2661 int (instead of void) like all others TYPE_up_ref() methods.
2662 So now these methods also check the return value of CRYPTO_atomic_add(),
2663 and the validity of object reference counter.
2665 *fdasilvayy@gmail.com*
2667 * With Windows Visual Studio builds, the .pdb files are installed
2668 alongside the installed libraries and executables. For a static
2669 library installation, ossl_static.pdb is the associate compiler
2670 generated .pdb file to be used when linking programs.
2674 * Remove openssl.spec. Packaging files belong with the packagers.
2678 * Automatic Darwin/OSX configuration has had a refresh, it will now
2679 recognise x86_64 architectures automatically. You can still decide
2680 to build for a different bitness with the environment variable
2681 KERNEL_BITS (can be 32 or 64), for example:
2683 KERNEL_BITS=32 ./config
2687 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2688 256 bit AES and HMAC with SHA256.
2692 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2696 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2700 * To enable users to have their own config files and build file templates,
2701 Configure looks in the directory indicated by the environment variable
2702 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2703 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2704 name and is used as is.
2708 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2709 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2710 X509_CERT_FILE_CTX was removed.
2714 * "shared" builds are now the default. To create only static libraries use
2715 the "no-shared" Configure option.
2719 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2720 All of these option have not worked for some while and are fundamental
2725 * Make various cleanup routines no-ops and mark them as deprecated. Most
2726 global cleanup functions are no longer required because they are handled
2727 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2728 Explicitly de-initing can cause problems (e.g. where a library that uses
2729 OpenSSL de-inits, but an application is still using it). The affected
2730 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2731 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2732 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2733 COMP_zlib_cleanup().
2737 * --strict-warnings no longer enables runtime debugging options
2738 such as REF_DEBUG. Instead, debug options are automatically
2739 enabled with '--debug' builds.
2741 *Andy Polyakov, Emilia Käsper*
2743 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2744 have been moved out of the public header files. New functions for managing
2745 these have been added.
2749 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2750 objects have been moved out of the public header files. New
2751 functions for managing these have been added.
2755 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2756 have been moved out of the public header files. New functions for managing
2757 these have been added.
2761 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2762 moved out of the public header files. New functions for managing these
2767 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2771 * Removed the mk1mf build scripts.
2775 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2776 it is always safe to #include a header now.
2780 * Removed the aged BC-32 config and all its supporting scripts
2784 * Removed support for Ultrix, Netware, and OS/2.
2788 * Add support for HKDF.
2790 *Alessandro Ghedini*
2792 * Add support for blake2b and blake2s
2796 * Added support for "pipelining". Ciphers that have the
2797 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2798 encryptions/decryptions simultaneously. There are currently no built-in
2799 ciphers with this property but the expectation is that engines will be able
2800 to offer it to significantly improve throughput. Support has been extended
2801 into libssl so that multiple records for a single connection can be
2802 processed in one go (for >=TLS 1.1).
2806 * Added the AFALG engine. This is an async capable engine which is able to
2807 offload work to the Linux kernel. In this initial version it only supports
2808 AES128-CBC. The kernel must be version 4.1.0 or greater.
2812 * OpenSSL now uses a new threading API. It is no longer necessary to
2813 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2814 are two supported threading models: pthreads and windows threads. It is
2815 also possible to configure OpenSSL at compile time for "no-threads". The
2816 old threading API should no longer be used. The functions have been
2817 replaced with "no-op" compatibility macros.
2819 *Alessandro Ghedini, Matt Caswell*
2821 * Modify behavior of ALPN to invoke callback after SNI/servername
2822 callback, such that updates to the SSL_CTX affect ALPN.
2826 * Add SSL_CIPHER queries for authentication and key-exchange.
2830 * Changes to the DEFAULT cipherlist:
2831 - Prefer (EC)DHE handshakes over plain RSA.
2832 - Prefer AEAD ciphers over legacy ciphers.
2833 - Prefer ECDSA over RSA when both certificates are available.
2834 - Prefer TLSv1.2 ciphers/PRF.
2835 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2840 * Change the ECC default curve list to be this, in order: x25519,
2841 secp256r1, secp521r1, secp384r1.
2845 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2846 disabled by default. They can be re-enabled using the
2847 enable-weak-ssl-ciphers option to Configure.
2851 * If the server has ALPN configured, but supports no protocols that the
2852 client advertises, send a fatal "no_application_protocol" alert.
2853 This behaviour is SHALL in RFC 7301, though it isn't universally
2854 implemented by other servers.
2858 * Add X25519 support.
2859 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2860 for public and private key encoding using the format documented in
2861 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2862 key generation and key derivation.
2864 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2869 * Deprecate SRP_VBASE_get_by_user.
2870 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2871 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2872 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2873 seed, even if the seed is configured.
2875 Users should use SRP_VBASE_get1_by_user instead. Note that in
2876 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2877 also that even though configuring the SRP seed attempts to hide
2878 invalid usernames by continuing the handshake with fake
2879 credentials, this behaviour is not constant time and no strong
2880 guarantees are made that the handshake is indistinguishable from
2881 that of a valid user.
2885 * Configuration change; it's now possible to build dynamic engines
2886 without having to build shared libraries and vice versa. This
2887 only applies to the engines in engines/, those in crypto/engine/
2888 will always be built into libcrypto (i.e. "static").
2890 Building dynamic engines is enabled by default; to disable, use
2891 the configuration option "disable-dynamic-engine".
2893 The only requirements for building dynamic engines are the
2894 presence of the DSO module and building with position independent
2895 code, so they will also automatically be disabled if configuring
2896 with "disable-dso" or "disable-pic".
2898 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2899 are also taken away from openssl/opensslconf.h, as they are
2904 * Configuration change; if there is a known flag to compile
2905 position independent code, it will always be applied on the
2906 libcrypto and libssl object files, and never on the application
2907 object files. This means other libraries that use routines from
2908 libcrypto / libssl can be made into shared libraries regardless
2909 of how OpenSSL was configured.
2911 If this isn't desirable, the configuration options "disable-pic"
2912 or "no-pic" can be used to disable the use of PIC. This will
2913 also disable building shared libraries and dynamic engines.
2917 * Removed JPAKE code. It was experimental and has no wide use.
2921 * The INSTALL_PREFIX Makefile variable has been renamed to
2922 DESTDIR. That makes for less confusion on what this variable
2923 is for. Also, the configuration option --install_prefix is
2928 * Heartbeat for TLS has been removed and is disabled by default
2929 for DTLS; configure with enable-heartbeats. Code that uses the
2930 old #define's might need to be updated.
2932 *Emilia Käsper, Rich Salz*
2934 * Rename REF_CHECK to REF_DEBUG.
2938 * New "unified" build system
2940 The "unified" build system is aimed to be a common system for all
2941 platforms we support. With it comes new support for VMS.
2943 This system builds supports building in a different directory tree
2944 than the source tree. It produces one Makefile (for unix family
2945 or lookalikes), or one descrip.mms (for VMS).
2947 The source of information to make the Makefile / descrip.mms is
2948 small files called 'build.info', holding the necessary
2949 information for each directory with source to compile, and a
2950 template in Configurations, like unix-Makefile.tmpl or
2953 With this change, the library names were also renamed on Windows
2954 and on VMS. They now have names that are closer to the standard
2955 on Unix, and include the major version number, and in certain
2956 cases, the architecture they are built for. See "Notes on shared
2957 libraries" in INSTALL.
2959 We rely heavily on the perl module Text::Template.
2963 * Added support for auto-initialisation and de-initialisation of the library.
2964 OpenSSL no longer requires explicit init or deinit routines to be called,
2965 except in certain circumstances. See the OPENSSL_init_crypto() and
2966 OPENSSL_init_ssl() man pages for further information.
2970 * The arguments to the DTLSv1_listen function have changed. Specifically the
2971 "peer" argument is now expected to be a BIO_ADDR object.
2973 * Rewrite of BIO networking library. The BIO library lacked consistent
2974 support of IPv6, and adding it required some more extensive
2975 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2976 which hold all types of addresses and chains of address information.
2977 It also introduces a new API, with functions like BIO_socket,
2978 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2979 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2980 have been adapted accordingly.
2984 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2989 * CRIME protection: disable compression by default, even if OpenSSL is
2990 compiled with zlib enabled. Applications can still enable compression
2991 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2992 using the SSL_CONF library to configure compression.
2996 * The signature of the session callback configured with
2997 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2998 was explicitly marked as `const unsigned char*` instead of
3003 * Always DPURIFY. Remove the use of uninitialized memory in the
3004 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3008 * Removed many obsolete configuration items, including
3009 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3010 MD2_CHAR, MD2_INT, MD2_LONG
3012 IDEA_SHORT, IDEA_LONG
3013 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3015 *Rich Salz, with advice from Andy Polyakov*
3017 * Many BN internals have been moved to an internal header file.
3019 *Rich Salz with help from Andy Polyakov*
3021 * Configuration and writing out the results from it has changed.
3022 Files such as Makefile include/openssl/opensslconf.h and are now
3023 produced through general templates, such as Makefile.in and
3024 crypto/opensslconf.h.in and some help from the perl module
3027 Also, the center of configuration information is no longer
3028 Makefile. Instead, Configure produces a perl module in
3029 configdata.pm which holds most of the config data (in the hash
3030 table %config), the target data that comes from the target
3031 configuration in one of the `Configurations/*.conf` files (in
3036 * To clarify their intended purposes, the Configure options
3037 --prefix and --openssldir change their semantics, and become more
3038 straightforward and less interdependent.
3040 --prefix shall be used exclusively to give the location INSTALLTOP
3041 where programs, scripts, libraries, include files and manuals are
3042 going to be installed. The default is now /usr/local.
3044 --openssldir shall be used exclusively to give the default
3045 location OPENSSLDIR where certificates, private keys, CRLs are
3046 managed. This is also where the default openssl.cnf gets
3048 If the directory given with this option is a relative path, the
3049 values of both the --prefix value and the --openssldir value will
3050 be combined to become OPENSSLDIR.
3051 The default for --openssldir is INSTALLTOP/ssl.
3053 Anyone who uses --openssldir to specify where OpenSSL is to be
3054 installed MUST change to use --prefix instead.
3058 * The GOST engine was out of date and therefore it has been removed. An up
3059 to date GOST engine is now being maintained in an external repository.
3060 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3061 support for GOST ciphersuites (these are only activated if a GOST engine
3066 * EGD is no longer supported by default; use enable-egd when
3069 *Ben Kaduk and Rich Salz*
3071 * The distribution now has Makefile.in files, which are used to
3072 create Makefile's when Configure is run. *Configure must be run
3073 before trying to build now.*
3077 * The return value for SSL_CIPHER_description() for error conditions
3082 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3084 Obtaining and performing DNSSEC validation of TLSA records is
3085 the application's responsibility. The application provides
3086 the TLSA records of its choice to OpenSSL, and these are then
3087 used to authenticate the peer.
3089 The TLSA records need not even come from DNS. They can, for
3090 example, be used to implement local end-entity certificate or
3091 trust-anchor "pinning", where the "pin" data takes the form
3092 of TLSA records, which can augment or replace verification
3093 based on the usual WebPKI public certification authorities.
3097 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3098 continues to support deprecated interfaces in default builds.
3099 However, applications are strongly advised to compile their
3100 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3101 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3102 or the 1.1.0 releases.
3104 In environments in which all applications have been ported to
3105 not use any deprecated interfaces OpenSSL's Configure script
3106 should be used with the --api=1.1.0 option to entirely remove
3107 support for the deprecated features from the library and
3108 unconditionally disable them in the installed headers.
3109 Essentially the same effect can be achieved with the "no-deprecated"
3110 argument to Configure, except that this will always restrict
3111 the build to just the latest API, rather than a fixed API
3114 As applications are ported to future revisions of the API,
3115 they should update their compile-time OPENSSL_API_COMPAT define
3116 accordingly, but in most cases should be able to continue to
3117 compile with later releases.
3119 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3120 0x10000000L and 0x00908000L, respectively. However those
3121 versions did not support the OPENSSL_API_COMPAT feature, and
3122 so applications are not typically tested for explicit support
3123 of just the undeprecated features of either release.
3127 * Add support for setting the minimum and maximum supported protocol.
3128 It can bet set via the SSL_set_min_proto_version() and
3129 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3130 MaxProtocol. It's recommended to use the new APIs to disable
3131 protocols instead of disabling individual protocols using
3132 SSL_set_options() or SSL_CONF's Protocol. This change also
3133 removes support for disabling TLS 1.2 in the OpenSSL TLS
3134 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3138 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3142 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3143 and integrates ECDSA and ECDH functionality into EC. Implementations can
3144 now redirect key generation and no longer need to convert to or from
3147 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3148 include the ec.h header file instead.
3152 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3153 ciphers who are no longer supported and drops support the ephemeral RSA key
3154 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3158 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3159 opaque. For HMAC_CTX, the following constructors and destructors
3162 HMAC_CTX *HMAC_CTX_new(void);
3163 void HMAC_CTX_free(HMAC_CTX *ctx);
3165 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3166 destroy such methods has been added. See EVP_MD_meth_new(3) and
3167 EVP_CIPHER_meth_new(3) for documentation.
3170 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3171 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3172 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3173 an already created structure.
3174 2) For consistency with the majority of our object creators and
3175 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3176 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3177 for deprecated builds.
3181 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3182 cryptographic operations to be performed asynchronously as long as an
3183 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3184 further details. Libssl has also had this capability integrated with the
3185 introduction of the new mode SSL_MODE_ASYNC and associated error
3186 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3187 pages. This work was developed in partnership with Intel Corp.
3191 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3192 always enabled now. If you want to disable the support you should
3193 exclude it using the list of supported ciphers. This also means that the
3194 "-no_ecdhe" option has been removed from s_server.
3198 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3199 SSL_{CTX_}set1_curves() which can set a list.
3203 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3204 curve you want to support using SSL_{CTX_}set1_curves().
3208 * State machine rewrite. The state machine code has been significantly
3209 refactored in order to remove much duplication of code and solve issues
3210 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3211 further details). This change does have some associated API changes.
3212 Notably the SSL_state() function has been removed and replaced by
3213 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3214 SSL_set_state() has been removed altogether. The previous handshake states
3215 defined in ssl.h and ssl3.h have also been removed.
3219 * All instances of the string "ssleay" in the public API were replaced
3220 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3221 Some error codes related to internal RSA_eay API's were renamed.
3225 * The demo files in crypto/threads were moved to demo/threads.
3229 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3232 *Matt Caswell, Rich Salz*
3234 * New ASN.1 embed macro.
3236 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3237 structure is not allocated: it is part of the parent. That is instead of
3245 This reduces memory fragmentation and make it impossible to accidentally
3246 set a mandatory field to NULL.
3248 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3249 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3250 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3255 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3259 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3260 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3261 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3262 DES and RC4 ciphersuites.
3266 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3267 This changes the decoding behaviour for some invalid messages,
3268 though the change is mostly in the more lenient direction, and
3269 legacy behaviour is preserved as much as possible.
3273 * Fix no-stdio build.
3274 *David Woodhouse <David.Woodhouse@intel.com> and also*
3275 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3277 * New testing framework
3278 The testing framework has been largely rewritten and is now using
3279 perl and the perl modules Test::Harness and an extended variant of
3280 Test::More called OpenSSL::Test to do its work. All test scripts in
3281 test/ have been rewritten into test recipes, and all direct calls to
3282 executables in test/Makefile have become individual recipes using the
3283 simplified testing OpenSSL::Test::Simple.
3285 For documentation on our testing modules, do:
3287 perldoc test/testlib/OpenSSL/Test/Simple.pm
3288 perldoc test/testlib/OpenSSL/Test.pm
3292 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3293 are used; the latter aborts on memory leaks (usually checked on exit).
3294 Some undocumented "set malloc, etc., hooks" functions were removed
3295 and others were changed. All are now documented.
3299 * In DSA_generate_parameters_ex, if the provided seed is too short,
3302 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3304 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3305 from RFC4279, RFC4785, RFC5487, RFC5489.
3307 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3308 original RSA_PSK patch.
3312 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3313 era flag was never set throughout the codebase (only read). Also removed
3314 SSL3_FLAGS_POP_BUFFER which was only used if
3315 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3319 * Changed the default name options in the "ca", "crl", "req" and "x509"
3320 to be "oneline" instead of "compat".
3324 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3325 not aware of clients that still exhibit this bug, and the workaround
3326 hasn't been working properly for a while.
3330 * The return type of BIO_number_read() and BIO_number_written() as well as
3331 the corresponding num_read and num_write members in the BIO structure has
3332 changed from unsigned long to uint64_t. On platforms where an unsigned
3333 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3338 * Given the pervasive nature of TLS extensions it is inadvisable to run
3339 OpenSSL without support for them. It also means that maintaining
3340 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3341 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3345 * Removed support for the two export grade static DH ciphersuites
3346 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3347 were newly added (along with a number of other static DH ciphersuites) to
3348 1.0.2. However the two export ones have *never* worked since they were
3349 introduced. It seems strange in any case to be adding new export
3350 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3354 * Version negotiation has been rewritten. In particular SSLv23_method(),
3355 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3356 and turned into macros which simply call the new preferred function names
3357 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3358 should use the new names instead. Also as part of this change the ssl23.h
3359 header file has been removed.
3363 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3364 code and the associated standard is no longer considered fit-for-purpose.
3368 * RT2547 was closed. When generating a private key, try to make the
3369 output file readable only by the owner. This behavior change might
3370 be noticeable when interacting with other software.
3372 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3377 * Added HTTP GET support to the ocsp command.
3381 * Changed default digest for the dgst and enc commands from MD5 to
3386 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3390 * Added support for TLS extended master secret from
3391 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3392 initial patch which was a great help during development.
3396 * All libssl internal structures have been removed from the public header
3397 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3398 now redundant). Users should not attempt to access internal structures
3399 directly. Instead they should use the provided API functions.
3403 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3404 Access to deprecated functions can be re-enabled by running config with
3405 "enable-deprecated". In addition applications wishing to use deprecated
3406 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3407 will, by default, disable some transitive includes that previously existed
3408 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3412 * Added support for OCB mode. OpenSSL has been granted a patent license
3413 compatible with the OpenSSL license for use of OCB. Details are available
3414 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3415 for OCB can be removed by calling config with no-ocb.
3419 * SSLv2 support has been removed. It still supports receiving a SSLv2
3420 compatible client hello.
3424 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3425 done while fixing the error code for the key-too-small case.
3427 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3429 * CA.sh has been removed; use CA.pl instead.
3433 * Removed old DES API.
3437 * Remove various unsupported platforms:
3443 Sinix/ReliantUNIX RM400
3448 16-bit platforms such as WIN16
3452 * Clean up OPENSSL_NO_xxx #define's
3453 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3454 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3455 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3456 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3457 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3458 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3459 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3460 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3461 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3462 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3466 * Cleaned up dead code
3467 Remove all but one '#ifdef undef' which is to be looked at.
3471 * Clean up calling of xxx_free routines.
3472 Just like free(), fix most of the xxx_free routines to accept
3473 NULL. Remove the non-null checks from callers. Save much code.
3477 * Add secure heap for storage of private keys (when possible).
3478 Add BIO_s_secmem(), CBIGNUM, etc.
3479 Contributed by Akamai Technologies under our Corporate CLA.
3483 * Experimental support for a new, fast, unbiased prime candidate generator,
3484 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3486 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3488 * New output format NSS in the sess_id command line tool. This allows
3489 exporting the session id and the master key in NSS keylog format.
3491 *Martin Kaiser <martin@kaiser.cx>*
3493 * Harmonize version and its documentation. -f flag is used to display
3496 *mancha <mancha1@zoho.com>*
3498 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3499 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3501 *mancha <mancha1@zoho.com>*
3503 * Fix some double frees. These are not thought to be exploitable.
3505 *mancha <mancha1@zoho.com>*
3507 * A missing bounds check in the handling of the TLS heartbeat extension
3508 can be used to reveal up to 64k of memory to a connected client or
3511 Thanks for Neel Mehta of Google Security for discovering this bug and to
3512 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3513 preparing the fix [CVE-2014-0160][]
3515 *Adam Langley, Bodo Moeller*
3517 * Fix for the attack described in the paper "Recovering OpenSSL
3518 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3519 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3520 <http://eprint.iacr.org/2014/140>
3522 Thanks to Yuval Yarom and Naomi Benger for discovering this
3523 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3525 *Yuval Yarom and Naomi Benger*
3527 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3528 this fixes a limitation in previous versions of OpenSSL.
3532 * Experimental encrypt-then-mac support.
3534 Experimental support for encrypt then mac from
3535 draft-gutmann-tls-encrypt-then-mac-02.txt
3537 To enable it set the appropriate extension number (0x42 for the test
3538 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3540 For non-compliant peers (i.e. just about everything) this should have no
3543 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3547 * Add EVP support for key wrapping algorithms, to avoid problems with
3548 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3549 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3550 algorithms and include tests cases.
3554 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3559 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3560 MGF1 digest and OAEP label.
3564 * Make openssl verify return errors.
3566 *Chris Palmer <palmer@google.com> and Ben Laurie*
3568 * New function ASN1_TIME_diff to calculate the difference between two
3569 ASN1_TIME structures or one structure and the current time.
3573 * Update fips_test_suite to support multiple command line options. New
3574 test to induce all self test errors in sequence and check expected
3579 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3580 sign or verify all in one operation.
3584 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3585 test programs and fips_test_suite. Includes functionality to parse
3586 the minimal script output of fipsalgest.pl directly.
3590 * Add authorisation parameter to FIPS_module_mode_set().
3594 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3598 * Use separate DRBG fields for internal and external flags. New function
3599 FIPS_drbg_health_check() to perform on demand health checking. Add
3600 generation tests to fips_test_suite with reduced health check interval to
3601 demonstrate periodic health checking. Add "nodh" option to
3602 fips_test_suite to skip very slow DH test.
3606 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3611 * More extensive health check for DRBG checking many more failure modes.
3612 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3613 combination: call this in fips_test_suite.
3617 * Add support for canonical generation of DSA parameter 'g'. See
3620 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3621 POST to handle HMAC cases.
3625 * Add functions FIPS_module_version() and FIPS_module_version_text()
3626 to return numerical and string versions of the FIPS module number.
3630 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3631 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3632 outside the validated module in the FIPS capable OpenSSL.
3636 * Minor change to DRBG entropy callback semantics. In some cases
3637 there is no multiple of the block length between min_len and
3638 max_len. Allow the callback to return more than max_len bytes
3639 of entropy but discard any extra: it is the callback's responsibility
3640 to ensure that the extra data discarded does not impact the
3641 requested amount of entropy.
3645 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3646 information in FIPS186-3, SP800-57 and SP800-131A.
3650 * CCM support via EVP. Interface is very similar to GCM case except we
3651 must supply all data in one chunk (i.e. no update, final) and the
3652 message length must be supplied if AAD is used. Add algorithm test
3657 * Initial version of POST overhaul. Add POST callback to allow the status
3658 of POST to be monitored and/or failures induced. Modify fips_test_suite
3659 to use callback. Always run all selftests even if one fails.
3663 * XTS support including algorithm test driver in the fips_gcmtest program.
3664 Note: this does increase the maximum key length from 32 to 64 bytes but
3665 there should be no binary compatibility issues as existing applications
3666 will never use XTS mode.
3670 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3671 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3672 performs algorithm blocking for unapproved PRNG types. Also do not
3673 set PRNG type in FIPS_mode_set(): leave this to the application.
3674 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3675 the standard OpenSSL PRNG: set additional data to a date time vector.
3679 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3680 This shouldn't present any incompatibility problems because applications
3681 shouldn't be using these directly and any that are will need to rethink
3682 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3686 * Extensive self tests and health checking required by SP800-90 DRBG.
3687 Remove strength parameter from FIPS_drbg_instantiate and always
3688 instantiate at maximum supported strength.
3692 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3696 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3700 * New function DH_compute_key_padded() to compute a DH key and pad with
3701 leading zeroes if needed: this complies with SP800-56A et al.
3705 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3706 anything, incomplete, subject to change and largely untested at present.
3710 * Modify fipscanisteronly build option to only build the necessary object
3711 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3715 * Add experimental option FIPSSYMS to give all symbols in
3716 fipscanister.o and FIPS or fips prefix. This will avoid
3717 conflicts with future versions of OpenSSL. Add perl script
3718 util/fipsas.pl to preprocess assembly language source files
3719 and rename any affected symbols.
3723 * Add selftest checks and algorithm block of non-fips algorithms in
3724 FIPS mode. Remove DES2 from selftests.
3728 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3729 return internal method without any ENGINE dependencies. Add new
3730 tiny fips sign and verify functions.
3734 * New build option no-ec2m to disable characteristic 2 code.
3738 * New build option "fipscanisteronly". This only builds fipscanister.o
3739 and (currently) associated fips utilities. Uses the file Makefile.fips
3740 instead of Makefile.org as the prototype.
3744 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3745 Update fips_gcmtest to use IV generator.
3749 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3750 setting output buffer to NULL. The `*Final` function must be
3751 called although it will not retrieve any additional data. The tag
3752 can be set or retrieved with a ctrl. The IV length is by default 12
3753 bytes (96 bits) but can be set to an alternative value. If the IV
3754 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3759 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3760 underlying do_cipher function handles all cipher semantics itself
3761 including padding and finalisation. This is useful if (for example)
3762 an ENGINE cipher handles block padding itself. The behaviour of
3763 do_cipher is subtly changed if this flag is set: the return value
3764 is the number of characters written to the output buffer (zero is
3765 no longer an error code) or a negative error code. Also if the
3766 input buffer is NULL and length 0 finalisation should be performed.
3770 * If a candidate issuer certificate is already part of the constructed
3771 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3775 * Improve forward-security support: add functions
3777 void SSL_CTX_set_not_resumable_session_callback(
3778 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3779 void SSL_set_not_resumable_session_callback(
3780 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3782 for use by SSL/TLS servers; the callback function will be called whenever a
3783 new session is created, and gets to decide whether the session may be
3784 cached to make it resumable (return 0) or not (return 1). (As by the
3785 SSL/TLS protocol specifications, the session_id sent by the server will be
3786 empty to indicate that the session is not resumable; also, the server will
3787 not generate RFC 4507 (RFC 5077) session tickets.)
3789 A simple reasonable callback implementation is to return is_forward_secure.
3790 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3791 by the SSL/TLS server library, indicating whether it can provide forward
3794 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3796 * New -verify_name option in command line utilities to set verification
3801 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3802 Add CMAC pkey methods.
3806 * Experimental renegotiation in s_server -www mode. If the client
3807 browses /reneg connection is renegotiated. If /renegcert it is
3808 renegotiated requesting a certificate.
3812 * Add an "external" session cache for debugging purposes to s_server. This
3813 should help trace issues which normally are only apparent in deployed
3814 multi-process servers.
3818 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3819 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3820 BIO_set_cipher() and some obscure PEM functions were changed so they
3821 can now return an error. The RAND changes required a change to the
3822 RAND_METHOD structure.
3826 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3827 a gcc attribute to warn if the result of a function is ignored. This
3828 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3829 whose return value is often ignored.
3833 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3834 These allow SCTs (signed certificate timestamps) to be requested and
3835 validated when establishing a connection.
3837 *Rob Percival <robpercival@google.com>*
3842 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3844 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3845 used even when parsing explicit parameters, when loading a encoded key
3846 or calling `EC_GROUP_new_from_ecpkparameters()`/
3847 `EC_GROUP_new_from_ecparameters()`.
3848 This prevents bypass of security hardening and performance gains,
3849 especially for curves with specialized EC_METHODs.
3850 By default, if a key encoded with explicit parameters is loaded and later
3851 encoded, the output is still encoded with explicit parameters, even if
3852 internally a "named" EC_GROUP is used for computation.
3856 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3857 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3858 NULL. After this change, only the cofactor parameter can be NULL. It also
3859 does some minimal sanity checks on the passed order.
3864 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3865 An attack is simple, if the first CMS_recipientInfo is valid but the
3866 second CMS_recipientInfo is chosen ciphertext. If the second
3867 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3868 encryption key will be replaced by garbage, and the message cannot be
3869 decoded, but if the RSA decryption fails, the correct encryption key is
3870 used and the recipient will not notice the attack.
3871 As a work around for this potential attack the length of the decrypted
3872 key must be equal to the cipher default key length, in case the
3873 certifiate is not given and all recipientInfo are tried out.
3874 The old behaviour can be re-enabled in the CMS code by setting the
3875 CMS_DEBUG_DECRYPT flag.
3880 * Document issue with installation paths in diverse Windows builds
3882 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3883 binaries and run-time config file.
3888 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3890 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3891 This changes the size when using the genpkey app when no size is given. It
3892 fixes an omission in earlier changes that changed all RSA, DSA and DH
3893 generation apps to use 2048 bits by default.
3897 * Add FIPS support for Android Arm 64-bit
3899 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3900 Module in Version 2.0.10. For some reason, the corresponding target
3901 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3902 built with FIPS support on Android Arm 64-bit. This omission has been
3905 *Matthias St. Pierre*
3907 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3909 * 0-byte record padding oracle
3911 If an application encounters a fatal protocol error and then calls
3912 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3913 then OpenSSL can respond differently to the calling application if a 0 byte
3914 record is received with invalid padding compared to if a 0 byte record is
3915 received with an invalid MAC. If the application then behaves differently
3916 based on that in a way that is detectable to the remote peer, then this
3917 amounts to a padding oracle that could be used to decrypt data.
3919 In order for this to be exploitable "non-stitched" ciphersuites must be in
3920 use. Stitched ciphersuites are optimised implementations of certain
3921 commonly used ciphersuites. Also the application must call SSL_shutdown()
3922 twice even if a protocol error has occurred (applications should not do
3923 this but some do anyway).
3925 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3926 Aviram, with additional investigation by Steven Collison and Andrew
3927 Hourselt. It was reported to OpenSSL on 10th December 2018.
3932 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3936 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3938 * Microarchitecture timing vulnerability in ECC scalar multiplication
3940 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3941 shown to be vulnerable to a microarchitecture timing side channel attack.
3942 An attacker with sufficient access to mount local timing attacks during
3943 ECDSA signature generation could recover the private key.
3945 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3946 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3952 * Timing vulnerability in DSA signature generation
3954 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3955 timing side channel attack. An attacker could use variations in the signing
3956 algorithm to recover the private key.
3958 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3963 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3964 Module, accidentally introduced while backporting security fixes from the
3965 development branch and hindering the use of ECC in FIPS mode.
3969 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3971 * Client DoS due to large DH parameter
3973 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3974 malicious server can send a very large prime value to the client. This will
3975 cause the client to spend an unreasonably long period of time generating a
3976 key for this prime resulting in a hang until the client has finished. This
3977 could be exploited in a Denial Of Service attack.
3979 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3984 * Cache timing vulnerability in RSA Key Generation
3986 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3987 a cache timing side channel attack. An attacker with sufficient access to
3988 mount cache timing attacks during the RSA key generation process could
3989 recover the private key.
3991 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3992 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3997 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3998 parameter is no longer accepted, as it leads to a corrupt table. NULL
3999 pem_str is reserved for alias entries only.
4003 * Revert blinding in ECDSA sign and instead make problematic addition
4004 length-invariant. Switch even to fixed-length Montgomery multiplication.
4008 * Change generating and checking of primes so that the error rate of not
4009 being prime depends on the intended use based on the size of the input.
4010 For larger primes this will result in more rounds of Miller-Rabin.
4011 The maximal error rate for primes with more than 1080 bits is lowered
4014 *Kurt Roeckx, Annie Yousar*
4016 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4020 * Add blinding to ECDSA and DSA signatures to protect against side channel
4021 attacks discovered by Keegan Ryan (NCC Group).
4025 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4026 now allow empty (zero character) pass phrases.
4030 * Certificate time validation (X509_cmp_time) enforces stricter
4031 compliance with RFC 5280. Fractional seconds and timezone offsets
4032 are no longer allowed.
4036 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4038 * Constructed ASN.1 types with a recursive definition could exceed the stack
4040 Constructed ASN.1 types with a recursive definition (such as can be found
4041 in PKCS7) could eventually exceed the stack given malicious input with
4042 excessive recursion. This could result in a Denial Of Service attack. There
4043 are no such structures used within SSL/TLS that come from untrusted sources
4044 so this is considered safe.
4046 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4052 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4054 * Read/write after SSL object in error state
4056 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4057 mechanism. The intent was that if a fatal error occurred during a handshake
4058 then OpenSSL would move into the error state and would immediately fail if
4059 you attempted to continue the handshake. This works as designed for the
4060 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4061 SSL_connect()), however due to a bug it does not work correctly if
4062 SSL_read() or SSL_write() is called directly. In that scenario, if the
4063 handshake fails then a fatal error will be returned in the initial function
4064 call. If SSL_read()/SSL_write() is subsequently called by the application
4065 for the same SSL object then it will succeed and the data is passed without
4066 being decrypted/encrypted directly from the SSL/TLS record layer.
4068 In order to exploit this issue an application bug would have to be present
4069 that resulted in a call to SSL_read()/SSL_write() being issued after having
4070 already received a fatal error.
4072 This issue was reported to OpenSSL by David Benjamin (Google).
4077 * rsaz_1024_mul_avx2 overflow bug on x86_64
4079 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4080 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4081 Analysis suggests that attacks against RSA and DSA as a result of this
4082 defect would be very difficult to perform and are not believed likely.
4083 Attacks against DH1024 are considered just feasible, because most of the
4084 work necessary to deduce information about a private key may be performed
4085 offline. The amount of resources required for such an attack would be
4086 significant. However, for an attack on TLS to be meaningful, the server
4087 would have to share the DH1024 private key among multiple clients, which is
4088 no longer an option since CVE-2016-0701.
4090 This only affects processors that support the AVX2 but not ADX extensions
4091 like Intel Haswell (4th generation).
4093 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4094 was originally found via the OSS-Fuzz project.
4099 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4101 * bn_sqrx8x_internal carry bug on x86_64
4103 There is a carry propagating bug in the x86_64 Montgomery squaring
4104 procedure. No EC algorithms are affected. Analysis suggests that attacks
4105 against RSA and DSA as a result of this defect would be very difficult to
4106 perform and are not believed likely. Attacks against DH are considered just
4107 feasible (although very difficult) because most of the work necessary to
4108 deduce information about a private key may be performed offline. The amount
4109 of resources required for such an attack would be very significant and
4110 likely only accessible to a limited number of attackers. An attacker would
4111 additionally need online access to an unpatched system using the target
4112 private key in a scenario with persistent DH parameters and a private
4113 key that is shared between multiple clients.
4115 This only affects processors that support the BMI1, BMI2 and ADX extensions
4116 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4118 This issue was reported to OpenSSL by the OSS-Fuzz project.
4123 * Malformed X.509 IPAddressFamily could cause OOB read
4125 If an X.509 certificate has a malformed IPAddressFamily extension,
4126 OpenSSL could do a one-byte buffer overread. The most likely result
4127 would be an erroneous display of the certificate in text format.
4129 This issue was reported to OpenSSL by the OSS-Fuzz project.
4134 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4136 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4137 platform rather than 'mingw'.
4141 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4143 * Truncated packet could crash via OOB read
4145 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4146 cipher is being used, then a truncated packet can cause that host to
4147 perform an out-of-bounds read, usually resulting in a crash.
4149 This issue was reported to OpenSSL by Robert Święcki of Google.
4154 * BN_mod_exp may produce incorrect results on x86_64
4156 There is a carry propagating bug in the x86_64 Montgomery squaring
4157 procedure. No EC algorithms are affected. Analysis suggests that attacks
4158 against RSA and DSA as a result of this defect would be very difficult to
4159 perform and are not believed likely. Attacks against DH are considered just
4160 feasible (although very difficult) because most of the work necessary to
4161 deduce information about a private key may be performed offline. The amount
4162 of resources required for such an attack would be very significant and
4163 likely only accessible to a limited number of attackers. An attacker would
4164 additionally need online access to an unpatched system using the target
4165 private key in a scenario with persistent DH parameters and a private
4166 key that is shared between multiple clients. For example this can occur by
4167 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4168 similar to CVE-2015-3193 but must be treated as a separate problem.
4170 This issue was reported to OpenSSL by the OSS-Fuzz project.
4175 * Montgomery multiplication may produce incorrect results
4177 There is a carry propagating bug in the Broadwell-specific Montgomery
4178 multiplication procedure that handles input lengths divisible by, but
4179 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4180 and DH private keys are impossible. This is because the subroutine in
4181 question is not used in operations with the private key itself and an input
4182 of the attacker's direct choice. Otherwise the bug can manifest itself as
4183 transient authentication and key negotiation failures or reproducible
4184 erroneous outcome of public-key operations with specially crafted input.
4185 Among EC algorithms only Brainpool P-512 curves are affected and one
4186 presumably can attack ECDH key negotiation. Impact was not analyzed in
4187 detail, because pre-requisites for attack are considered unlikely. Namely
4188 multiple clients have to choose the curve in question and the server has to
4189 share the private key among them, neither of which is default behaviour.
4190 Even then only clients that chose the curve will be affected.
4192 This issue was publicly reported as transient failures and was not
4193 initially recognized as a security issue. Thanks to Richard Morgan for
4194 providing reproducible case.
4199 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4200 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4201 prevent issues where no progress is being made and the peer continually
4202 sends unrecognised record types, using up resources processing them.
4206 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4208 * Missing CRL sanity check
4210 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4211 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4212 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4214 This issue only affects the OpenSSL 1.0.2i
4219 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4221 * OCSP Status Request extension unbounded memory growth
4223 A malicious client can send an excessively large OCSP Status Request
4224 extension. If that client continually requests renegotiation, sending a
4225 large OCSP Status Request extension each time, then there will be unbounded
4226 memory growth on the server. This will eventually lead to a Denial Of
4227 Service attack through memory exhaustion. Servers with a default
4228 configuration are vulnerable even if they do not support OCSP. Builds using
4229 the "no-ocsp" build time option are not affected.
4231 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4236 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4239 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4245 * OOB write in MDC2_Update()
4247 An overflow can occur in MDC2_Update() either if called directly or
4248 through the EVP_DigestUpdate() function using MDC2. If an attacker
4249 is able to supply very large amounts of input data after a previous
4250 call to EVP_EncryptUpdate() with a partial block then a length check
4251 can overflow resulting in a heap corruption.
4253 The amount of data needed is comparable to SIZE_MAX which is impractical
4256 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4261 * Malformed SHA512 ticket DoS
4263 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4264 DoS attack where a malformed ticket will result in an OOB read which will
4267 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4268 a custom server callback and ticket lookup mechanism.
4270 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4275 * OOB write in BN_bn2dec()
4277 The function BN_bn2dec() does not check the return value of BN_div_word().
4278 This can cause an OOB write if an application uses this function with an
4279 overly large BIGNUM. This could be a problem if an overly large certificate
4280 or CRL is printed out from an untrusted source. TLS is not affected because
4281 record limits will reject an oversized certificate before it is parsed.
4283 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4288 * OOB read in TS_OBJ_print_bio()
4290 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4291 the total length the OID text representation would use and not the amount
4292 of data written. This will result in OOB reads when large OIDs are
4295 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4300 * Pointer arithmetic undefined behaviour
4302 Avoid some undefined pointer arithmetic
4304 A common idiom in the codebase is to check limits in the following manner:
4307 Where "p" points to some malloc'd data of SIZE bytes and
4310 "len" here could be from some externally supplied data (e.g. from a TLS
4313 The rules of C pointer arithmetic are such that "p + len" is only well
4314 defined where len <= SIZE. Therefore the above idiom is actually
4315 undefined behaviour.
4317 For example this could cause problems if some malloc implementation
4318 provides an address for "p" such that "p + len" actually overflows for
4319 values of len that are too big and therefore p + len < limit.
4321 This issue was reported to OpenSSL by Guido Vranken
4326 * Constant time flag not preserved in DSA signing
4328 Operations in the DSA signing algorithm should run in constant time in
4329 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4330 implementation means that a non-constant time codepath is followed for
4331 certain operations. This has been demonstrated through a cache-timing
4332 attack to be sufficient for an attacker to recover the private DSA key.
4334 This issue was reported by César Pereida (Aalto University), Billy Brumley
4335 (Tampere University of Technology), and Yuval Yarom (The University of
4336 Adelaide and NICTA).
4341 * DTLS buffered message DoS
4343 In a DTLS connection where handshake messages are delivered out-of-order
4344 those messages that OpenSSL is not yet ready to process will be buffered
4345 for later use. Under certain circumstances, a flaw in the logic means that
4346 those messages do not get removed from the buffer even though the handshake
4347 has been completed. An attacker could force up to approx. 15 messages to
4348 remain in the buffer when they are no longer required. These messages will
4349 be cleared when the DTLS connection is closed. The default maximum size for
4350 a message is 100k. Therefore the attacker could force an additional 1500k
4351 to be consumed per connection. By opening many simulataneous connections an
4352 attacker could cause a DoS attack through memory exhaustion.
4354 This issue was reported to OpenSSL by Quan Luo.
4359 * DTLS replay protection DoS
4361 A flaw in the DTLS replay attack protection mechanism means that records
4362 that arrive for future epochs update the replay protection "window" before
4363 the MAC for the record has been validated. This could be exploited by an
4364 attacker by sending a record for the next epoch (which does not have to
4365 decrypt or have a valid MAC), with a very large sequence number. This means
4366 that all subsequent legitimate packets are dropped causing a denial of
4367 service for a specific DTLS connection.
4369 This issue was reported to OpenSSL by the OCAP audit team.
4374 * Certificate message OOB reads
4376 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4377 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4378 theoretical DoS risk but this has not been observed in practice on common
4381 The messages affected are client certificate, client certificate request
4382 and server certificate. As a result the attack can only be performed
4383 against a client or a server which enables client authentication.
4385 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4390 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4392 * Prevent padding oracle in AES-NI CBC MAC check
4394 A MITM attacker can use a padding oracle attack to decrypt traffic
4395 when the connection uses an AES CBC cipher and the server support
4398 This issue was introduced as part of the fix for Lucky 13 padding
4399 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4400 constant time by making sure that always the same bytes are read and
4401 compared against either the MAC or padding bytes. But it no longer
4402 checked that there was enough data to have both the MAC and padding
4405 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4410 * Fix EVP_EncodeUpdate overflow
4412 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4413 Base64 encoding of binary data. If an attacker is able to supply very large
4414 amounts of input data then a length check can overflow resulting in a heap
4417 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4418 the `PEM_write_bio*` family of functions. These are mainly used within the
4419 OpenSSL command line applications, so any application which processes data
4420 from an untrusted source and outputs it as a PEM file should be considered
4421 vulnerable to this issue. User applications that call these APIs directly
4422 with large amounts of untrusted data may also be vulnerable.
4424 This issue was reported by Guido Vranken.
4429 * Fix EVP_EncryptUpdate overflow
4431 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4432 is able to supply very large amounts of input data after a previous call to
4433 EVP_EncryptUpdate() with a partial block then a length check can overflow
4434 resulting in a heap corruption. Following an analysis of all OpenSSL
4435 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4436 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4437 the first called function after an EVP_EncryptInit(), and therefore that
4438 specific call must be safe. The second form is where the length passed to
4439 EVP_EncryptUpdate() can be seen from the code to be some small value and
4440 therefore there is no possibility of an overflow. Since all instances are
4441 one of these two forms, it is believed that there can be no overflows in
4442 internal code due to this problem. It should be noted that
4443 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4444 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4445 of these calls have also been analysed too and it is believed there are no
4446 instances in internal usage where an overflow could occur.
4448 This issue was reported by Guido Vranken.
4453 * Prevent ASN.1 BIO excessive memory allocation
4455 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4456 a short invalid encoding can cause allocation of large amounts of memory
4457 potentially consuming excessive resources or exhausting memory.
4459 Any application parsing untrusted data through d2i BIO functions is
4460 affected. The memory based functions such as d2i_X509() are *not* affected.
4461 Since the memory based functions are used by the TLS library, TLS
4462 applications are not affected.
4464 This issue was reported by Brian Carpenter.
4471 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4472 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4473 in arbitrary stack data being returned in the buffer.
4475 This issue was reported by Guido Vranken.
4480 * Modify behavior of ALPN to invoke callback after SNI/servername
4481 callback, such that updates to the SSL_CTX affect ALPN.
4485 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4490 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4491 methods are enabled and ssl2 is disabled the methods return NULL.
4495 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4497 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4498 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4499 provide any "EXPORT" or "LOW" strength ciphers.
4503 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4504 is by default disabled at build-time. Builds that are not configured with
4505 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4506 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4507 will need to explicitly call either of:
4509 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4511 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4513 as appropriate. Even if either of those is used, or the application
4514 explicitly uses the version-specific SSLv2_method() or its client and
4515 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4516 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4517 ciphers, and SSLv2 56-bit DES are no longer available.
4522 * Fix a double-free in DSA code
4524 A double free bug was discovered when OpenSSL parses malformed DSA private
4525 keys and could lead to a DoS attack or memory corruption for applications
4526 that receive DSA private keys from untrusted sources. This scenario is
4529 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4535 * Disable SRP fake user seed to address a server memory leak.
4537 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4539 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4540 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4541 was changed to ignore the "fake user" SRP seed, even if the seed
4544 Users should use SRP_VBASE_get1_by_user instead. Note that in
4545 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4546 also that even though configuring the SRP seed attempts to hide
4547 invalid usernames by continuing the handshake with fake
4548 credentials, this behaviour is not constant time and no strong
4549 guarantees are made that the handshake is indistinguishable from
4550 that of a valid user.
4555 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4557 In the BN_hex2bn function the number of hex digits is calculated using an
4558 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4559 large values of `i` this can result in `bn_expand` not allocating any
4560 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4561 field as NULL leading to a subsequent NULL ptr deref. For very large values
4562 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4563 In this case memory is allocated to the internal BIGNUM data field, but it
4564 is insufficiently sized leading to heap corruption. A similar issue exists
4565 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4566 is ever called by user applications with very large untrusted hex/dec data.
4567 This is anticipated to be a rare occurrence.
4569 All OpenSSL internal usage of these functions use data that is not expected
4570 to be untrusted, e.g. config file data or application command line
4571 arguments. If user developed applications generate config file data based
4572 on untrusted data then it is possible that this could also lead to security
4573 consequences. This is also anticipated to be rare.
4575 This issue was reported to OpenSSL by Guido Vranken.
4580 * Fix memory issues in `BIO_*printf` functions
4582 The internal `fmtstr` function used in processing a "%s" format string in
4583 the `BIO_*printf` functions could overflow while calculating the length of a
4584 string and cause an OOB read when printing very long strings.
4586 Additionally the internal `doapr_outch` function can attempt to write to an
4587 OOB memory location (at an offset from the NULL pointer) in the event of a
4588 memory allocation failure. In 1.0.2 and below this could be caused where
4589 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4590 could be in processing a very long "%s" format string. Memory leaks can
4593 The first issue may mask the second issue dependent on compiler behaviour.
4594 These problems could enable attacks where large amounts of untrusted data
4595 is passed to the `BIO_*printf` functions. If applications use these functions
4596 in this way then they could be vulnerable. OpenSSL itself uses these
4597 functions when printing out human-readable dumps of ASN.1 data. Therefore
4598 applications that print this data could be vulnerable if the data is from
4599 untrusted sources. OpenSSL command line applications could also be
4600 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4601 as command line arguments.
4603 Libssl is not considered directly vulnerable. Additionally certificates etc
4604 received via remote connections via libssl are also unlikely to be able to
4605 trigger these issues because of message size limits enforced within libssl.
4607 This issue was reported to OpenSSL Guido Vranken.
4612 * Side channel attack on modular exponentiation
4614 A side-channel attack was found which makes use of cache-bank conflicts on
4615 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4616 of RSA keys. The ability to exploit this issue is limited as it relies on
4617 an attacker who has control of code in a thread running on the same
4618 hyper-threaded core as the victim thread which is performing decryptions.
4620 This issue was reported to OpenSSL by Yuval Yarom, The University of
4621 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4622 Nadia Heninger, University of Pennsylvania with more information at
4623 <http://cachebleed.info>.
4628 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4629 if no keysize is specified with default_bits. This fixes an
4630 omission in an earlier change that changed all RSA/DSA key generation
4631 apps to use 2048 bits by default.
4635 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4637 * DH small subgroups
4639 Historically OpenSSL only ever generated DH parameters based on "safe"
4640 primes. More recently (in version 1.0.2) support was provided for
4641 generating X9.42 style parameter files such as those required for RFC 5114
4642 support. The primes used in such files may not be "safe". Where an
4643 application is using DH configured with parameters based on primes that are
4644 not "safe" then an attacker could use this fact to find a peer's private
4645 DH exponent. This attack requires that the attacker complete multiple
4646 handshakes in which the peer uses the same private DH exponent. For example
4647 this could be used to discover a TLS server's private DH exponent if it's
4648 reusing the private DH exponent or it's using a static DH ciphersuite.
4650 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4651 TLS. It is not on by default. If the option is not set then the server
4652 reuses the same private DH exponent for the life of the server process and
4653 would be vulnerable to this attack. It is believed that many popular
4654 applications do set this option and would therefore not be at risk.
4656 The fix for this issue adds an additional check where a "q" parameter is
4657 available (as is the case in X9.42 based parameters). This detects the
4658 only known attack, and is the only possible defense for static DH
4659 ciphersuites. This could have some performance impact.
4661 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4662 default and cannot be disabled. This could have some performance impact.
4664 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4669 * SSLv2 doesn't block disabled ciphers
4671 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4672 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4673 been disabled, provided that the SSLv2 protocol was not also disabled via
4676 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4677 and Sebastian Schinzel.
4682 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4684 * BN_mod_exp may produce incorrect results on x86_64
4686 There is a carry propagating bug in the x86_64 Montgomery squaring
4687 procedure. No EC algorithms are affected. Analysis suggests that attacks
4688 against RSA and DSA as a result of this defect would be very difficult to
4689 perform and are not believed likely. Attacks against DH are considered just
4690 feasible (although very difficult) because most of the work necessary to
4691 deduce information about a private key may be performed offline. The amount
4692 of resources required for such an attack would be very significant and
4693 likely only accessible to a limited number of attackers. An attacker would
4694 additionally need online access to an unpatched system using the target
4695 private key in a scenario with persistent DH parameters and a private
4696 key that is shared between multiple clients. For example this can occur by
4697 default in OpenSSL DHE based SSL/TLS ciphersuites.
4699 This issue was reported to OpenSSL by Hanno Böck.
4704 * Certificate verify crash with missing PSS parameter
4706 The signature verification routines will crash with a NULL pointer
4707 dereference if presented with an ASN.1 signature using the RSA PSS
4708 algorithm and absent mask generation function parameter. Since these
4709 routines are used to verify certificate signature algorithms this can be
4710 used to crash any certificate verification operation and exploited in a
4711 DoS attack. Any application which performs certificate verification is
4712 vulnerable including OpenSSL clients and servers which enable client
4715 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4720 * X509_ATTRIBUTE memory leak
4722 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4723 memory. This structure is used by the PKCS#7 and CMS routines so any
4724 application which reads PKCS#7 or CMS data from untrusted sources is
4725 affected. SSL/TLS is not affected.
4727 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4733 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4734 This changes the decoding behaviour for some invalid messages,
4735 though the change is mostly in the more lenient direction, and
4736 legacy behaviour is preserved as much as possible.
4740 * In DSA_generate_parameters_ex, if the provided seed is too short,
4743 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4745 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4747 * Alternate chains certificate forgery
4749 During certificate verification, OpenSSL will attempt to find an
4750 alternative certificate chain if the first attempt to build such a chain
4751 fails. An error in the implementation of this logic can mean that an
4752 attacker could cause certain checks on untrusted certificates to be
4753 bypassed, such as the CA flag, enabling them to use a valid leaf
4754 certificate to act as a CA and "issue" an invalid certificate.
4756 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4761 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4763 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4764 incompatibility in the handling of HMAC. The previous ABI has now been
4769 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4771 * Malformed ECParameters causes infinite loop
4773 When processing an ECParameters structure OpenSSL enters an infinite loop
4774 if the curve specified is over a specially malformed binary polynomial
4777 This can be used to perform denial of service against any
4778 system which processes public keys, certificate requests or
4779 certificates. This includes TLS clients and TLS servers with
4780 client authentication enabled.
4782 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4787 * Exploitable out-of-bounds read in X509_cmp_time
4789 X509_cmp_time does not properly check the length of the ASN1_TIME
4790 string and can read a few bytes out of bounds. In addition,
4791 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4794 An attacker can use this to craft malformed certificates and CRLs of
4795 various sizes and potentially cause a segmentation fault, resulting in
4796 a DoS on applications that verify certificates or CRLs. TLS clients
4797 that verify CRLs are affected. TLS clients and servers with client
4798 authentication enabled may be affected if they use custom verification
4801 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4802 independently by Hanno Böck.
4807 * PKCS7 crash with missing EnvelopedContent
4809 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4810 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4811 with missing content and trigger a NULL pointer dereference on parsing.
4813 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4814 structures from untrusted sources are affected. OpenSSL clients and
4815 servers are not affected.
4817 This issue was reported to OpenSSL by Michal Zalewski (Google).
4822 * CMS verify infinite loop with unknown hash function
4824 When verifying a signedData message the CMS code can enter an infinite loop
4825 if presented with an unknown hash function OID. This can be used to perform
4826 denial of service against any system which verifies signedData messages using
4828 This issue was reported to OpenSSL by Johannes Bauer.
4833 * Race condition handling NewSessionTicket
4835 If a NewSessionTicket is received by a multi-threaded client when attempting to
4836 reuse a previous ticket then a race condition can occur potentially leading to
4837 a double free of the ticket data.
4842 * Only support 256-bit or stronger elliptic curves with the
4843 'ecdh_auto' setting (server) or by default (client). Of supported
4844 curves, prefer P-256 (both).
4848 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4850 * ClientHello sigalgs DoS fix
4852 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4853 invalid signature algorithms extension a NULL pointer dereference will
4854 occur. This can be exploited in a DoS attack against the server.
4856 This issue was was reported to OpenSSL by David Ramos of Stanford
4860 *Stephen Henson and Matt Caswell*
4862 * Multiblock corrupted pointer fix
4864 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4865 feature only applies on 64 bit x86 architecture platforms that support AES
4866 NI instructions. A defect in the implementation of "multiblock" can cause
4867 OpenSSL's internal write buffer to become incorrectly set to NULL when
4868 using non-blocking IO. Typically, when the user application is using a
4869 socket BIO for writing, this will only result in a failed connection.
4870 However if some other BIO is used then it is likely that a segmentation
4871 fault will be triggered, thus enabling a potential DoS attack.
4873 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4878 * Segmentation fault in DTLSv1_listen fix
4880 The DTLSv1_listen function is intended to be stateless and processes the
4881 initial ClientHello from many peers. It is common for user code to loop
4882 over the call to DTLSv1_listen until a valid ClientHello is received with
4883 an associated cookie. A defect in the implementation of DTLSv1_listen means
4884 that state is preserved in the SSL object from one invocation to the next
4885 that can lead to a segmentation fault. Errors processing the initial
4886 ClientHello can trigger this scenario. An example of such an error could be
4887 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4890 This issue was reported to OpenSSL by Per Allansson.
4895 * Segmentation fault in ASN1_TYPE_cmp fix
4897 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4898 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4899 certificate signature algorithm consistency this can be used to crash any
4900 certificate verification operation and exploited in a DoS attack. Any
4901 application which performs certificate verification is vulnerable including
4902 OpenSSL clients and servers which enable client authentication.
4907 * Segmentation fault for invalid PSS parameters fix
4909 The signature verification routines will crash with a NULL pointer
4910 dereference if presented with an ASN.1 signature using the RSA PSS
4911 algorithm and invalid parameters. Since these routines are used to verify
4912 certificate signature algorithms this can be used to crash any
4913 certificate verification operation and exploited in a DoS attack. Any
4914 application which performs certificate verification is vulnerable including
4915 OpenSSL clients and servers which enable client authentication.
4917 This issue was was reported to OpenSSL by Brian Carpenter.
4922 * ASN.1 structure reuse memory corruption fix
4924 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4925 memory corruption via an invalid write. Such reuse is and has been
4926 strongly discouraged and is believed to be rare.
4928 Applications that parse structures containing CHOICE or ANY DEFINED BY
4929 components may be affected. Certificate parsing (d2i_X509 and related
4930 functions) are however not affected. OpenSSL clients and servers are
4936 * PKCS7 NULL pointer dereferences fix
4938 The PKCS#7 parsing code does not handle missing outer ContentInfo
4939 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4940 missing content and trigger a NULL pointer dereference on parsing.
4942 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4943 otherwise parse PKCS#7 structures from untrusted sources are
4944 affected. OpenSSL clients and servers are not affected.
4946 This issue was reported to OpenSSL by Michal Zalewski (Google).
4951 * DoS via reachable assert in SSLv2 servers fix
4953 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4954 servers that both support SSLv2 and enable export cipher suites by sending
4955 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4957 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4958 (OpenSSL development team).
4963 * Empty CKE with client auth and DHE fix
4965 If client auth is used then a server can seg fault in the event of a DHE
4966 ciphersuite being selected and a zero length ClientKeyExchange message
4967 being sent by the client. This could be exploited in a DoS attack.
4972 * Handshake with unseeded PRNG fix
4974 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4975 with an unseeded PRNG. The conditions are:
4976 - The client is on a platform where the PRNG has not been seeded
4977 automatically, and the user has not seeded manually
4978 - A protocol specific client method version has been used (i.e. not
4979 SSL_client_methodv23)
4980 - A ciphersuite is used that does not require additional random data from
4981 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4983 If the handshake succeeds then the client random that has been used will
4984 have been generated from a PRNG with insufficient entropy and therefore the
4985 output may be predictable.
4987 For example using the following command with an unseeded openssl will
4988 succeed on an unpatched platform:
4990 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4995 * Use After Free following d2i_ECPrivatekey error fix
4997 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4998 could cause a use after free condition. This, in turn, could cause a double
4999 free in several private key parsing functions (such as d2i_PrivateKey
5000 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5001 for applications that receive EC private keys from untrusted
5002 sources. This scenario is considered rare.
5004 This issue was discovered by the BoringSSL project and fixed in their
5010 * X509_to_X509_REQ NULL pointer deref fix
5012 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5013 the certificate key is invalid. This function is rarely used in practice.
5015 This issue was discovered by Brian Carpenter.
5020 * Removed the export ciphers from the DEFAULT ciphers
5024 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5026 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5027 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5028 So far those who have to target multiple platforms would compromise
5029 and argue that binary targeting say ARMv5 would still execute on
5030 ARMv8. "Universal" build resolves this compromise by providing
5031 near-optimal performance even on newer platforms.
5035 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5036 (other platforms pending).
5038 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5040 * Add support for the SignedCertificateTimestampList certificate and
5041 OCSP response extensions from RFC6962.
5045 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5046 for corner cases. (Certain input points at infinity could lead to
5047 bogus results, with non-infinity inputs mapped to infinity too.)
5051 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5052 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5053 common cases are optimized and there still is room for further
5054 improvements. Vector Permutation AES for Altivec is also added.
5058 * Add support for little-endian ppc64 Linux target.
5060 *Marcelo Cerri (IBM)*
5062 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5063 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5064 are optimized and there still is room for further improvements.
5065 Both 32- and 64-bit modes are supported.
5067 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5069 * Improved ARMv7 NEON support.
5073 * Support for SPARC Architecture 2011 crypto extensions, first
5074 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5075 SHA256/512, MD5, GHASH and modular exponentiation.
5077 *Andy Polyakov, David Miller*
5079 * Accelerated modular exponentiation for Intel processors, a.k.a.
5082 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5084 * Support for new and upcoming Intel processors, including AVX2,
5085 BMI and SHA ISA extensions. This includes additional "stitched"
5086 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5089 This work was sponsored by Intel Corp.
5093 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5094 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5095 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5099 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5100 this fixes a limitation in previous versions of OpenSSL.
5104 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5105 MGF1 digest and OAEP label.
5109 * Add EVP support for key wrapping algorithms, to avoid problems with
5110 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5111 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5112 algorithms and include tests cases.
5116 * Add functions to allocate and set the fields of an ECDSA_METHOD
5119 *Douglas E. Engert, Steve Henson*
5121 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5122 difference in days and seconds between two tm or ASN1_TIME structures.
5126 * Add -rev test option to s_server to just reverse order of characters
5127 received by client and send back to server. Also prints an abbreviated
5128 summary of the connection parameters.
5132 * New option -brief for s_client and s_server to print out a brief summary
5133 of connection parameters.
5137 * Add callbacks for arbitrary TLS extensions.
5139 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5141 * New option -crl_download in several openssl utilities to download CRLs
5142 from CRLDP extension in certificates.
5146 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5150 * New function X509_CRL_diff to generate a delta CRL from the difference
5151 of two full CRLs. Add support to "crl" utility.
5155 * New functions to set lookup_crls function and to retrieve
5156 X509_STORE from X509_STORE_CTX.
5160 * Print out deprecated issuer and subject unique ID fields in
5165 * Extend OCSP I/O functions so they can be used for simple general purpose
5166 HTTP as well as OCSP. New wrapper function which can be used to download
5167 CRLs using the OCSP API.
5171 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5175 * `SSL_CONF*` functions. These provide a common framework for application
5176 configuration using configuration files or command lines.
5180 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5181 message callback and prints the results. Needs compile time option
5182 "enable-ssl-trace". New options to s_client and s_server to enable
5187 * New ctrl and macro to retrieve supported points extensions.
5188 Print out extension in s_server and s_client.
5192 * New functions to retrieve certificate signature and signature
5197 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5202 * New Suite B modes for TLS code. These use and enforce the requirements
5203 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5204 only use Suite B curves. The Suite B modes can be set by using the
5205 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5209 * New chain verification flags for Suite B levels of security. Check
5210 algorithms are acceptable when flags are set in X509_verify_cert.
5214 * Make tls1_check_chain return a set of flags indicating checks passed
5215 by a certificate chain. Add additional tests to handle client
5216 certificates: checks for matching certificate type and issuer name
5221 * If an attempt is made to use a signature algorithm not in the peer
5222 preference list abort the handshake. If client has no suitable
5223 signature algorithms in response to a certificate request do not
5224 use the certificate.
5228 * If server EC tmp key is not in client preference list abort handshake.
5232 * Add support for certificate stores in CERT structure. This makes it
5233 possible to have different stores per SSL structure or one store in
5234 the parent SSL_CTX. Include distinct stores for certificate chain
5235 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5236 to build and store a certificate chain in CERT structure: returning
5237 an error if the chain cannot be built: this will allow applications
5238 to test if a chain is correctly configured.
5240 Note: if the CERT based stores are not set then the parent SSL_CTX
5241 store is used to retain compatibility with existing behaviour.
5245 * New function ssl_set_client_disabled to set a ciphersuite disabled
5246 mask based on the current session, check mask when sending client
5247 hello and checking the requested ciphersuite.
5251 * New ctrls to retrieve and set certificate types in a certificate
5252 request message. Print out received values in s_client. If certificate
5253 types is not set with custom values set sensible values based on
5254 supported signature algorithms.
5258 * Support for distinct client and server supported signature algorithms.
5262 * Add certificate callback. If set this is called whenever a certificate
5263 is required by client or server. An application can decide which
5264 certificate chain to present based on arbitrary criteria: for example
5265 supported signature algorithms. Add very simple example to s_server.
5266 This fixes many of the problems and restrictions of the existing client
5267 certificate callback: for example you can now clear an existing
5268 certificate and specify the whole chain.
5272 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5273 the certificate can be used for (if anything). Set valid_flags field
5274 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5275 to have similar checks in it.
5277 Add new "cert_flags" field to CERT structure and include a "strict mode".
5278 This enforces some TLS certificate requirements (such as only permitting
5279 certificate signature algorithms contained in the supported algorithms
5280 extension) which some implementations ignore: this option should be used
5281 with caution as it could cause interoperability issues.
5285 * Update and tidy signature algorithm extension processing. Work out
5286 shared signature algorithms based on preferences and peer algorithms
5287 and print them out in s_client and s_server. Abort handshake if no
5288 shared signature algorithms.
5292 * Add new functions to allow customised supported signature algorithms
5293 for SSL and SSL_CTX structures. Add options to s_client and s_server
5298 * New function SSL_certs_clear() to delete all references to certificates
5299 from an SSL structure. Before this once a certificate had been added
5300 it couldn't be removed.
5304 * Integrate hostname, email address and IP address checking with certificate
5305 verification. New verify options supporting checking in openssl utility.
5309 * Fixes and wildcard matching support to hostname and email checking
5310 functions. Add manual page.
5312 *Florian Weimer (Red Hat Product Security Team)*
5314 * New functions to check a hostname email or IP address against a
5315 certificate. Add options x509 utility to print results of checks against
5320 * Fix OCSP checking.
5322 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5324 * Initial experimental support for explicitly trusted non-root CAs.
5325 OpenSSL still tries to build a complete chain to a root but if an
5326 intermediate CA has a trust setting included that is used. The first
5327 setting is used: whether to trust (e.g., -addtrust option to the x509
5332 * Add -trusted_first option which attempts to find certificates in the
5333 trusted store even if an untrusted chain is also supplied.
5337 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5338 platform support for Linux and Android.
5342 * Support for linux-x32, ILP32 environment in x86_64 framework.
5346 * Experimental multi-implementation support for FIPS capable OpenSSL.
5347 When in FIPS mode the approved implementations are used as normal,
5348 when not in FIPS mode the internal unapproved versions are used instead.
5349 This means that the FIPS capable OpenSSL isn't forced to use the
5350 (often lower performance) FIPS implementations outside FIPS mode.
5354 * Transparently support X9.42 DH parameters when calling
5355 PEM_read_bio_DHparameters. This means existing applications can handle
5356 the new parameter format automatically.
5360 * Initial experimental support for X9.42 DH parameter format: mainly
5361 to support use of 'q' parameter for RFC5114 parameters.
5365 * Add DH parameters from RFC5114 including test data to dhtest.
5369 * Support for automatic EC temporary key parameter selection. If enabled
5370 the most preferred EC parameters are automatically used instead of
5371 hardcoded fixed parameters. Now a server just has to call:
5372 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5373 support ECDH and use the most appropriate parameters.
5377 * Enhance and tidy EC curve and point format TLS extension code. Use
5378 static structures instead of allocation if default values are used.
5379 New ctrls to set curves we wish to support and to retrieve shared curves.
5380 Print out shared curves in s_server. New options to s_server and s_client
5381 to set list of supported curves.
5385 * New ctrls to retrieve supported signature algorithms and
5386 supported curve values as an array of NIDs. Extend openssl utility
5387 to print out received values.
5391 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5392 between NIDs and the more common NIST names such as "P-256". Enhance
5393 ecparam utility and ECC method to recognise the NIST names for curves.
5397 * Enhance SSL/TLS certificate chain handling to support different
5398 chains for each certificate instead of one chain in the parent SSL_CTX.
5402 * Support for fixed DH ciphersuite client authentication: where both
5403 server and client use DH certificates with common parameters.
5407 * Support for fixed DH ciphersuites: those requiring DH server
5412 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5414 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5415 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5416 X509_CINF_get_signature were reverted post internal team review.
5421 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5423 * OCSP Status Request extension unbounded memory growth
5425 A malicious client can send an excessively large OCSP Status Request
5426 extension. If that client continually requests renegotiation, sending a
5427 large OCSP Status Request extension each time, then there will be unbounded
5428 memory growth on the server. This will eventually lead to a Denial Of
5429 Service attack through memory exhaustion. Servers with a default
5430 configuration are vulnerable even if they do not support OCSP. Builds using
5431 the "no-ocsp" build time option are not affected.
5433 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5438 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5441 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5447 * OOB write in MDC2_Update()
5449 An overflow can occur in MDC2_Update() either if called directly or
5450 through the EVP_DigestUpdate() function using MDC2. If an attacker
5451 is able to supply very large amounts of input data after a previous
5452 call to EVP_EncryptUpdate() with a partial block then a length check
5453 can overflow resulting in a heap corruption.
5455 The amount of data needed is comparable to SIZE_MAX which is impractical
5458 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5463 * Malformed SHA512 ticket DoS
5465 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5466 DoS attack where a malformed ticket will result in an OOB read which will
5469 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5470 a custom server callback and ticket lookup mechanism.
5472 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5477 * OOB write in BN_bn2dec()
5479 The function BN_bn2dec() does not check the return value of BN_div_word().
5480 This can cause an OOB write if an application uses this function with an
5481 overly large BIGNUM. This could be a problem if an overly large certificate
5482 or CRL is printed out from an untrusted source. TLS is not affected because
5483 record limits will reject an oversized certificate before it is parsed.
5485 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5490 * OOB read in TS_OBJ_print_bio()
5492 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5493 the total length the OID text representation would use and not the amount
5494 of data written. This will result in OOB reads when large OIDs are
5497 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5502 * Pointer arithmetic undefined behaviour
5504 Avoid some undefined pointer arithmetic
5506 A common idiom in the codebase is to check limits in the following manner:
5509 Where "p" points to some malloc'd data of SIZE bytes and
5512 "len" here could be from some externally supplied data (e.g. from a TLS
5515 The rules of C pointer arithmetic are such that "p + len" is only well
5516 defined where len <= SIZE. Therefore the above idiom is actually
5517 undefined behaviour.
5519 For example this could cause problems if some malloc implementation
5520 provides an address for "p" such that "p + len" actually overflows for
5521 values of len that are too big and therefore p + len < limit.
5523 This issue was reported to OpenSSL by Guido Vranken
5528 * Constant time flag not preserved in DSA signing
5530 Operations in the DSA signing algorithm should run in constant time in
5531 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5532 implementation means that a non-constant time codepath is followed for
5533 certain operations. This has been demonstrated through a cache-timing
5534 attack to be sufficient for an attacker to recover the private DSA key.
5536 This issue was reported by César Pereida (Aalto University), Billy Brumley
5537 (Tampere University of Technology), and Yuval Yarom (The University of
5538 Adelaide and NICTA).
5543 * DTLS buffered message DoS
5545 In a DTLS connection where handshake messages are delivered out-of-order
5546 those messages that OpenSSL is not yet ready to process will be buffered
5547 for later use. Under certain circumstances, a flaw in the logic means that
5548 those messages do not get removed from the buffer even though the handshake
5549 has been completed. An attacker could force up to approx. 15 messages to
5550 remain in the buffer when they are no longer required. These messages will
5551 be cleared when the DTLS connection is closed. The default maximum size for
5552 a message is 100k. Therefore the attacker could force an additional 1500k
5553 to be consumed per connection. By opening many simulataneous connections an
5554 attacker could cause a DoS attack through memory exhaustion.
5556 This issue was reported to OpenSSL by Quan Luo.
5561 * DTLS replay protection DoS
5563 A flaw in the DTLS replay attack protection mechanism means that records
5564 that arrive for future epochs update the replay protection "window" before
5565 the MAC for the record has been validated. This could be exploited by an
5566 attacker by sending a record for the next epoch (which does not have to
5567 decrypt or have a valid MAC), with a very large sequence number. This means
5568 that all subsequent legitimate packets are dropped causing a denial of
5569 service for a specific DTLS connection.
5571 This issue was reported to OpenSSL by the OCAP audit team.
5576 * Certificate message OOB reads
5578 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5579 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5580 theoretical DoS risk but this has not been observed in practice on common
5583 The messages affected are client certificate, client certificate request
5584 and server certificate. As a result the attack can only be performed
5585 against a client or a server which enables client authentication.
5587 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5592 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5594 * Prevent padding oracle in AES-NI CBC MAC check
5596 A MITM attacker can use a padding oracle attack to decrypt traffic
5597 when the connection uses an AES CBC cipher and the server support
5600 This issue was introduced as part of the fix for Lucky 13 padding
5601 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5602 constant time by making sure that always the same bytes are read and
5603 compared against either the MAC or padding bytes. But it no longer
5604 checked that there was enough data to have both the MAC and padding
5607 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5612 * Fix EVP_EncodeUpdate overflow
5614 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5615 Base64 encoding of binary data. If an attacker is able to supply very large
5616 amounts of input data then a length check can overflow resulting in a heap
5619 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5620 the `PEM_write_bio*` family of functions. These are mainly used within the
5621 OpenSSL command line applications, so any application which processes data
5622 from an untrusted source and outputs it as a PEM file should be considered
5623 vulnerable to this issue. User applications that call these APIs directly
5624 with large amounts of untrusted data may also be vulnerable.
5626 This issue was reported by Guido Vranken.
5631 * Fix EVP_EncryptUpdate overflow
5633 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5634 is able to supply very large amounts of input data after a previous call to
5635 EVP_EncryptUpdate() with a partial block then a length check can overflow
5636 resulting in a heap corruption. Following an analysis of all OpenSSL
5637 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5638 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5639 the first called function after an EVP_EncryptInit(), and therefore that
5640 specific call must be safe. The second form is where the length passed to
5641 EVP_EncryptUpdate() can be seen from the code to be some small value and
5642 therefore there is no possibility of an overflow. Since all instances are
5643 one of these two forms, it is believed that there can be no overflows in
5644 internal code due to this problem. It should be noted that
5645 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5646 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5647 of these calls have also been analysed too and it is believed there are no
5648 instances in internal usage where an overflow could occur.
5650 This issue was reported by Guido Vranken.
5655 * Prevent ASN.1 BIO excessive memory allocation
5657 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5658 a short invalid encoding can casuse allocation of large amounts of memory
5659 potentially consuming excessive resources or exhausting memory.
5661 Any application parsing untrusted data through d2i BIO functions is
5662 affected. The memory based functions such as d2i_X509() are *not* affected.
5663 Since the memory based functions are used by the TLS library, TLS
5664 applications are not affected.
5666 This issue was reported by Brian Carpenter.
5673 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5674 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5675 in arbitrary stack data being returned in the buffer.
5677 This issue was reported by Guido Vranken.
5682 * Modify behavior of ALPN to invoke callback after SNI/servername
5683 callback, such that updates to the SSL_CTX affect ALPN.
5687 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5692 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5693 methods are enabled and ssl2 is disabled the methods return NULL.
5697 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5699 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5700 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5701 provide any "EXPORT" or "LOW" strength ciphers.
5705 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5706 is by default disabled at build-time. Builds that are not configured with
5707 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5708 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5709 will need to explicitly call either of:
5711 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5713 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5715 as appropriate. Even if either of those is used, or the application
5716 explicitly uses the version-specific SSLv2_method() or its client and
5717 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5718 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5719 ciphers, and SSLv2 56-bit DES are no longer available.
5724 * Fix a double-free in DSA code
5726 A double free bug was discovered when OpenSSL parses malformed DSA private
5727 keys and could lead to a DoS attack or memory corruption for applications
5728 that receive DSA private keys from untrusted sources. This scenario is
5731 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5737 * Disable SRP fake user seed to address a server memory leak.
5739 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5741 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5742 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5743 was changed to ignore the "fake user" SRP seed, even if the seed
5746 Users should use SRP_VBASE_get1_by_user instead. Note that in
5747 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5748 also that even though configuring the SRP seed attempts to hide
5749 invalid usernames by continuing the handshake with fake
5750 credentials, this behaviour is not constant time and no strong
5751 guarantees are made that the handshake is indistinguishable from
5752 that of a valid user.
5757 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5759 In the BN_hex2bn function the number of hex digits is calculated using an
5760 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5761 large values of `i` this can result in `bn_expand` not allocating any
5762 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5763 field as NULL leading to a subsequent NULL ptr deref. For very large values
5764 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5765 In this case memory is allocated to the internal BIGNUM data field, but it
5766 is insufficiently sized leading to heap corruption. A similar issue exists
5767 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5768 is ever called by user applications with very large untrusted hex/dec data.
5769 This is anticipated to be a rare occurrence.
5771 All OpenSSL internal usage of these functions use data that is not expected
5772 to be untrusted, e.g. config file data or application command line
5773 arguments. If user developed applications generate config file data based
5774 on untrusted data then it is possible that this could also lead to security
5775 consequences. This is also anticipated to be rare.
5777 This issue was reported to OpenSSL by Guido Vranken.
5782 * Fix memory issues in `BIO_*printf` functions
5784 The internal `fmtstr` function used in processing a "%s" format string in
5785 the `BIO_*printf` functions could overflow while calculating the length of a
5786 string and cause an OOB read when printing very long strings.
5788 Additionally the internal `doapr_outch` function can attempt to write to an
5789 OOB memory location (at an offset from the NULL pointer) in the event of a
5790 memory allocation failure. In 1.0.2 and below this could be caused where
5791 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5792 could be in processing a very long "%s" format string. Memory leaks can
5795 The first issue may mask the second issue dependent on compiler behaviour.
5796 These problems could enable attacks where large amounts of untrusted data
5797 is passed to the `BIO_*printf` functions. If applications use these functions
5798 in this way then they could be vulnerable. OpenSSL itself uses these
5799 functions when printing out human-readable dumps of ASN.1 data. Therefore
5800 applications that print this data could be vulnerable if the data is from
5801 untrusted sources. OpenSSL command line applications could also be
5802 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5803 as command line arguments.
5805 Libssl is not considered directly vulnerable. Additionally certificates etc
5806 received via remote connections via libssl are also unlikely to be able to
5807 trigger these issues because of message size limits enforced within libssl.
5809 This issue was reported to OpenSSL Guido Vranken.
5814 * Side channel attack on modular exponentiation
5816 A side-channel attack was found which makes use of cache-bank conflicts on
5817 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5818 of RSA keys. The ability to exploit this issue is limited as it relies on
5819 an attacker who has control of code in a thread running on the same
5820 hyper-threaded core as the victim thread which is performing decryptions.
5822 This issue was reported to OpenSSL by Yuval Yarom, The University of
5823 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5824 Nadia Heninger, University of Pennsylvania with more information at
5825 <http://cachebleed.info>.
5830 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5831 if no keysize is specified with default_bits. This fixes an
5832 omission in an earlier change that changed all RSA/DSA key generation
5833 apps to use 2048 bits by default.
5837 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5839 * Protection for DH small subgroup attacks
5841 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5842 switched on by default and cannot be disabled. This could have some
5847 * SSLv2 doesn't block disabled ciphers
5849 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5850 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5851 been disabled, provided that the SSLv2 protocol was not also disabled via
5854 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5855 and Sebastian Schinzel.
5860 * Reject DH handshakes with parameters shorter than 1024 bits.
5864 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5866 * Certificate verify crash with missing PSS parameter
5868 The signature verification routines will crash with a NULL pointer
5869 dereference if presented with an ASN.1 signature using the RSA PSS
5870 algorithm and absent mask generation function parameter. Since these
5871 routines are used to verify certificate signature algorithms this can be
5872 used to crash any certificate verification operation and exploited in a
5873 DoS attack. Any application which performs certificate verification is
5874 vulnerable including OpenSSL clients and servers which enable client
5877 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5882 * X509_ATTRIBUTE memory leak
5884 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5885 memory. This structure is used by the PKCS#7 and CMS routines so any
5886 application which reads PKCS#7 or CMS data from untrusted sources is
5887 affected. SSL/TLS is not affected.
5889 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5895 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5896 This changes the decoding behaviour for some invalid messages,
5897 though the change is mostly in the more lenient direction, and
5898 legacy behaviour is preserved as much as possible.
5902 * In DSA_generate_parameters_ex, if the provided seed is too short,
5903 use a random seed, as already documented.
5905 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5907 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5909 * Alternate chains certificate forgery
5911 During certificate verfification, OpenSSL will attempt to find an
5912 alternative certificate chain if the first attempt to build such a chain
5913 fails. An error in the implementation of this logic can mean that an
5914 attacker could cause certain checks on untrusted certificates to be
5915 bypassed, such as the CA flag, enabling them to use a valid leaf
5916 certificate to act as a CA and "issue" an invalid certificate.
5918 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5924 * Race condition handling PSK identify hint
5926 If PSK identity hints are received by a multi-threaded client then
5927 the values are wrongly updated in the parent SSL_CTX structure. This can
5928 result in a race condition potentially leading to a double free of the
5934 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5936 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5937 incompatibility in the handling of HMAC. The previous ABI has now been
5940 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5942 * Malformed ECParameters causes infinite loop
5944 When processing an ECParameters structure OpenSSL enters an infinite loop
5945 if the curve specified is over a specially malformed binary polynomial
5948 This can be used to perform denial of service against any
5949 system which processes public keys, certificate requests or
5950 certificates. This includes TLS clients and TLS servers with
5951 client authentication enabled.
5953 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5958 * Exploitable out-of-bounds read in X509_cmp_time
5960 X509_cmp_time does not properly check the length of the ASN1_TIME
5961 string and can read a few bytes out of bounds. In addition,
5962 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5965 An attacker can use this to craft malformed certificates and CRLs of
5966 various sizes and potentially cause a segmentation fault, resulting in
5967 a DoS on applications that verify certificates or CRLs. TLS clients
5968 that verify CRLs are affected. TLS clients and servers with client
5969 authentication enabled may be affected if they use custom verification
5972 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5973 independently by Hanno Böck.
5978 * PKCS7 crash with missing EnvelopedContent
5980 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5981 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5982 with missing content and trigger a NULL pointer dereference on parsing.
5984 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5985 structures from untrusted sources are affected. OpenSSL clients and
5986 servers are not affected.
5988 This issue was reported to OpenSSL by Michal Zalewski (Google).
5993 * CMS verify infinite loop with unknown hash function
5995 When verifying a signedData message the CMS code can enter an infinite loop
5996 if presented with an unknown hash function OID. This can be used to perform
5997 denial of service against any system which verifies signedData messages using
5999 This issue was reported to OpenSSL by Johannes Bauer.
6004 * Race condition handling NewSessionTicket
6006 If a NewSessionTicket is received by a multi-threaded client when attempting to
6007 reuse a previous ticket then a race condition can occur potentially leading to
6008 a double free of the ticket data.
6013 * Reject DH handshakes with parameters shorter than 768 bits.
6015 *Kurt Roeckx and Emilia Kasper*
6017 * dhparam: generate 2048-bit parameters by default.
6019 *Kurt Roeckx and Emilia Kasper*
6021 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6023 * Segmentation fault in ASN1_TYPE_cmp fix
6025 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6026 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6027 certificate signature algorithm consistency this can be used to crash any
6028 certificate verification operation and exploited in a DoS attack. Any
6029 application which performs certificate verification is vulnerable including
6030 OpenSSL clients and servers which enable client authentication.
6035 * ASN.1 structure reuse memory corruption fix
6037 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6038 memory corruption via an invalid write. Such reuse is and has been
6039 strongly discouraged and is believed to be rare.
6041 Applications that parse structures containing CHOICE or ANY DEFINED BY
6042 components may be affected. Certificate parsing (d2i_X509 and related
6043 functions) are however not affected. OpenSSL clients and servers are
6049 * PKCS7 NULL pointer dereferences fix
6051 The PKCS#7 parsing code does not handle missing outer ContentInfo
6052 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6053 missing content and trigger a NULL pointer dereference on parsing.
6055 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6056 otherwise parse PKCS#7 structures from untrusted sources are
6057 affected. OpenSSL clients and servers are not affected.
6059 This issue was reported to OpenSSL by Michal Zalewski (Google).
6064 * DoS via reachable assert in SSLv2 servers fix
6066 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6067 servers that both support SSLv2 and enable export cipher suites by sending
6068 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6070 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6071 (OpenSSL development team).
6076 * Use After Free following d2i_ECPrivatekey error fix
6078 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6079 could cause a use after free condition. This, in turn, could cause a double
6080 free in several private key parsing functions (such as d2i_PrivateKey
6081 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6082 for applications that receive EC private keys from untrusted
6083 sources. This scenario is considered rare.
6085 This issue was discovered by the BoringSSL project and fixed in their
6091 * X509_to_X509_REQ NULL pointer deref fix
6093 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6094 the certificate key is invalid. This function is rarely used in practice.
6096 This issue was discovered by Brian Carpenter.
6101 * Removed the export ciphers from the DEFAULT ciphers
6105 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6107 * Build fixes for the Windows and OpenVMS platforms
6109 *Matt Caswell and Richard Levitte*
6111 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6113 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6114 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6115 dereference. This could lead to a Denial Of Service attack. Thanks to
6116 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6121 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6122 dtls1_buffer_record function under certain conditions. In particular this
6123 could occur if an attacker sent repeated DTLS records with the same
6124 sequence number but for the next epoch. The memory leak could be exploited
6125 by an attacker in a Denial of Service attack through memory exhaustion.
6126 Thanks to Chris Mueller for reporting this issue.
6131 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6132 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6133 method would be set to NULL which could later result in a NULL pointer
6134 dereference. Thanks to Frank Schmirler for reporting this issue.
6139 * Abort handshake if server key exchange message is omitted for ephemeral
6142 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6143 reporting this issue.
6148 * Remove non-export ephemeral RSA code on client and server. This code
6149 violated the TLS standard by allowing the use of temporary RSA keys in
6150 non-export ciphersuites and could be used by a server to effectively
6151 downgrade the RSA key length used to a value smaller than the server
6152 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6153 INRIA or reporting this issue.
6158 * Fixed issue where DH client certificates are accepted without verification.
6159 An OpenSSL server will accept a DH certificate for client authentication
6160 without the certificate verify message. This effectively allows a client to
6161 authenticate without the use of a private key. This only affects servers
6162 which trust a client certificate authority which issues certificates
6163 containing DH keys: these are extremely rare and hardly ever encountered.
6164 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6170 * Ensure that the session ID context of an SSL is updated when its
6171 SSL_CTX is updated via SSL_set_SSL_CTX.
6173 The session ID context is typically set from the parent SSL_CTX,
6174 and can vary with the CTX.
6178 * Fix various certificate fingerprint issues.
6180 By using non-DER or invalid encodings outside the signed portion of a
6181 certificate the fingerprint can be changed without breaking the signature.
6182 Although no details of the signed portion of the certificate can be changed
6183 this can cause problems with some applications: e.g. those using the
6184 certificate fingerprint for blacklists.
6186 1. Reject signatures with non zero unused bits.
6188 If the BIT STRING containing the signature has non zero unused bits reject
6189 the signature. All current signature algorithms require zero unused bits.
6191 2. Check certificate algorithm consistency.
6193 Check the AlgorithmIdentifier inside TBS matches the one in the
6194 certificate signature. NB: this will result in signature failure
6195 errors for some broken certificates.
6197 Thanks to Konrad Kraszewski from Google for reporting this issue.
6199 3. Check DSA/ECDSA signatures use DER.
6201 Re-encode DSA/ECDSA signatures and compare with the original received
6202 signature. Return an error if there is a mismatch.
6204 This will reject various cases including garbage after signature
6205 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6206 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6207 (negative or with leading zeroes).
6209 Further analysis was conducted and fixes were developed by Stephen Henson
6210 of the OpenSSL core team.
6216 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6217 results on some platforms, including x86_64. This bug occurs at random
6218 with a very low probability, and is not known to be exploitable in any
6219 way, though its exact impact is difficult to determine. Thanks to Pieter
6220 Wuille (Blockstream) who reported this issue and also suggested an initial
6221 fix. Further analysis was conducted by the OpenSSL development team and
6222 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6223 the OpenSSL core team.
6228 * Do not resume sessions on the server if the negotiated protocol
6229 version does not match the session's version. Resuming with a different
6230 version, while not strictly forbidden by the RFC, is of questionable
6231 sanity and breaks all known clients.
6233 *David Benjamin, Emilia Käsper*
6235 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6236 early CCS messages during renegotiation. (Note that because
6237 renegotiation is encrypted, this early CCS was not exploitable.)
6241 * Tighten client-side session ticket handling during renegotiation:
6242 ensure that the client only accepts a session ticket if the server sends
6243 the extension anew in the ServerHello. Previously, a TLS client would
6244 reuse the old extension state and thus accept a session ticket if one was
6245 announced in the initial ServerHello.
6247 Similarly, ensure that the client requires a session ticket if one
6248 was advertised in the ServerHello. Previously, a TLS client would
6249 ignore a missing NewSessionTicket message.
6253 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6257 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6258 sends a carefully crafted handshake message, to cause OpenSSL to fail
6259 to free up to 64k of memory causing a memory leak. This could be
6260 exploited in a Denial Of Service attack. This issue affects OpenSSL
6261 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6262 whether SRTP is used or configured. Implementations of OpenSSL that
6263 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6265 The fix was developed by the OpenSSL team.
6270 * Session Ticket Memory Leak.
6272 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6273 integrity of that ticket is first verified. In the event of a session
6274 ticket integrity check failing, OpenSSL will fail to free memory
6275 causing a memory leak. By sending a large number of invalid session
6276 tickets an attacker could exploit this issue in a Denial Of Service
6282 * Build option no-ssl3 is incomplete.
6284 When OpenSSL is configured with "no-ssl3" as a build option, servers
6285 could accept and complete a SSL 3.0 handshake, and clients could be
6286 configured to send them.
6289 *Akamai and the OpenSSL team*
6291 * Add support for TLS_FALLBACK_SCSV.
6292 Client applications doing fallback retries should call
6293 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6296 *Adam Langley, Bodo Moeller*
6298 * Add additional DigestInfo checks.
6300 Re-encode DigestInto in DER and check against the original when
6301 verifying RSA signature: this will reject any improperly encoded
6302 DigestInfo structures.
6304 Note: this is a precautionary measure and no attacks are currently known.
6308 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6310 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6311 SRP code can be overrun an internal buffer. Add sanity check that
6312 g, A, B < N to SRP code.
6314 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6315 Group for discovering this issue.
6320 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6321 TLS 1.0 instead of higher protocol versions when the ClientHello message
6322 is badly fragmented. This allows a man-in-the-middle attacker to force a
6323 downgrade to TLS 1.0 even if both the server and the client support a
6324 higher protocol version, by modifying the client's TLS records.
6326 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6327 researching this issue.
6332 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6333 to a denial of service attack. A malicious server can crash the client
6334 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6335 ciphersuite and sending carefully crafted handshake messages.
6337 Thanks to Felix Gröbert (Google) for discovering and researching this
6343 * By sending carefully crafted DTLS packets an attacker could cause openssl
6344 to leak memory. This can be exploited through a Denial of Service attack.
6345 Thanks to Adam Langley for discovering and researching this issue.
6350 * An attacker can force openssl to consume large amounts of memory whilst
6351 processing DTLS handshake messages. This can be exploited through a
6352 Denial of Service attack.
6353 Thanks to Adam Langley for discovering and researching this issue.
6358 * An attacker can force an error condition which causes openssl to crash
6359 whilst processing DTLS packets due to memory being freed twice. This
6360 can be exploited through a Denial of Service attack.
6361 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6367 * If a multithreaded client connects to a malicious server using a resumed
6368 session and the server sends an ec point format extension it could write
6369 up to 255 bytes to freed memory.
6371 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6377 * A malicious server can crash an OpenSSL client with a null pointer
6378 dereference (read) by specifying an SRP ciphersuite even though it was not
6379 properly negotiated with the client. This can be exploited through a
6380 Denial of Service attack.
6382 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6383 discovering and researching this issue.
6388 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6389 X509_name_oneline, X509_name_print_ex et al. to leak some information
6390 from the stack. Applications may be affected if they echo pretty printing
6391 output to the attacker.
6393 Thanks to Ivan Fratric (Google) for discovering this issue.
6396 *Emilia Käsper, and Steve Henson*
6398 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6399 for corner cases. (Certain input points at infinity could lead to
6400 bogus results, with non-infinity inputs mapped to infinity too.)
6404 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6406 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6407 handshake can force the use of weak keying material in OpenSSL
6408 SSL/TLS clients and servers.
6410 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6411 researching this issue. [CVE-2014-0224][]
6413 *KIKUCHI Masashi, Steve Henson*
6415 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6416 OpenSSL DTLS client the code can be made to recurse eventually crashing
6419 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6422 *Imre Rad, Steve Henson*
6424 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6425 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6426 client or server. This is potentially exploitable to run arbitrary
6427 code on a vulnerable client or server.
6429 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6431 *Jüri Aedla, Steve Henson*
6433 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6434 are subject to a denial of service attack.
6436 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6437 this issue. [CVE-2014-3470][]
6439 *Felix Gröbert, Ivan Fratric, Steve Henson*
6441 * Harmonize version and its documentation. -f flag is used to display
6444 *mancha <mancha1@zoho.com>*
6446 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6447 in i2d_ECPrivateKey.
6449 *mancha <mancha1@zoho.com>*
6451 * Fix some double frees. These are not thought to be exploitable.
6453 *mancha <mancha1@zoho.com>*
6455 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6457 * A missing bounds check in the handling of the TLS heartbeat extension
6458 can be used to reveal up to 64k of memory to a connected client or
6461 Thanks for Neel Mehta of Google Security for discovering this bug and to
6462 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6463 preparing the fix [CVE-2014-0160][]
6465 *Adam Langley, Bodo Moeller*
6467 * Fix for the attack described in the paper "Recovering OpenSSL
6468 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6469 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6470 <http://eprint.iacr.org/2014/140>
6472 Thanks to Yuval Yarom and Naomi Benger for discovering this
6473 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6475 *Yuval Yarom and Naomi Benger*
6477 * TLS pad extension: draft-agl-tls-padding-03
6479 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6480 TLS client Hello record length value would otherwise be > 255 and
6481 less that 512 pad with a dummy extension containing zeroes so it
6482 is at least 512 bytes long.
6484 *Adam Langley, Steve Henson*
6486 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6488 * Fix for TLS record tampering bug. A carefully crafted invalid
6489 handshake could crash OpenSSL with a NULL pointer exception.
6490 Thanks to Anton Johansson for reporting this issues.
6493 * Keep original DTLS digest and encryption contexts in retransmission
6494 structures so we can use the previous session parameters if they need
6495 to be resent. [CVE-2013-6450][]
6499 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6500 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6501 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6502 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6503 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6504 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6506 *Rob Stradling, Adam Langley*
6508 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6510 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6511 supporting platforms or when small records were transferred.
6513 *Andy Polyakov, Steve Henson*
6515 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6517 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6519 This addresses the flaw in CBC record processing discovered by
6520 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6521 at: <http://www.isg.rhul.ac.uk/tls/>
6523 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6524 Security Group at Royal Holloway, University of London
6525 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6526 Emilia Käsper for the initial patch.
6529 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6531 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6532 ciphersuites which can be exploited in a denial of service attack.
6533 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6534 and detecting this bug and to Wolfgang Ettlinger
6535 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6540 * Return an error when checking OCSP signatures when key is NULL.
6541 This fixes a DoS attack. [CVE-2013-0166][]
6545 * Make openssl verify return errors.
6547 *Chris Palmer <palmer@google.com> and Ben Laurie*
6549 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6550 the right response is stapled. Also change SSL_get_certificate()
6551 so it returns the certificate actually sent.
6552 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6554 *Rob Stradling <rob.stradling@comodo.com>*
6556 * Fix possible deadlock when decoding public keys.
6560 * Don't use TLS 1.0 record version number in initial client hello
6565 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6567 * Sanity check record length before skipping explicit IV in TLS
6568 1.2, 1.1 and DTLS to fix DoS attack.
6570 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6571 fuzzing as a service testing platform.
6576 * Initialise tkeylen properly when encrypting CMS messages.
6577 Thanks to Solar Designer of Openwall for reporting this issue.
6581 * In FIPS mode don't try to use composite ciphers as they are not
6586 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6588 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6589 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6590 mean any application compiled against OpenSSL 1.0.0 headers setting
6591 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6592 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6593 0x10000000L Any application which was previously compiled against
6594 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6595 will need to be recompiled as a result. Letting be results in
6596 inability to disable specifically TLS 1.1 and in client context,
6597 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6601 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6602 disable just protocol X, but all protocols above X *if* there are
6603 protocols *below* X still enabled. In more practical terms it means
6604 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6605 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6606 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6611 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6613 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6614 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6615 in CRYPTO_realloc_clean.
6617 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6618 issue and to Adam Langley <agl@chromium.org> for fixing it.
6621 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6623 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6627 * Workarounds for some broken servers that "hang" if a client hello
6628 record length exceeds 255 bytes.
6630 1. Do not use record version number > TLS 1.0 in initial client
6631 hello: some (but not all) hanging servers will now work.
6632 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6633 the number of ciphers sent in the client hello. This should be
6634 set to an even number, such as 50, for example by passing:
6635 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6636 Most broken servers should now work.
6637 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6638 TLS 1.2 client support entirely.
6642 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6646 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6648 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6649 STRING form instead of a DigestInfo.
6653 * The format used for MDC2 RSA signatures is inconsistent between EVP
6654 and the RSA_sign/RSA_verify functions. This was made more apparent when
6655 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6656 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6657 the correct format in RSA_verify so both forms transparently work.
6661 * Some servers which support TLS 1.0 can choke if we initially indicate
6662 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6663 encrypted premaster secret. As a workaround use the maximum permitted
6664 client version in client hello, this should keep such servers happy
6665 and still work with previous versions of OpenSSL.
6669 * Add support for TLS/DTLS heartbeats.
6671 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6673 * Add support for SCTP.
6675 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6677 * Improved PRNG seeding for VOS.
6679 *Paul Green <Paul.Green@stratus.com>*
6681 * Extensive assembler packs updates, most notably:
6683 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6684 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6685 - x86_64: bit-sliced AES implementation;
6686 - ARM: NEON support, contemporary platforms optimizations;
6687 - s390x: z196 support;
6688 - `*`: GHASH and GF(2^m) multiplication implementations;
6692 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6693 (removal of unnecessary code)
6695 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6697 * Add TLS key material exporter from RFC 5705.
6701 * Add DTLS-SRTP negotiation from RFC 5764.
6705 * Add Next Protocol Negotiation,
6706 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6707 disabled with a no-npn flag to config or Configure. Code donated
6710 *Adam Langley <agl@google.com> and Ben Laurie*
6712 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6713 NIST-P256, NIST-P521, with constant-time single point multiplication on
6714 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6715 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6716 Code made available under Apache License version 2.0.
6718 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6719 line to include this in your build of OpenSSL, and run "make depend" (or
6720 "make update"). This enables the following EC_METHODs:
6722 EC_GFp_nistp224_method()
6723 EC_GFp_nistp256_method()
6724 EC_GFp_nistp521_method()
6726 EC_GROUP_new_by_curve_name() will automatically use these (while
6727 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6730 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6732 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6733 all platforms. Move ssize_t definition from e_os.h to the public
6734 header file e_os2.h as it now appears in public header file cms.h
6738 * New -sigopt option to the ca, req and x509 utilities. Additional
6739 signature parameters can be passed using this option and in
6744 * Add RSA PSS signing function. This will generate and set the
6745 appropriate AlgorithmIdentifiers for PSS based on those in the
6746 corresponding EVP_MD_CTX structure. No application support yet.
6750 * Support for companion algorithm specific ASN1 signing routines.
6751 New function ASN1_item_sign_ctx() signs a pre-initialised
6752 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6753 the appropriate parameters.
6757 * Add new algorithm specific ASN1 verification initialisation function
6758 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6759 handling will be the same no matter what EVP_PKEY_METHOD is used.
6760 Add a PSS handler to support verification of PSS signatures: checked
6761 against a number of sample certificates.
6765 * Add signature printing for PSS. Add PSS OIDs.
6767 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6769 * Add algorithm specific signature printing. An individual ASN1 method
6770 can now print out signatures instead of the standard hex dump.
6772 More complex signatures (e.g. PSS) can print out more meaningful
6773 information. Include DSA version that prints out the signature
6778 * Password based recipient info support for CMS library: implementing
6783 * Split password based encryption into PBES2 and PBKDF2 functions. This
6784 neatly separates the code into cipher and PBE sections and is required
6785 for some algorithms that split PBES2 into separate pieces (such as
6786 password based CMS).
6790 * Session-handling fixes:
6791 - Fix handling of connections that are resuming with a session ID,
6792 but also support Session Tickets.
6793 - Fix a bug that suppressed issuing of a new ticket if the client
6794 presented a ticket with an expired session.
6795 - Try to set the ticket lifetime hint to something reasonable.
6796 - Make tickets shorter by excluding irrelevant information.
6797 - On the client side, don't ignore renewed tickets.
6799 *Adam Langley, Bodo Moeller (Google)*
6801 * Fix PSK session representation.
6805 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6807 This work was sponsored by Intel.
6811 * Add GCM support to TLS library. Some custom code is needed to split
6812 the IV between the fixed (from PRF) and explicit (from TLS record)
6813 portions. This adds all GCM ciphersuites supported by RFC5288 and
6814 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6815 add a special AESGCM string for GCM only.
6819 * Expand range of ctrls for AES GCM. Permit setting invocation
6820 field on decrypt and retrieval of invocation field only on encrypt.
6824 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6825 As required by RFC5289 these ciphersuites cannot be used if for
6826 versions of TLS earlier than 1.2.
6830 * For FIPS capable OpenSSL interpret a NULL default public key method
6831 as unset and return the appropriate default but do *not* set the default.
6832 This means we can return the appropriate method in applications that
6833 switch between FIPS and non-FIPS modes.
6837 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6838 ENGINE is used then we cannot handle that in the FIPS module so we
6839 keep original code iff non-FIPS operations are allowed.
6843 * Add -attime option to openssl utilities.
6845 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6847 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6851 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6852 FIPS EC methods unconditionally for now.
6856 * New build option no-ec2m to disable characteristic 2 code.
6860 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6861 all cases can be covered as some introduce binary incompatibilities.
6865 * Redirect RSA operations to FIPS module including keygen,
6866 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6870 * Add similar low level API blocking to ciphers.
6874 * Low level digest APIs are not approved in FIPS mode: any attempt
6875 to use these will cause a fatal error. Applications that *really* want
6876 to use them can use the `private_*` version instead.
6880 * Redirect cipher operations to FIPS module for FIPS builds.
6884 * Redirect digest operations to FIPS module for FIPS builds.
6888 * Update build system to add "fips" flag which will link in fipscanister.o
6889 for static and shared library builds embedding a signature if needed.
6893 * Output TLS supported curves in preference order instead of numerical
6894 order. This is currently hardcoded for the highest order curves first.
6895 This should be configurable so applications can judge speed vs strength.
6899 * Add TLS v1.2 server support for client authentication.
6903 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6908 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6909 FIPS modules versions.
6913 * Add TLS v1.2 client side support for client authentication. Keep cache
6914 of handshake records longer as we don't know the hash algorithm to use
6915 until after the certificate request message is received.
6919 * Initial TLS v1.2 client support. Add a default signature algorithms
6920 extension including all the algorithms we support. Parse new signature
6921 format in client key exchange. Relax some ECC signing restrictions for
6922 TLS v1.2 as indicated in RFC5246.
6926 * Add server support for TLS v1.2 signature algorithms extension. Switch
6927 to new signature format when needed using client digest preference.
6928 All server ciphersuites should now work correctly in TLS v1.2. No client
6929 support yet and no support for client certificates.
6933 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6934 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6935 ciphersuites. At present only RSA key exchange ciphersuites work with
6936 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6937 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6938 and version checking.
6942 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6943 with this defined it will not be affected by any changes to ssl internal
6944 structures. Add several utility functions to allow openssl application
6945 to work with OPENSSL_NO_SSL_INTERN defined.
6949 * A long standing patch to add support for SRP from EdelWeb (Peter
6950 Sylvester and Christophe Renou) was integrated.
6951 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6952 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6955 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6959 * Permit abbreviated handshakes when renegotiating using the function
6960 SSL_renegotiate_abbreviated().
6962 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6964 * Add call to ENGINE_register_all_complete() to
6965 ENGINE_load_builtin_engines(), so some implementations get used
6966 automatically instead of needing explicit application support.
6970 * Add support for TLS key exporter as described in RFC5705.
6972 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6974 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6975 a few changes are required:
6977 Add SSL_OP_NO_TLSv1_1 flag.
6978 Add TLSv1_1 methods.
6979 Update version checking logic to handle version 1.1.
6980 Add explicit IV handling (ported from DTLS code).
6981 Add command line options to s_client/s_server.
6988 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6990 * X509_ATTRIBUTE memory leak
6992 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6993 memory. This structure is used by the PKCS#7 and CMS routines so any
6994 application which reads PKCS#7 or CMS data from untrusted sources is
6995 affected. SSL/TLS is not affected.
6997 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7003 * Race condition handling PSK identify hint
7005 If PSK identity hints are received by a multi-threaded client then
7006 the values are wrongly updated in the parent SSL_CTX structure. This can
7007 result in a race condition potentially leading to a double free of the
7013 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7015 * Malformed ECParameters causes infinite loop
7017 When processing an ECParameters structure OpenSSL enters an infinite loop
7018 if the curve specified is over a specially malformed binary polynomial
7021 This can be used to perform denial of service against any
7022 system which processes public keys, certificate requests or
7023 certificates. This includes TLS clients and TLS servers with
7024 client authentication enabled.
7026 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7031 * Exploitable out-of-bounds read in X509_cmp_time
7033 X509_cmp_time does not properly check the length of the ASN1_TIME
7034 string and can read a few bytes out of bounds. In addition,
7035 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7038 An attacker can use this to craft malformed certificates and CRLs of
7039 various sizes and potentially cause a segmentation fault, resulting in
7040 a DoS on applications that verify certificates or CRLs. TLS clients
7041 that verify CRLs are affected. TLS clients and servers with client
7042 authentication enabled may be affected if they use custom verification
7045 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7046 independently by Hanno Böck.
7051 * PKCS7 crash with missing EnvelopedContent
7053 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7054 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7055 with missing content and trigger a NULL pointer dereference on parsing.
7057 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7058 structures from untrusted sources are affected. OpenSSL clients and
7059 servers are not affected.
7061 This issue was reported to OpenSSL by Michal Zalewski (Google).
7066 * CMS verify infinite loop with unknown hash function
7068 When verifying a signedData message the CMS code can enter an infinite loop
7069 if presented with an unknown hash function OID. This can be used to perform
7070 denial of service against any system which verifies signedData messages using
7072 This issue was reported to OpenSSL by Johannes Bauer.
7077 * Race condition handling NewSessionTicket
7079 If a NewSessionTicket is received by a multi-threaded client when attempting to
7080 reuse a previous ticket then a race condition can occur potentially leading to
7081 a double free of the ticket data.
7086 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7088 * Segmentation fault in ASN1_TYPE_cmp fix
7090 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7091 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7092 certificate signature algorithm consistency this can be used to crash any
7093 certificate verification operation and exploited in a DoS attack. Any
7094 application which performs certificate verification is vulnerable including
7095 OpenSSL clients and servers which enable client authentication.
7100 * ASN.1 structure reuse memory corruption fix
7102 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7103 memory corruption via an invalid write. Such reuse is and has been
7104 strongly discouraged and is believed to be rare.
7106 Applications that parse structures containing CHOICE or ANY DEFINED BY
7107 components may be affected. Certificate parsing (d2i_X509 and related
7108 functions) are however not affected. OpenSSL clients and servers are
7114 * PKCS7 NULL pointer dereferences fix
7116 The PKCS#7 parsing code does not handle missing outer ContentInfo
7117 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7118 missing content and trigger a NULL pointer dereference on parsing.
7120 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7121 otherwise parse PKCS#7 structures from untrusted sources are
7122 affected. OpenSSL clients and servers are not affected.
7124 This issue was reported to OpenSSL by Michal Zalewski (Google).
7129 * DoS via reachable assert in SSLv2 servers fix
7131 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7132 servers that both support SSLv2 and enable export cipher suites by sending
7133 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7135 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7136 (OpenSSL development team).
7141 * Use After Free following d2i_ECPrivatekey error fix
7143 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7144 could cause a use after free condition. This, in turn, could cause a double
7145 free in several private key parsing functions (such as d2i_PrivateKey
7146 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7147 for applications that receive EC private keys from untrusted
7148 sources. This scenario is considered rare.
7150 This issue was discovered by the BoringSSL project and fixed in their
7156 * X509_to_X509_REQ NULL pointer deref fix
7158 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7159 the certificate key is invalid. This function is rarely used in practice.
7161 This issue was discovered by Brian Carpenter.
7166 * Removed the export ciphers from the DEFAULT ciphers
7170 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7172 * Build fixes for the Windows and OpenVMS platforms
7174 *Matt Caswell and Richard Levitte*
7176 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7178 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7179 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7180 dereference. This could lead to a Denial Of Service attack. Thanks to
7181 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7186 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7187 dtls1_buffer_record function under certain conditions. In particular this
7188 could occur if an attacker sent repeated DTLS records with the same
7189 sequence number but for the next epoch. The memory leak could be exploited
7190 by an attacker in a Denial of Service attack through memory exhaustion.
7191 Thanks to Chris Mueller for reporting this issue.
7196 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7197 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7198 method would be set to NULL which could later result in a NULL pointer
7199 dereference. Thanks to Frank Schmirler for reporting this issue.
7204 * Abort handshake if server key exchange message is omitted for ephemeral
7207 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7208 reporting this issue.
7213 * Remove non-export ephemeral RSA code on client and server. This code
7214 violated the TLS standard by allowing the use of temporary RSA keys in
7215 non-export ciphersuites and could be used by a server to effectively
7216 downgrade the RSA key length used to a value smaller than the server
7217 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7218 INRIA or reporting this issue.
7223 * Fixed issue where DH client certificates are accepted without verification.
7224 An OpenSSL server will accept a DH certificate for client authentication
7225 without the certificate verify message. This effectively allows a client to
7226 authenticate without the use of a private key. This only affects servers
7227 which trust a client certificate authority which issues certificates
7228 containing DH keys: these are extremely rare and hardly ever encountered.
7229 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7235 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7236 results on some platforms, including x86_64. This bug occurs at random
7237 with a very low probability, and is not known to be exploitable in any
7238 way, though its exact impact is difficult to determine. Thanks to Pieter
7239 Wuille (Blockstream) who reported this issue and also suggested an initial
7240 fix. Further analysis was conducted by the OpenSSL development team and
7241 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7242 the OpenSSL core team.
7247 * Fix various certificate fingerprint issues.
7249 By using non-DER or invalid encodings outside the signed portion of a
7250 certificate the fingerprint can be changed without breaking the signature.
7251 Although no details of the signed portion of the certificate can be changed
7252 this can cause problems with some applications: e.g. those using the
7253 certificate fingerprint for blacklists.
7255 1. Reject signatures with non zero unused bits.
7257 If the BIT STRING containing the signature has non zero unused bits reject
7258 the signature. All current signature algorithms require zero unused bits.
7260 2. Check certificate algorithm consistency.
7262 Check the AlgorithmIdentifier inside TBS matches the one in the
7263 certificate signature. NB: this will result in signature failure
7264 errors for some broken certificates.
7266 Thanks to Konrad Kraszewski from Google for reporting this issue.
7268 3. Check DSA/ECDSA signatures use DER.
7270 Reencode DSA/ECDSA signatures and compare with the original received
7271 signature. Return an error if there is a mismatch.
7273 This will reject various cases including garbage after signature
7274 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7275 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7276 (negative or with leading zeroes).
7278 Further analysis was conducted and fixes were developed by Stephen Henson
7279 of the OpenSSL core team.
7285 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7287 * Session Ticket Memory Leak.
7289 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7290 integrity of that ticket is first verified. In the event of a session
7291 ticket integrity check failing, OpenSSL will fail to free memory
7292 causing a memory leak. By sending a large number of invalid session
7293 tickets an attacker could exploit this issue in a Denial Of Service
7299 * Build option no-ssl3 is incomplete.
7301 When OpenSSL is configured with "no-ssl3" as a build option, servers
7302 could accept and complete a SSL 3.0 handshake, and clients could be
7303 configured to send them.
7306 *Akamai and the OpenSSL team*
7308 * Add support for TLS_FALLBACK_SCSV.
7309 Client applications doing fallback retries should call
7310 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7313 *Adam Langley, Bodo Moeller*
7315 * Add additional DigestInfo checks.
7317 Reencode DigestInto in DER and check against the original when
7318 verifying RSA signature: this will reject any improperly encoded
7319 DigestInfo structures.
7321 Note: this is a precautionary measure and no attacks are currently known.
7325 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7327 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7328 to a denial of service attack. A malicious server can crash the client
7329 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7330 ciphersuite and sending carefully crafted handshake messages.
7332 Thanks to Felix Gröbert (Google) for discovering and researching this
7338 * By sending carefully crafted DTLS packets an attacker could cause openssl
7339 to leak memory. This can be exploited through a Denial of Service attack.
7340 Thanks to Adam Langley for discovering and researching this issue.
7345 * An attacker can force openssl to consume large amounts of memory whilst
7346 processing DTLS handshake messages. This can be exploited through a
7347 Denial of Service attack.
7348 Thanks to Adam Langley for discovering and researching this issue.
7353 * An attacker can force an error condition which causes openssl to crash
7354 whilst processing DTLS packets due to memory being freed twice. This
7355 can be exploited through a Denial of Service attack.
7356 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7362 * If a multithreaded client connects to a malicious server using a resumed
7363 session and the server sends an ec point format extension it could write
7364 up to 255 bytes to freed memory.
7366 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7372 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7373 X509_name_oneline, X509_name_print_ex et al. to leak some information
7374 from the stack. Applications may be affected if they echo pretty printing
7375 output to the attacker.
7377 Thanks to Ivan Fratric (Google) for discovering this issue.
7380 *Emilia Käsper, and Steve Henson*
7382 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7383 for corner cases. (Certain input points at infinity could lead to
7384 bogus results, with non-infinity inputs mapped to infinity too.)
7388 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7390 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7391 handshake can force the use of weak keying material in OpenSSL
7392 SSL/TLS clients and servers.
7394 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7395 researching this issue. [CVE-2014-0224][]
7397 *KIKUCHI Masashi, Steve Henson*
7399 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7400 OpenSSL DTLS client the code can be made to recurse eventually crashing
7403 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7406 *Imre Rad, Steve Henson*
7408 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7409 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7410 client or server. This is potentially exploitable to run arbitrary
7411 code on a vulnerable client or server.
7413 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7415 *Jüri Aedla, Steve Henson*
7417 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7418 are subject to a denial of service attack.
7420 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7421 this issue. [CVE-2014-3470][]
7423 *Felix Gröbert, Ivan Fratric, Steve Henson*
7425 * Harmonize version and its documentation. -f flag is used to display
7428 *mancha <mancha1@zoho.com>*
7430 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7431 in i2d_ECPrivateKey.
7433 *mancha <mancha1@zoho.com>*
7435 * Fix some double frees. These are not thought to be exploitable.
7437 *mancha <mancha1@zoho.com>*
7439 * Fix for the attack described in the paper "Recovering OpenSSL
7440 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7441 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7442 <http://eprint.iacr.org/2014/140>
7444 Thanks to Yuval Yarom and Naomi Benger for discovering this
7445 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7447 *Yuval Yarom and Naomi Benger*
7449 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7451 * Keep original DTLS digest and encryption contexts in retransmission
7452 structures so we can use the previous session parameters if they need
7453 to be resent. [CVE-2013-6450][]
7457 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7458 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7459 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7460 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7461 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7462 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7464 *Rob Stradling, Adam Langley*
7466 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7468 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7470 This addresses the flaw in CBC record processing discovered by
7471 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7472 at: <http://www.isg.rhul.ac.uk/tls/>
7474 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7475 Security Group at Royal Holloway, University of London
7476 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7477 Emilia Käsper for the initial patch.
7480 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7482 * Return an error when checking OCSP signatures when key is NULL.
7483 This fixes a DoS attack. [CVE-2013-0166][]
7487 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7488 the right response is stapled. Also change SSL_get_certificate()
7489 so it returns the certificate actually sent.
7490 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7491 (This is a backport)
7493 *Rob Stradling <rob.stradling@comodo.com>*
7495 * Fix possible deadlock when decoding public keys.
7499 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7501 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7504 * Sanity check record length before skipping explicit IV in DTLS
7507 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7508 fuzzing as a service testing platform.
7513 * Initialise tkeylen properly when encrypting CMS messages.
7514 Thanks to Solar Designer of Openwall for reporting this issue.
7518 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7520 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7521 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7522 in CRYPTO_realloc_clean.
7524 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7525 issue and to Adam Langley <agl@chromium.org> for fixing it.
7528 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7530 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7532 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7533 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7534 content decryption and always return the same error. Note: this attack
7535 needs on average 2^20 messages so it only affects automated senders. The
7536 old behaviour can be re-enabled in the CMS code by setting the
7537 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7538 an MMA defence is not necessary.
7539 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7540 this issue. [CVE-2012-0884][]
7544 * Fix CVE-2011-4619: make sure we really are receiving a
7545 client hello before rejecting multiple SGC restarts. Thanks to
7546 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7550 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7552 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7553 Thanks to Antonio Martin, Enterprise Secure Access Research and
7554 Development, Cisco Systems, Inc. for discovering this bug and
7555 preparing a fix. [CVE-2012-0050][]
7559 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7561 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7562 of the Vaudenay padding oracle attack on CBC mode encryption
7563 which enables an efficient plaintext recovery attack against
7564 the OpenSSL implementation of DTLS. Their attack exploits timing
7565 differences arising during decryption processing. A research
7566 paper describing this attack can be found at:
7567 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7568 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7569 Security Group at Royal Holloway, University of London
7570 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7571 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7572 for preparing the fix. [CVE-2011-4108][]
7574 *Robin Seggelmann, Michael Tuexen*
7576 * Clear bytes used for block padding of SSL 3.0 records.
7579 *Adam Langley (Google)*
7581 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7582 Kadianakis <desnacked@gmail.com> for discovering this issue and
7583 Adam Langley for preparing the fix. [CVE-2011-4619][]
7585 *Adam Langley (Google)*
7587 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7589 *Andrey Kulikov <amdeich@gmail.com>*
7591 * Prevent malformed RFC3779 data triggering an assertion failure.
7592 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7593 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7595 *Rob Austein <sra@hactrn.net>*
7597 * Improved PRNG seeding for VOS.
7599 *Paul Green <Paul.Green@stratus.com>*
7601 * Fix ssl_ciph.c set-up race.
7603 *Adam Langley (Google)*
7605 * Fix spurious failures in ecdsatest.c.
7607 *Emilia Käsper (Google)*
7609 * Fix the BIO_f_buffer() implementation (which was mixing different
7610 interpretations of the `..._len` fields).
7612 *Adam Langley (Google)*
7614 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7615 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7616 threads won't reuse the same blinding coefficients.
7618 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7619 lock to call BN_BLINDING_invert_ex, and avoids one use of
7620 BN_BLINDING_update for each BN_BLINDING structure (previously,
7621 the last update always remained unused).
7623 *Emilia Käsper (Google)*
7625 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7627 *Bob Buckholz (Google)*
7629 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7631 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7632 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7634 *Kaspar Brand <ossl@velox.ch>*
7636 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7637 for multi-threaded use of ECDH. [CVE-2011-3210][]
7639 *Adam Langley (Google)*
7641 * Fix x509_name_ex_d2i memory leak on bad inputs.
7645 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7646 signature public key algorithm by using OID xref utilities instead.
7647 Before this you could only use some ECC ciphersuites with SHA1 only.
7651 * Add protection against ECDSA timing attacks as mentioned in the paper
7652 by Billy Bob Brumley and Nicola Tuveri, see:
7653 <http://eprint.iacr.org/2011/232.pdf>
7655 *Billy Bob Brumley and Nicola Tuveri*
7657 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7659 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7661 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7663 * Fix bug in string printing code: if *any* escaping is enabled we must
7664 escape the escape character (backslash) or the resulting string is
7669 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7671 * Disable code workaround for ancient and obsolete Netscape browsers
7672 and servers: an attacker can use it in a ciphersuite downgrade attack.
7673 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7677 * Fixed J-PAKE implementation error, originally discovered by
7678 Sebastien Martini, further info and confirmation from Stefan
7679 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7683 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7685 * Fix extension code to avoid race conditions which can result in a buffer
7686 overrun vulnerability: resumed sessions must not be modified as they can
7687 be shared by multiple threads. CVE-2010-3864
7691 * Fix WIN32 build system to correctly link an ENGINE directory into
7696 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7698 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7701 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7703 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7705 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7706 context. The operation can be customised via the ctrl mechanism in
7707 case ENGINEs want to include additional functionality.
7711 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7715 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7716 output hashes compatible with older versions of OpenSSL.
7718 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7720 * Fix compression algorithm handling: if resuming a session use the
7721 compression algorithm of the resumed session instead of determining
7722 it from client hello again. Don't allow server to change algorithm.
7726 * Add load_crls() function to apps tidying load_certs() too. Add option
7727 to verify utility to allow additional CRLs to be included.
7731 * Update OCSP request code to permit adding custom headers to the request:
7732 some responders need this.
7736 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7739 *Julia Lawall <julia@diku.dk>*
7741 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7742 needlessly dereferenced structures, used obsolete functions and
7743 didn't handle all updated verify codes correctly.
7747 * Disable MD2 in the default configuration.
7751 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7752 indicate the initial BIO being pushed or popped. This makes it possible
7753 to determine whether the BIO is the one explicitly called or as a result
7754 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7755 it handles reference counts correctly and doesn't zero out the I/O bio
7756 when it is not being explicitly popped. WARNING: applications which
7757 included workarounds for the old buggy behaviour will need to be modified
7758 or they could free up already freed BIOs.
7762 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7763 renaming to all platforms (within the 0.9.8 branch, this was
7764 done conditionally on Netware platforms to avoid a name clash).
7766 *Guenter <lists@gknw.net>*
7768 * Add ECDHE and PSK support to DTLS.
7770 *Michael Tuexen <tuexen@fh-muenster.de>*
7772 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7777 * Add "missing" function EVP_MD_flags() (without this the only way to
7778 retrieve a digest flags is by accessing the structure directly. Update
7779 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7780 or cipher is registered as in the "from" argument. Print out all
7781 registered digests in the dgst usage message instead of manually
7782 attempting to work them out.
7786 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7787 this allows the use of compression and extensions. Change default cipher
7788 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7789 by default unless an application cipher string requests it.
7793 * Alter match criteria in PKCS12_parse(). It used to try to use local
7794 key ids to find matching certificates and keys but some PKCS#12 files
7795 don't follow the (somewhat unwritten) rules and this strategy fails.
7796 Now just gather all certificates together and the first private key
7797 then look for the first certificate that matches the key.
7801 * Support use of registered digest and cipher names for dgst and cipher
7802 commands instead of having to add each one as a special case. So now
7809 openssl dgst -sha256 foo
7811 and this works for ENGINE based algorithms too.
7815 * Update Gost ENGINE to support parameter files.
7817 *Victor B. Wagner <vitus@cryptocom.ru>*
7819 * Support GeneralizedTime in ca utility.
7821 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7823 * Enhance the hash format used for certificate directory links. The new
7824 form uses the canonical encoding (meaning equivalent names will work
7825 even if they aren't identical) and uses SHA1 instead of MD5. This form
7826 is incompatible with the older format and as a result c_rehash should
7827 be used to rebuild symbolic links.
7831 * Make PKCS#8 the default write format for private keys, replacing the
7832 traditional format. This form is standardised, more secure and doesn't
7833 include an implicit MD5 dependency.
7837 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7838 committed to OpenSSL should pass this lot as a minimum.
7842 * Add session ticket override functionality for use by EAP-FAST.
7844 *Jouni Malinen <j@w1.fi>*
7846 * Modify HMAC functions to return a value. Since these can be implemented
7847 in an ENGINE errors can occur.
7851 * Type-checked OBJ_bsearch_ex.
7855 * Type-checked OBJ_bsearch. Also some constification necessitated
7856 by type-checking. Still to come: TXT_DB, bsearch(?),
7857 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7862 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7863 seconds to a tm structure directly, instead of going through OS
7864 specific date routines. This avoids any issues with OS routines such
7865 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7866 and X509_time_adj_ex() to cover the extended range. The existing
7867 X509_time_adj() is still usable and will no longer have any date issues.
7871 * Delta CRL support. New use deltas option which will attempt to locate
7872 and search any appropriate delta CRLs available.
7874 This work was sponsored by Google.
7878 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7879 code and add additional score elements. Validate alternate CRL paths
7880 as part of the CRL checking and indicate a new error "CRL path validation
7881 error" in this case. Applications wanting additional details can use
7882 the verify callback and check the new "parent" field. If this is not
7883 NULL CRL path validation is taking place. Existing applications won't
7884 see this because it requires extended CRL support which is off by
7887 This work was sponsored by Google.
7891 * Support for freshest CRL extension.
7893 This work was sponsored by Google.
7897 * Initial indirect CRL support. Currently only supported in the CRLs
7898 passed directly and not via lookup. Process certificate issuer
7899 CRL entry extension and lookup CRL entries by bother issuer name
7900 and serial number. Check and process CRL issuer entry in IDP extension.
7902 This work was sponsored by Google.
7906 * Add support for distinct certificate and CRL paths. The CRL issuer
7907 certificate is validated separately in this case. Only enabled if
7908 an extended CRL support flag is set: this flag will enable additional
7909 CRL functionality in future.
7911 This work was sponsored by Google.
7915 * Add support for policy mappings extension.
7917 This work was sponsored by Google.
7921 * Fixes to pathlength constraint, self issued certificate handling,
7922 policy processing to align with RFC3280 and PKITS tests.
7924 This work was sponsored by Google.
7928 * Support for name constraints certificate extension. DN, email, DNS
7929 and URI types are currently supported.
7931 This work was sponsored by Google.
7935 * To cater for systems that provide a pointer-based thread ID rather
7936 than numeric, deprecate the current numeric thread ID mechanism and
7937 replace it with a structure and associated callback type. This
7938 mechanism allows a numeric "hash" to be extracted from a thread ID in
7939 either case, and on platforms where pointers are larger than 'long',
7940 mixing is done to help ensure the numeric 'hash' is usable even if it
7941 can't be guaranteed unique. The default mechanism is to use "&errno"
7942 as a pointer-based thread ID to distinguish between threads.
7944 Applications that want to provide their own thread IDs should now use
7945 CRYPTO_THREADID_set_callback() to register a callback that will call
7946 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7948 Note that ERR_remove_state() is now deprecated, because it is tied
7949 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7950 to free the current thread's error state should be replaced by
7951 ERR_remove_thread_state(NULL).
7953 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7954 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7955 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7956 application was previously providing a numeric thread callback that
7957 was inappropriate for distinguishing threads, then uniqueness might
7958 have been obtained with &errno that happened immediately in the
7959 intermediate development versions of OpenSSL; this is no longer the
7960 case, the numeric thread callback will now override the automatic use
7963 *Geoff Thorpe, with help from Bodo Moeller*
7965 * Initial support for different CRL issuing certificates. This covers a
7966 simple case where the self issued certificates in the chain exist and
7967 the real CRL issuer is higher in the existing chain.
7969 This work was sponsored by Google.
7973 * Removed effectively defunct crypto/store from the build.
7977 * Revamp of STACK to provide stronger type-checking. Still to come:
7978 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7979 ASN1_STRING, CONF_VALUE.
7983 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7984 RAM on SSL connections. This option can save about 34k per idle SSL.
7988 * Revamp of LHASH to provide stronger type-checking. Still to come:
7989 STACK, TXT_DB, bsearch, qsort.
7993 * Initial support for Cryptographic Message Syntax (aka CMS) based
7994 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7995 support for data, signedData, compressedData, digestedData and
7996 encryptedData, envelopedData types included. Scripts to check against
7997 RFC4134 examples draft and interop and consistency checks of many
7998 content types and variants.
8002 * Add options to enc utility to support use of zlib compression BIO.
8006 * Extend mk1mf to support importing of options and assembly language
8007 files from Configure script, currently only included in VC-WIN32.
8008 The assembly language rules can now optionally generate the source
8009 files from the associated perl scripts.
8013 * Implement remaining functionality needed to support GOST ciphersuites.
8014 Interop testing has been performed using CryptoPro implementations.
8016 *Victor B. Wagner <vitus@cryptocom.ru>*
8018 * s390x assembler pack.
8022 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8027 * Implement Opaque PRF Input TLS extension as specified in
8028 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8029 official specification yet and no extension type assignment by
8030 IANA exists, this extension (for now) will have to be explicitly
8031 enabled when building OpenSSL by providing the extension number
8032 to use. For example, specify an option
8034 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8036 to the "config" or "Configure" script to enable the extension,
8037 assuming extension number 0x9527 (which is a completely arbitrary
8038 and unofficial assignment based on the MD5 hash of the Internet
8039 Draft). Note that by doing so, you potentially lose
8040 interoperability with other TLS implementations since these might
8041 be using the same extension number for other purposes.
8043 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8044 opaque PRF input value to use in the handshake. This will create
8045 an internal copy of the length-'len' string at 'src', and will
8046 return non-zero for success.
8048 To get more control and flexibility, provide a callback function
8051 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8052 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8056 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8059 Callback function 'cb' will be called in handshakes, and is
8060 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8061 Argument 'arg' is for application purposes (the value as given to
8062 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8063 be provided to the callback function). The callback function
8064 has to return non-zero to report success: usually 1 to use opaque
8065 PRF input just if possible, or 2 to enforce use of the opaque PRF
8066 input. In the latter case, the library will abort the handshake
8067 if opaque PRF input is not successfully negotiated.
8069 Arguments 'peerinput' and 'len' given to the callback function
8070 will always be NULL and 0 in the case of a client. A server will
8071 see the client's opaque PRF input through these variables if
8072 available (NULL and 0 otherwise). Note that if the server
8073 provides an opaque PRF input, the length must be the same as the
8074 length of the client's opaque PRF input.
8076 Note that the callback function will only be called when creating
8077 a new session (session resumption can resume whatever was
8078 previously negotiated), and will not be called in SSL 2.0
8079 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8080 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8081 for applications that need to enforce opaque PRF input.
8085 * Update ssl code to support digests other than SHA1+MD5 for handshake
8088 *Victor B. Wagner <vitus@cryptocom.ru>*
8090 * Add RFC4507 support to OpenSSL. This includes the corrections in
8091 RFC4507bis. The encrypted ticket format is an encrypted encoded
8092 SSL_SESSION structure, that way new session features are automatically
8095 If a client application caches session in an SSL_SESSION structure
8096 support is transparent because tickets are now stored in the encoded
8099 The SSL_CTX structure automatically generates keys for ticket
8100 protection in servers so again support should be possible
8101 with no application modification.
8103 If a client or server wishes to disable RFC4507 support then the option
8104 SSL_OP_NO_TICKET can be set.
8106 Add a TLS extension debugging callback to allow the contents of any client
8107 or server extensions to be examined.
8109 This work was sponsored by Google.
8113 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8114 OpenSSL should now compile cleanly on gcc 4.2
8116 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8118 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8119 support including streaming MAC support: this is required for GOST
8120 ciphersuite support.
8122 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8124 * Add option -stream to use PKCS#7 streaming in smime utility. New
8125 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8126 to output in BER and PEM format.
8130 * Experimental support for use of HMAC via EVP_PKEY interface. This
8131 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8132 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8133 ENGINE support for HMAC keys which are unextractable. New -mac and
8134 -macopt options to dgst utility.
8138 * New option -sigopt to dgst utility. Update dgst to use
8139 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8140 alternative signing parameters such as X9.31 or PSS in the dgst
8145 * Change ssl_cipher_apply_rule(), the internal function that does
8146 the work each time a ciphersuite string requests enabling
8147 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8148 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8149 the order of disabled ciphersuites such that those ciphersuites
8150 that most recently went from enabled to disabled not only stay
8151 in order with respect to each other, but also have higher priority
8152 than other disabled ciphersuites the next time ciphersuites are
8155 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8156 the same ciphersuites as with "HIGH" alone, but in a specific
8157 order where the PSK ciphersuites come first (since they are the
8158 most recently disabled ciphersuites when "HIGH" is parsed).
8160 Also, change ssl_create_cipher_list() (using this new
8161 functionality) such that between otherwise identical
8162 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8167 * Change ssl_create_cipher_list() so that it automatically
8168 arranges the ciphersuites in reasonable order before starting
8169 to process the rule string. Thus, the definition for "DEFAULT"
8170 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8171 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8172 This makes it much easier to arrive at a reasonable default order
8173 in applications for which anonymous ciphers are OK (meaning
8174 that you can't actually use DEFAULT).
8176 *Bodo Moeller; suggested by Victor Duchovni*
8178 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8179 processing) into multiple integers instead of setting
8180 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8181 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8182 (These masks as well as the individual bit definitions are hidden
8183 away into the non-exported interface ssl/ssl_locl.h, so this
8184 change to the definition of the SSL_CIPHER structure shouldn't
8185 affect applications.) This give us more bits for each of these
8186 categories, so there is no longer a need to coagulate AES128 and
8187 AES256 into a single algorithm bit, and to coagulate Camellia128
8188 and Camellia256 into a single algorithm bit, which has led to all
8191 Thus, among other things, the kludge introduced in 0.9.7m and
8192 0.9.8e for masking out AES256 independently of AES128 or masking
8193 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8195 With the change, we also introduce new ciphersuite aliases that
8196 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8201 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8202 Use the leftmost N bytes of the signature input if the input is
8203 larger than the prime q (with N being the size in bytes of q).
8207 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8208 it yet and it is largely untested.
8212 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8216 * Initial incomplete changes to avoid need for function casts in OpenSSL
8217 some compilers (gcc 4.2 and later) reject their use. Safestack is
8218 reimplemented. Update ASN1 to avoid use of legacy functions.
8222 * Win32/64 targets are linked with Winsock2.
8226 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8227 to external functions. This can be used to increase CRL handling
8228 efficiency especially when CRLs are very large by (for example) storing
8229 the CRL revoked certificates in a database.
8233 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8234 new CRLs added to a directory can be used. New command line option
8235 -verify_return_error to s_client and s_server. This causes real errors
8236 to be returned by the verify callback instead of carrying on no matter
8237 what. This reflects the way a "real world" verify callback would behave.
8241 * GOST engine, supporting several GOST algorithms and public key formats.
8242 Kindly donated by Cryptocom.
8246 * Partial support for Issuing Distribution Point CRL extension. CRLs
8247 partitioned by DP are handled but no indirect CRL or reason partitioning
8248 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8249 selected via a scoring technique which handles IDP and AKID in CRLs.
8253 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8254 will ultimately be used for all verify operations: this will remove the
8255 X509_STORE dependency on certificate verification and allow alternative
8256 lookup methods. X509_STORE based implementations of these two callbacks.
8260 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8261 Modify get_crl() to find a valid (unexpired) CRL if possible.
8265 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8266 this would be called X509_CRL_cmp() but that name is already used by
8267 a function that just compares CRL issuer names. Cache several CRL
8268 extensions in X509_CRL structure and cache CRLDP in X509.
8272 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8273 this maps equivalent X509_NAME structures into a consistent structure.
8274 Name comparison can then be performed rapidly using memcmp().
8278 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8283 * Allow digests to supply their own micalg string for S/MIME type using
8284 the ctrl EVP_MD_CTRL_MICALG.
8288 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8289 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8290 ctrl. It can then customise the structure before and/or after signing
8295 * New function OBJ_add_sigid() to allow application defined signature OIDs
8296 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8297 to free up any added signature OIDs.
8301 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8302 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8303 digest and cipher tables. New options added to openssl utility:
8304 list-message-digest-algorithms and list-cipher-algorithms.
8308 * Change the array representation of binary polynomials: the list
8309 of degrees of non-zero coefficients is now terminated with -1.
8310 Previously it was terminated with 0, which was also part of the
8311 value; thus, the array representation was not applicable to
8312 polynomials where t^0 has coefficient zero. This change makes
8313 the array representation useful in a more general context.
8317 * Various modifications and fixes to SSL/TLS cipher string
8318 handling. For ECC, the code now distinguishes between fixed ECDH
8319 with RSA certificates on the one hand and with ECDSA certificates
8320 on the other hand, since these are separate ciphersuites. The
8321 unused code for Fortezza ciphersuites has been removed.
8323 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8324 (not "ECDHE"). For consistency with the code for DH
8325 certificates, use of ECDH certificates is now considered ECDH
8326 authentication, not RSA or ECDSA authentication (the latter is
8327 merely the CA's signing algorithm and not actively used in the
8330 The temporary ciphersuite alias "ECCdraft" is no longer
8331 available, and ECC ciphersuites are no longer excluded from "ALL"
8332 and "DEFAULT". The following aliases now exist for RFC 4492
8333 ciphersuites, most of these by analogy with the DH case:
8335 kECDHr - ECDH cert, signed with RSA
8336 kECDHe - ECDH cert, signed with ECDSA
8337 kECDH - ECDH cert (signed with either RSA or ECDSA)
8338 kEECDH - ephemeral ECDH
8339 ECDH - ECDH cert or ephemeral ECDH
8345 AECDH - anonymous ECDH
8346 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8350 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8351 Use correct micalg parameters depending on digest(s) in signed message.
8355 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8356 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8360 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8361 an engine to register a method. Add ENGINE lookups for methods and
8362 functional reference processing.
8366 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8367 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8372 * New -resign option to smime utility. This adds one or more signers
8373 to an existing PKCS#7 signedData structure. Also -md option to use an
8374 alternative message digest algorithm for signing.
8378 * Tidy up PKCS#7 routines and add new functions to make it easier to
8379 create PKCS7 structures containing multiple signers. Update smime
8380 application to support multiple signers.
8384 * New -macalg option to pkcs12 utility to allow setting of an alternative
8389 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8390 Reorganize PBE internals to lookup from a static table using NIDs,
8391 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8392 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8393 PRF which will be automatically used with PBES2.
8397 * Replace the algorithm specific calls to generate keys in "req" with the
8402 * Update PKCS#7 enveloped data routines to use new API. This is now
8403 supported by any public key method supporting the encrypt operation. A
8404 ctrl is added to allow the public key algorithm to examine or modify
8405 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8410 * Add a ctrl to asn1 method to allow a public key algorithm to express
8411 a default digest type to use. In most cases this will be SHA1 but some
8412 algorithms (such as GOST) need to specify an alternative digest. The
8413 return value indicates how strong the preference is 1 means optional and
8414 2 is mandatory (that is it is the only supported type). Modify
8415 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8416 use the default md. Update openssl utilities to use the default digest
8417 type for signing if it is not explicitly indicated.
8421 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8422 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8423 signing method from the key type. This effectively removes the link
8424 between digests and public key types.
8428 * Add an OID cross reference table and utility functions. Its purpose is to
8429 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8430 rsaEncryption. This will allow some of the algorithm specific hackery
8431 needed to use the correct OID to be removed.
8435 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8436 structures for PKCS7_sign(). They are now set up by the relevant public
8441 * Add provisional EC pkey method with support for ECDSA and ECDH.
8445 * Add support for key derivation (agreement) in the API, DH method and
8450 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8451 public and private key formats. As a side effect these add additional
8452 command line functionality not previously available: DSA signatures can be
8453 generated and verified using pkeyutl and DH key support and generation in
8460 *Oliver Tappe <zooey@hirschkaefer.de>*
8462 * New make target "install_html_docs" installs HTML renditions of the
8465 *Oliver Tappe <zooey@hirschkaefer.de>*
8467 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8468 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8469 support key and parameter generation and add initial key generation
8470 functionality for RSA.
8474 * Add functions for main EVP_PKEY_method operations. The undocumented
8475 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8476 `EVP_PKEY_{encrypt,decrypt}_old`.
8480 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8481 key API, doesn't do much yet.
8485 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8486 public key algorithms. New option to openssl utility:
8487 "list-public-key-algorithms" to print out info.
8491 * Implement the Supported Elliptic Curves Extension for
8492 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8496 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8497 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8501 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8502 utilities such as rsa, dsa, dsaparam etc except they process any key
8507 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8508 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8509 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8514 * Initial support for pluggable public key ASN1.
8515 De-spaghettify the public key ASN1 handling. Move public and private
8516 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8517 algorithm specific handling to a single module within the relevant
8518 algorithm directory. Add functions to allow (near) opaque processing
8519 of public and private key structures.
8523 * Implement the Supported Point Formats Extension for
8524 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8528 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8529 for the psk identity [hint] and the psk callback functions to the
8530 SSL_SESSION, SSL and SSL_CTX structure.
8533 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8537 SSL_CTX_use_psk_identity_hint
8538 SSL_get_psk_identity_hint
8539 SSL_get_psk_identity
8540 SSL_use_psk_identity_hint
8542 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8544 * Add RFC 3161 compliant time stamp request creation, response generation
8545 and response verification functionality.
8547 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8549 * Add initial support for TLS extensions, specifically for the server_name
8550 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8551 have new members for a host name. The SSL data structure has an
8552 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8553 stored in that context to allow for session resumption, even after the
8554 SSL has been switched to a new SSL_CTX in reaction to a client's
8555 server_name extension.
8557 New functions (subject to change):
8559 SSL_get_servername()
8560 SSL_get_servername_type()
8563 New CTRL codes and macros (subject to change):
8565 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8566 - SSL_CTX_set_tlsext_servername_callback()
8567 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8568 - SSL_CTX_set_tlsext_servername_arg()
8569 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8571 openssl s_client has a new '-servername ...' option.
8573 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8574 '-key2 ...', '-servername_fatal' (subject to change). This allows
8575 testing the HostName extension for a specific single host name ('-cert'
8576 and '-key' remain fallbacks for handshakes without HostName
8577 negotiation). If the unrecognized_name alert has to be sent, this by
8578 default is a warning; it becomes fatal with the '-servername_fatal'
8581 *Peter Sylvester, Remy Allais, Christophe Renou*
8583 * Whirlpool hash implementation is added.
8587 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8588 bn(64,32). Because of instruction set limitations it doesn't have
8589 any negative impact on performance. This was done mostly in order
8590 to make it possible to share assembler modules, such as bn_mul_mont
8591 implementations, between 32- and 64-bit builds without hassle.
8595 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8596 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8601 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8602 dedicated Montgomery multiplication procedure, is introduced.
8603 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8604 "64-bit" performance on certain 32-bit targets.
8608 * New option SSL_OP_NO_COMP to disable use of compression selectively
8609 in SSL structures. New SSL ctrl to set maximum send fragment size.
8610 Save memory by setting the I/O buffer sizes dynamically instead of
8611 using the maximum available value.
8615 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8616 in addition to the text details.
8620 * Very, very preliminary EXPERIMENTAL support for printing of general
8621 ASN1 structures. This currently produces rather ugly output and doesn't
8622 handle several customised structures at all.
8626 * Integrated support for PVK file format and some related formats such
8627 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8628 these in the 'rsa' and 'dsa' utilities.
8632 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8636 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8637 place for the (very old) "NETSCAPE" format certificates which are now
8638 handled using new ASN1 code equivalents.
8642 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8643 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8644 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8648 * Modify CRL distribution points extension code to print out previously
8649 unsupported fields. Enhance extension setting code to allow setting of
8654 * Add print and set support for Issuing Distribution Point CRL extension.
8658 * Change 'Configure' script to enable Camellia by default.
8665 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8667 * When rejecting SSL/TLS records due to an incorrect version number, never
8668 update s->server with a new major version number. As of
8669 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8670 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8671 the previous behavior could result in a read attempt at NULL when
8672 receiving specific incorrect SSL/TLS records once record payload
8673 protection is active. [CVE-2010-0740][]
8675 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8677 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8678 could be crashed if the relevant tables were not present (e.g. chrooted).
8680 *Tomas Hoger <thoger@redhat.com>*
8682 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8684 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8686 *Martin Olsson, Neel Mehta*
8688 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8689 accommodate for stack sorting, always a write lock!).
8693 * On some versions of WIN32 Heap32Next is very slow. This can cause
8694 excessive delays in the RAND_poll(): over a minute. As a workaround
8695 include a time check in the inner Heap32Next loop too.
8699 * The code that handled flushing of data in SSL/TLS originally used the
8700 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8701 the problem outlined in PR#1949. The fix suggested there however can
8702 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8703 of Apache). So instead simplify the code to flush unconditionally.
8704 This should be fine since flushing with no data to flush is a no op.
8708 * Handle TLS versions 2.0 and later properly and correctly use the
8709 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8710 off ancient servers have a habit of sticking around for a while...
8714 * Modify compression code so it frees up structures without using the
8715 ex_data callbacks. This works around a problem where some applications
8716 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8717 restarting) then use compression (e.g. SSL with compression) later.
8718 This results in significant per-connection memory leaks and
8719 has caused some security issues including CVE-2008-1678 and
8724 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8725 change when encrypting or decrypting.
8729 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8730 connect and renegotiate with servers which do not support RI.
8731 Until RI is more widely deployed this option is enabled by default.
8735 * Add "missing" ssl ctrls to clear options and mode.
8739 * If client attempts to renegotiate and doesn't support RI respond with
8740 a no_renegotiation alert as required by RFC5746. Some renegotiating
8741 TLS clients will continue a connection gracefully when they receive
8742 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8743 waiting for a server hello which it will never receive. Now we treat a
8744 received no_renegotiation alert as a fatal error. This is because
8745 applications requesting a renegotiation might well expect it to succeed
8746 and would have no code in place to handle the server denying it so the
8747 only safe thing to do is to terminate the connection.
8751 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8752 peer supports secure renegotiation and 0 otherwise. Print out peer
8753 renegotiation support in s_client/s_server.
8757 * Replace the highly broken and deprecated SPKAC certification method with
8758 the updated NID creation version. This should correctly handle UTF8.
8762 * Implement RFC5746. Re-enable renegotiation but require the extension
8763 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8764 turns out to be a bad idea. It has been replaced by
8765 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8766 SSL_CTX_set_options(). This is really not recommended unless you
8767 know what you are doing.
8769 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8771 * Fixes to stateless session resumption handling. Use initial_ctx when
8772 issuing and attempting to decrypt tickets in case it has changed during
8773 servername handling. Use a non-zero length session ID when attempting
8774 stateless session resumption: this makes it possible to determine if
8775 a resumption has occurred immediately after receiving server hello
8776 (several places in OpenSSL subtly assume this) instead of later in
8781 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8782 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8783 fixes for a few places where the return code is not checked
8786 *Julia Lawall <julia@diku.dk>*
8788 * Add --strict-warnings option to Configure script to include devteam
8789 warnings in other configurations.
8793 * Add support for --libdir option and LIBDIR variable in makefiles. This
8794 makes it possible to install openssl libraries in locations which
8795 have names other than "lib", for example "/usr/lib64" which some
8798 *Steve Henson, based on patch from Jeremy Utley*
8800 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8801 X690 8.9.12 and can produce some misleading textual output of OIDs.
8803 *Steve Henson, reported by Dan Kaminsky*
8805 * Delete MD2 from algorithm tables. This follows the recommendation in
8806 several standards that it is not used in new applications due to
8807 several cryptographic weaknesses. For binary compatibility reasons
8808 the MD2 API is still compiled in by default.
8812 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8817 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8818 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8821 *Guenter <lists@gknw.net>*
8823 * Fix the server certificate chain building code to use X509_verify_cert(),
8824 it used to have an ad-hoc builder which was unable to cope with anything
8825 other than a simple chain.
8827 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8829 * Don't check self signed certificate signatures in X509_verify_cert()
8830 by default (a flag can override this): it just wastes time without
8831 adding any security. As a useful side effect self signed root CAs
8832 with non-FIPS digests are now usable in FIPS mode.
8836 * In dtls1_process_out_of_seq_message() the check if the current message
8837 is already buffered was missing. For every new message was memory
8838 allocated, allowing an attacker to perform an denial of service attack
8839 with sending out of seq handshake messages until there is no memory
8840 left. Additionally every future message was buffered, even if the
8841 sequence number made no sense and would be part of another handshake.
8842 So only messages with sequence numbers less than 10 in advance will be
8843 buffered. [CVE-2009-1378][]
8845 *Robin Seggelmann, discovered by Daniel Mentz*
8847 * Records are buffered if they arrive with a future epoch to be
8848 processed after finishing the corresponding handshake. There is
8849 currently no limitation to this buffer allowing an attacker to perform
8850 a DOS attack with sending records with future epochs until there is no
8851 memory left. This patch adds the pqueue_size() function to determine
8852 the size of a buffer and limits the record buffer to 100 entries.
8855 *Robin Seggelmann, discovered by Daniel Mentz*
8857 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8858 parent structure is freed. [CVE-2009-1379][]
8862 * Handle non-blocking I/O properly in SSL_shutdown() call.
8864 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8866 * Add `2.5.4.*` OIDs
8868 *Ilya O. <vrghost@gmail.com>*
8870 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8872 * Disable renegotiation completely - this fixes a severe security
8873 problem [CVE-2009-3555][] at the cost of breaking all
8874 renegotiation. Renegotiation can be re-enabled by setting
8875 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8876 run-time. This is really not recommended unless you know what
8881 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8883 * Don't set val to NULL when freeing up structures, it is freed up by
8884 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8885 zeroing past the valid field. [CVE-2009-0789][]
8887 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8889 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8890 checked correctly. This would allow some invalid signed attributes to
8891 appear to verify correctly. [CVE-2009-0591][]
8893 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8895 * Reject UniversalString and BMPString types with invalid lengths. This
8896 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8897 a legal length. [CVE-2009-0590][]
8901 * Set S/MIME signing as the default purpose rather than setting it
8902 unconditionally. This allows applications to override it at the store
8907 * Permit restricted recursion of ASN1 strings. This is needed in practice
8908 to handle some structures.
8912 * Improve efficiency of mem_gets: don't search whole buffer each time
8915 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8917 * New -hex option for openssl rand.
8921 * Print out UTF8String and NumericString when parsing ASN1.
8925 * Support NumericString type for name components.
8929 * Allow CC in the environment to override the automatically chosen
8930 compiler. Note that nothing is done to ensure flags work with the
8935 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8937 * Properly check EVP_VerifyFinal() and similar return values
8940 *Ben Laurie, Bodo Moeller, Google Security Team*
8942 * Enable TLS extensions by default.
8946 * Allow the CHIL engine to be loaded, whether the application is
8947 multithreaded or not. (This does not release the developer from the
8948 obligation to set up the dynamic locking callbacks.)
8950 *Sander Temme <sander@temme.net>*
8952 * Use correct exit code if there is an error in dgst command.
8954 *Steve Henson; problem pointed out by Roland Dirlewanger*
8956 * Tweak Configure so that you need to say "experimental-jpake" to enable
8957 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8961 * Add experimental JPAKE support, including demo authentication in
8962 s_client and s_server.
8966 * Set the comparison function in v3_addr_canonize().
8968 *Rob Austein <sra@hactrn.net>*
8970 * Add support for XMPP STARTTLS in s_client.
8972 *Philip Paeps <philip@freebsd.org>*
8974 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8975 to ensure that even with this option, only ciphersuites in the
8976 server's preference list will be accepted. (Note that the option
8977 applies only when resuming a session, so the earlier behavior was
8978 just about the algorithm choice for symmetric cryptography.)
8982 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8984 * Fix NULL pointer dereference if a DTLS server received
8985 ChangeCipherSpec as first record [CVE-2009-1386][].
8989 * Fix a state transition in s3_srvr.c and d1_srvr.c
8990 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8994 * The fix in 0.9.8c that supposedly got rid of unsafe
8995 double-checked locking was incomplete for RSA blinding,
8996 addressing just one layer of what turns out to have been
8997 doubly unsafe triple-checked locking.
8999 So now fix this for real by retiring the MONT_HELPER macro
9000 in crypto/rsa/rsa_eay.c.
9002 *Bodo Moeller; problem pointed out by Marius Schilder*
9004 * Various precautionary measures:
9006 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9008 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9009 (NB: This would require knowledge of the secret session ticket key
9010 to exploit, in which case you'd be SOL either way.)
9012 - Change bn_nist.c so that it will properly handle input BIGNUMs
9013 outside the expected range.
9015 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9018 *Neel Mehta, Bodo Moeller*
9020 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9021 the load fails. Useful for distros.
9023 *Ben Laurie and the FreeBSD team*
9025 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9029 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9033 * Expand ENGINE to support engine supplied SSL client certificate functions.
9035 This work was sponsored by Logica.
9039 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9040 keystores. Support for SSL/TLS client authentication too.
9041 Not compiled unless enable-capieng specified to Configure.
9043 This work was sponsored by Logica.
9047 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9048 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9049 attribute creation routines such as certificate requests and PKCS#12
9054 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9056 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9057 handshake which could lead to a client crash as found using the
9058 Codenomicon TLS test suite [CVE-2008-1672][]
9060 *Steve Henson, Mark Cox*
9062 * Fix double free in TLS server name extensions which could lead to
9063 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
9067 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9069 Clear the error queue to ensure that error entries left from
9070 older function calls do not interfere with the correct operation.
9072 *Lutz Jaenicke, Erik de Castro Lopo*
9074 * Remove root CA certificates of commercial CAs:
9076 The OpenSSL project does not recommend any specific CA and does not
9077 have any policy with respect to including or excluding any CA.
9078 Therefore it does not make any sense to ship an arbitrary selection
9079 of root CA certificates with the OpenSSL software.
9083 * RSA OAEP patches to fix two separate invalid memory reads.
9084 The first one involves inputs when 'lzero' is greater than
9085 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9086 before the beginning of from). The second one involves inputs where
9087 the 'db' section contains nothing but zeroes (there is a one-byte
9088 invalid read after the end of 'db').
9090 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9092 * Partial backport from 0.9.9-dev:
9094 Introduce bn_mul_mont (dedicated Montgomery multiplication
9095 procedure) as a candidate for BIGNUM assembler implementation.
9096 While 0.9.9-dev uses assembler for various architectures, only
9097 x86_64 is available by default here in the 0.9.8 branch, and
9098 32-bit x86 is available through a compile-time setting.
9100 To try the 32-bit x86 assembler implementation, use Configure
9101 option "enable-montasm" (which exists only for this backport).
9103 As "enable-montasm" for 32-bit x86 disclaims code stability
9104 anyway, in this constellation we activate additional code
9105 backported from 0.9.9-dev for further performance improvements,
9106 namely BN_from_montgomery_word. (To enable this otherwise,
9107 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9109 *Andy Polyakov (backport partially by Bodo Moeller)*
9111 * Add TLS session ticket callback. This allows an application to set
9112 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9113 values. This is useful for key rollover for example where several key
9114 sets may exist with different names.
9118 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9119 This was broken until now in 0.9.8 releases, such that the only way
9120 a registered ENGINE could be used (assuming it initialises
9121 successfully on the host) was to explicitly set it as the default
9122 for the relevant algorithms. This is in contradiction with 0.9.7
9123 behaviour and the documentation. With this fix, when an ENGINE is
9124 registered into a given algorithm's table of implementations, the
9125 'uptodate' flag is reset so that auto-discovery will be used next
9126 time a new context for that algorithm attempts to select an
9129 *Ian Lister (tweaked by Geoff Thorpe)*
9131 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9132 implementation in the following ways:
9134 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9137 Lack of BER streaming support means one pass streaming processing is
9138 only supported if data is detached: setting the streaming flag is
9139 ignored for embedded content.
9141 CMS support is disabled by default and must be explicitly enabled
9142 with the enable-cms configuration option.
9146 * Update the GMP engine glue to do direct copies between BIGNUM and
9147 mpz_t when openssl and GMP use the same limb size. Otherwise the
9148 existing "conversion via a text string export" trick is still used.
9150 *Paul Sheer <paulsheer@gmail.com>*
9152 * Zlib compression BIO. This is a filter BIO which compressed and
9153 uncompresses any data passed through it.
9157 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9158 RFC3394 compatible AES key wrapping.
9162 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9163 sets string data without copying. X509_ALGOR_set0() and
9164 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9165 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9166 from an X509_ATTRIBUTE structure optionally checking it occurs only
9167 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9172 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9173 to get the expected BN_FLG_CONSTTIME behavior.
9175 *Bodo Moeller (Google)*
9179 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9180 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9181 - added some more tests to do_tests.pl
9182 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9183 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9184 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9185 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9186 - various changes to netware.pl to enable gcc-cross builds on Win32
9188 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9189 - various changes to fix missing prototype warnings
9190 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9191 - added AES, WHIRLPOOL and CPUID assembler code to build files
9192 - added missing AES assembler make rules to mk1mf.pl
9193 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9195 *Guenter Knauf <eflash@gmx.net>*
9197 * Implement certificate status request TLS extension defined in RFC3546.
9198 A client can set the appropriate parameters and receive the encoded
9199 OCSP response via a callback. A server can query the supplied parameters
9200 and set the encoded OCSP response in the callback. Add simplified examples
9201 to s_client and s_server.
9205 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9208 + Binary incompatibility of ssl_ctx_st structure
9209 + DTLS interoperation with non-compliant servers
9210 + Don't call get_session_cb() without proposed session
9211 + Fix ia64 assembler code
9213 *Andy Polyakov, Steve Henson*
9215 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9217 * DTLS Handshake overhaul. There were longstanding issues with
9218 OpenSSL DTLS implementation, which were making it impossible for
9219 RFC 4347 compliant client to communicate with OpenSSL server.
9220 Unfortunately just fixing these incompatibilities would "cut off"
9221 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9222 server keeps tolerating non RFC compliant syntax. The opposite is
9223 not true, 0.9.8f client can not communicate with earlier server.
9224 This update even addresses CVE-2007-4995.
9228 * Changes to avoid need for function casts in OpenSSL: some compilers
9229 (gcc 4.2 and later) reject their use.
9230 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9233 * Add RFC4507 support to OpenSSL. This includes the corrections in
9234 RFC4507bis. The encrypted ticket format is an encrypted encoded
9235 SSL_SESSION structure, that way new session features are automatically
9238 If a client application caches session in an SSL_SESSION structure
9239 support is transparent because tickets are now stored in the encoded
9242 The SSL_CTX structure automatically generates keys for ticket
9243 protection in servers so again support should be possible
9244 with no application modification.
9246 If a client or server wishes to disable RFC4507 support then the option
9247 SSL_OP_NO_TICKET can be set.
9249 Add a TLS extension debugging callback to allow the contents of any client
9250 or server extensions to be examined.
9252 This work was sponsored by Google.
9256 * Add initial support for TLS extensions, specifically for the server_name
9257 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9258 have new members for a host name. The SSL data structure has an
9259 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9260 stored in that context to allow for session resumption, even after the
9261 SSL has been switched to a new SSL_CTX in reaction to a client's
9262 server_name extension.
9264 New functions (subject to change):
9266 SSL_get_servername()
9267 SSL_get_servername_type()
9270 New CTRL codes and macros (subject to change):
9272 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9273 - SSL_CTX_set_tlsext_servername_callback()
9274 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9275 - SSL_CTX_set_tlsext_servername_arg()
9276 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9278 openssl s_client has a new '-servername ...' option.
9280 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9281 '-key2 ...', '-servername_fatal' (subject to change). This allows
9282 testing the HostName extension for a specific single host name ('-cert'
9283 and '-key' remain fallbacks for handshakes without HostName
9284 negotiation). If the unrecognized_name alert has to be sent, this by
9285 default is a warning; it becomes fatal with the '-servername_fatal'
9288 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9290 * Add AES and SSE2 assembly language support to VC++ build.
9294 * Mitigate attack on final subtraction in Montgomery reduction.
9298 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9299 (which previously caused an internal error).
9303 * Squeeze another 10% out of IGE mode when in != out.
9307 * AES IGE mode speedup.
9309 *Dean Gaudet (Google)*
9311 * Add the Korean symmetric 128-bit cipher SEED (see
9312 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9313 add SEED ciphersuites from RFC 4162:
9315 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9316 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9317 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9318 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9320 To minimize changes between patchlevels in the OpenSSL 0.9.8
9321 series, SEED remains excluded from compilation unless OpenSSL
9322 is configured with 'enable-seed'.
9324 *KISA, Bodo Moeller*
9326 * Mitigate branch prediction attacks, which can be practical if a
9327 single processor is shared, allowing a spy process to extract
9328 information. For detailed background information, see
9329 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9330 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9331 and Necessary Software Countermeasures"). The core of the change
9332 are new versions BN_div_no_branch() and
9333 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9334 respectively, which are slower, but avoid the security-relevant
9335 conditional branches. These are automatically called by BN_div()
9336 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9337 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9338 remove a conditional branch.
9340 BN_FLG_CONSTTIME is the new name for the previous
9341 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9342 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9343 in the exponent causes BN_mod_exp_mont() to use the alternative
9344 implementation in BN_mod_exp_mont_consttime().) The old name
9345 remains as a deprecated alias.
9347 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9348 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9349 constant-time implementations for more than just exponentiation.
9350 Here too the old name is kept as a deprecated alias.
9352 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9353 the BN_BLINDING structure gets an independent copy of the
9354 modulus. This means that the previous `BIGNUM *m` argument to
9355 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9356 essentially becomes `const BIGNUM *m`, although we can't actually
9357 change this in the header file before 0.9.9. It allows
9358 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9359 enable BN_FLG_CONSTTIME.
9361 *Matthew D Wood (Intel Corp)*
9363 * In the SSL/TLS server implementation, be strict about session ID
9364 context matching (which matters if an application uses a single
9365 external cache for different purposes). Previously,
9366 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9367 set. This did ensure strict client verification, but meant that,
9368 with applications using a single external cache for quite
9369 different requirements, clients could circumvent ciphersuite
9370 restrictions for a given session ID context by starting a session
9371 in a different context.
9375 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9376 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9377 authentication-only ciphersuites.
9381 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9382 not complete and could lead to a possible single byte overflow
9383 [CVE-2007-5135][] [Ben Laurie]
9385 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9387 * Since AES128 and AES256 (and similarly Camellia128 and
9388 Camellia256) share a single mask bit in the logic of
9389 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9390 kludge to work properly if AES128 is available and AES256 isn't
9391 (or if Camellia128 is available and Camellia256 isn't).
9395 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9396 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9397 When a point or a seed is encoded in a BIT STRING, we need to
9398 prevent the removal of trailing zero bits to get the proper DER
9399 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9400 of a NamedBitList, for which trailing 0 bits need to be removed.)
9404 * Have SSL/TLS server implementation tolerate "mismatched" record
9405 protocol version while receiving ClientHello even if the
9406 ClientHello is fragmented. (The server can't insist on the
9407 particular protocol version it has chosen before the ServerHello
9408 message has informed the client about his choice.)
9412 * Add RFC 3779 support.
9414 *Rob Austein for ARIN, Ben Laurie*
9416 * Load error codes if they are not already present instead of using a
9417 static variable. This allows them to be cleanly unloaded and reloaded.
9418 Improve header file function name parsing.
9422 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9423 or CAPABILITY handshake as required by RFCs.
9427 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9429 * Introduce limits to prevent malicious keys being able to
9430 cause a denial of service. [CVE-2006-2940][]
9432 *Steve Henson, Bodo Moeller*
9434 * Fix ASN.1 parsing of certain invalid structures that can result
9435 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9437 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9438 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9440 * Fix SSL client code which could crash if connecting to a
9441 malicious SSLv2 server. [CVE-2006-4343][]
9443 *Tavis Ormandy and Will Drewry, Google Security Team*
9445 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9446 match only those. Before that, "AES256-SHA" would be interpreted
9447 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9448 the same strength classification in 0.9.7h) as we currently only
9449 have a single AES bit in the ciphersuite description bitmap.
9450 That change, however, also applied to ciphersuite strings such as
9451 "RC4-MD5" that intentionally matched multiple ciphersuites --
9452 namely, SSL 2.0 ciphersuites in addition to the more common ones
9453 from SSL 3.0/TLS 1.0.
9455 So we change the selection algorithm again: Naming an explicit
9456 ciphersuite selects this one ciphersuite, and any other similar
9457 ciphersuite (same bitmap) from *other* protocol versions.
9458 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9459 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9461 Since SSL 2.0 does not have any ciphersuites for which the
9462 128/256 bit distinction would be relevant, this works for now.
9463 The proper fix will be to use different bits for AES128 and
9464 AES256, which would have avoided the problems from the beginning;
9465 however, bits are scarce, so we can only do this in a new release
9466 (not just a patchlevel) when we can change the SSL_CIPHER
9467 definition to split the single 'unsigned long mask' bitmap into
9468 multiple values to extend the available space.
9472 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9474 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9475 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9477 * Add AES IGE and biIGE modes.
9481 * Change the Unix randomness entropy gathering to use poll() when
9482 possible instead of select(), since the latter has some
9483 undesirable limitations.
9485 *Darryl Miles via Richard Levitte and Bodo Moeller*
9487 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9488 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9489 cannot be implicitly activated as part of, e.g., the "AES" alias.
9490 However, please upgrade to OpenSSL 0.9.9[-dev] for
9491 non-experimental use of the ECC ciphersuites to get TLS extension
9492 support, which is required for curve and point format negotiation
9493 to avoid potential handshake problems.
9497 * Disable rogue ciphersuites:
9499 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9500 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9501 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9503 The latter two were purportedly from
9504 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9507 Also deactivate the remaining ciphersuites from
9508 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9509 unofficial, and the ID has long expired.
9513 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9514 dual-core machines) and other potential thread-safety issues.
9518 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9519 versions), which is now available for royalty-free use
9520 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9521 Also, add Camellia TLS ciphersuites from RFC 4132.
9523 To minimize changes between patchlevels in the OpenSSL 0.9.8
9524 series, Camellia remains excluded from compilation unless OpenSSL
9525 is configured with 'enable-camellia'.
9529 * Disable the padding bug check when compression is in use. The padding
9530 bug check assumes the first packet is of even length, this is not
9531 necessarily true if compression is enabled and can result in false
9532 positives causing handshake failure. The actual bug test is ancient
9533 code so it is hoped that implementations will either have fixed it by
9534 now or any which still have the bug do not support compression.
9538 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9540 * When applying a cipher rule check to see if string match is an explicit
9541 cipher suite and only match that one cipher suite if it is.
9545 * Link in manifests for VC++ if needed.
9547 *Austin Ziegler <halostatue@gmail.com>*
9549 * Update support for ECC-based TLS ciphersuites according to
9550 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9551 TLS extensions, which are supported starting with the 0.9.9
9552 branch, not in the OpenSSL 0.9.8 branch).
9556 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9557 opaque EVP_CIPHER_CTX handling.
9561 * Fixes and enhancements to zlib compression code. We now only use
9562 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9563 to conform with the standards mentioned here:
9564 <http://www.zlib.net/DLL_FAQ.txt>
9565 Static zlib linking now works on Windows and the new --with-zlib-include
9566 --with-zlib-lib options to Configure can be used to supply the location
9567 of the headers and library. Gracefully handle case where zlib library
9572 * Several fixes and enhancements to the OID generation code. The old code
9573 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9574 handle numbers larger than ULONG_MAX, truncated printing and had a
9575 non standard OBJ_obj2txt() behaviour.
9579 * Add support for building of engines under engine/ as shared libraries
9580 under VC++ build system.
9584 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9585 Hopefully, we will not see any false combination of paths any more.
9589 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9591 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9592 (part of SSL_OP_ALL). This option used to disable the
9593 countermeasure against man-in-the-middle protocol-version
9594 rollback in the SSL 2.0 server implementation, which is a bad
9595 idea. [CVE-2005-2969][]
9597 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9598 for Information Security, National Institute of Advanced Industrial
9599 Science and Technology [AIST], Japan)*
9601 * Add two function to clear and return the verify parameter flags.
9605 * Keep cipherlists sorted in the source instead of sorting them at
9606 runtime, thus removing the need for a lock.
9610 * Avoid some small subgroup attacks in Diffie-Hellman.
9612 *Nick Mathewson and Ben Laurie*
9614 * Add functions for well-known primes.
9618 * Extended Windows CE support.
9620 *Satoshi Nakamura and Andy Polyakov*
9622 * Initialize SSL_METHOD structures at compile time instead of during
9623 runtime, thus removing the need for a lock.
9627 * Make PKCS7_decrypt() work even if no certificate is supplied by
9628 attempting to decrypt each encrypted key in turn. Add support to
9633 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9635 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9638 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9642 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9643 key into the same file any more.
9647 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9651 * Add -utf8 command line and config file option to 'ca'.
9653 *Stefan <stf@udoma.org*
9655 * Removed the macro des_crypt(), as it seems to conflict with some
9656 libraries. Use DES_crypt().
9660 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9661 involves renaming the source and generated shared-libs for
9662 both. The engines will accept the corrected or legacy ids
9663 ('ncipher' and '4758_cca' respectively) when binding. NB,
9664 this only applies when building 'shared'.
9666 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9668 * Add attribute functions to EVP_PKEY structure. Modify
9669 PKCS12_create() to recognize a CSP name attribute and
9670 use it. Make -CSP option work again in pkcs12 utility.
9674 * Add new functionality to the bn blinding code:
9675 - automatic re-creation of the BN_BLINDING parameters after
9676 a fixed number of uses (currently 32)
9677 - add new function for parameter creation
9678 - introduce flags to control the update behaviour of the
9679 BN_BLINDING parameters
9680 - hide BN_BLINDING structure
9681 Add a second BN_BLINDING slot to the RSA structure to improve
9682 performance when a single RSA object is shared among several
9687 * Add support for DTLS.
9689 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9691 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9692 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9696 * Remove buggy and incomplete DH cert support from
9697 ssl/ssl_rsa.c and ssl/s3_both.c
9701 * Use SHA-1 instead of MD5 as the default digest algorithm for
9702 the apps/openssl applications.
9706 * Compile clean with "-Wall -Wmissing-prototypes
9707 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9708 DEBUG_SAFESTACK must also be set.
9712 * Change ./Configure so that certain algorithms can be disabled by default.
9713 The new counterpiece to "no-xxx" is "enable-xxx".
9715 The patented RC5 and MDC2 algorithms will now be disabled unless
9716 "enable-rc5" and "enable-mdc2", respectively, are specified.
9718 (IDEA remains enabled despite being patented. This is because IDEA
9719 is frequently required for interoperability, and there is no license
9720 fee for non-commercial use. As before, "no-idea" can be used to
9721 avoid this algorithm.)
9725 * Add processing of proxy certificates (see RFC 3820). This work was
9726 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9727 EGEE (Enabling Grids for E-science in Europe).
9731 * RC4 performance overhaul on modern architectures/implementations, such
9732 as Intel P4, IA-64 and AMD64.
9736 * New utility extract-section.pl. This can be used specify an alternative
9737 section number in a pod file instead of having to treat each file as
9738 a separate case in Makefile. This can be done by adding two lines to the
9741 =for comment openssl_section:XXX
9743 The blank line is mandatory.
9747 * New arguments -certform, -keyform and -pass for s_client and s_server
9748 to allow alternative format key and certificate files and passphrase
9753 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9754 update associated structures and add various utility functions.
9756 Add new policy related verify parameters, include policy checking in
9757 standard verify code. Enhance 'smime' application with extra parameters
9758 to support policy checking and print out.
9762 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9763 Nehemiah processors. These extensions support AES encryption in hardware
9764 as well as RNG (though RNG support is currently disabled).
9766 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9768 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9772 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9774 *Andy Polyakov and a number of other people*
9776 * Improved PowerPC platform support. Most notably BIGNUM assembler
9777 implementation contributed by IBM.
9779 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9781 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9782 exponent rather than 'unsigned long'. There is a corresponding change to
9783 the new 'rsa_keygen' element of the RSA_METHOD structure.
9785 *Jelte Jansen, Geoff Thorpe*
9787 * Functionality for creating the initial serial number file is now
9788 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9790 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9791 number file to 1, which is bound to cause problems. To avoid
9792 the problems while respecting compatibility between different 0.9.7
9793 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9794 CA.pl for serial number initialization. With the new release 0.9.8,
9795 we can fix the problem directly in the 'ca' utility.)
9799 * Reduced header interdependencies by declaring more opaque objects in
9800 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9801 give fewer recursive includes, which could break lazy source code - so
9802 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9803 developers should define this symbol when building and using openssl to
9804 ensure they track the recommended behaviour, interfaces, [etc], but
9805 backwards-compatible behaviour prevails when this isn't defined.
9809 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9813 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9814 This will generate a random key of the appropriate length based on the
9815 cipher context. The EVP_CIPHER can provide its own random key generation
9816 routine to support keys of a specific form. This is used in the des and
9817 3des routines to generate a key of the correct parity. Update S/MIME
9818 code to use new functions and hence generate correct parity DES keys.
9819 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9820 valid (weak or incorrect parity).
9824 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9825 as looking them up. This is useful when the verified structure may contain
9826 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9827 present unless the new PKCS7_NO_CRL flag is asserted.
9831 * Extend ASN1 oid configuration module. It now additionally accepts the
9834 shortName = some long name, 1.2.3.4
9838 * Reimplemented the BN_CTX implementation. There is now no more static
9839 limitation on the number of variables it can handle nor the depth of the
9840 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9841 information can now expand as required, and rather than having a single
9842 static array of bignums, BN_CTX now uses a linked-list of such arrays
9843 allowing it to expand on demand whilst maintaining the usefulness of
9844 BN_CTX's "bundling".
9848 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9849 to allow all RSA operations to function using a single BN_CTX.
9853 * Preliminary support for certificate policy evaluation and checking. This
9854 is initially intended to pass the tests outlined in "Conformance Testing
9855 of Relying Party Client Certificate Path Processing Logic" v1.07.
9859 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9860 remained unused and not that useful. A variety of other little bignum
9861 tweaks and fixes have also been made continuing on from the audit (see
9866 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9867 associated ASN1, EVP and SSL functions and old ASN1 macros.
9871 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9872 and this should never fail. So the return value from the use of
9873 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9874 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9878 * BN_CTX_get() should return zero-valued bignums, providing the same
9879 initialised value as BN_new().
9881 *Geoff Thorpe, suggested by Ulf Möller*
9883 * Support for inhibitAnyPolicy certificate extension.
9887 * An audit of the BIGNUM code is underway, for which debugging code is
9888 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9889 is considered valid when processing BIGNUMs, and causes execution to
9890 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9891 further steps are taken to deliberately pollute unused data in BIGNUM
9892 structures to try and expose faulty code further on. For now, openssl will
9893 (in its default mode of operation) continue to tolerate the inconsistent
9894 forms that it has tolerated in the past, but authors and packagers should
9895 consider trying openssl and their own applications when compiled with
9896 these debugging symbols defined. It will help highlight potential bugs in
9897 their own code, and will improve the test coverage for OpenSSL itself. At
9898 some point, these tighter rules will become openssl's default to improve
9899 maintainability, though the assert()s and other overheads will remain only
9900 in debugging configurations. See bn.h for more details.
9902 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9904 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9905 that can only be obtained through BN_CTX_new() (which implicitly
9906 initialises it). The presence of this function only made it possible
9907 to overwrite an existing structure (and cause memory leaks).
9911 * Because of the callback-based approach for implementing LHASH as a
9912 template type, lh_insert() adds opaque objects to hash-tables and
9913 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9914 to clean up those corresponding objects before destroying the hash table
9915 (and losing the object pointers). So some over-zealous constifications in
9916 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9917 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9918 prototyped to have "const" restrictions on the object pointers they are
9919 given (and so aren't required to cast them away any more).
9923 * The tmdiff.h API was so ugly and minimal that our own timing utility
9924 (speed) prefers to use its own implementation. The two implementations
9925 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9926 its object type properly exposed (MS_TM) instead of casting to/from
9927 `char *`. This may still change yet if someone realises MS_TM and
9929 aren't necessarily the greatest nomenclatures - but this is what was used
9930 internally to the implementation so I've used that for now.
9934 * Ensure that deprecated functions do not get compiled when
9935 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9936 the self-tests were still using deprecated key-generation functions so
9937 these have been updated also.
9941 * Reorganise PKCS#7 code to separate the digest location functionality
9942 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9943 New function PKCS7_set_digest() to set the digest type for PKCS#7
9944 digestedData type. Add additional code to correctly generate the
9945 digestedData type and add support for this type in PKCS7 initialization
9950 * New function PKCS7_set0_type_other() this initializes a PKCS7
9951 structure of type "other".
9955 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9956 sure the loop does correctly stop and breaking ("division by zero")
9957 modulus operations are not performed. The (pre-generated) prime
9958 table crypto/bn/bn_prime.h was already correct, but it could not be
9959 re-generated on some platforms because of the "division by zero"
9960 situation in the script.
9962 *Ralf S. Engelschall*
9964 * Update support for ECC-based TLS ciphersuites according to
9965 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9966 SHA-1 now is only used for "small" curves (where the
9967 representation of a field element takes up to 24 bytes); for
9968 larger curves, the field element resulting from ECDH is directly
9969 used as premaster secret.
9971 *Douglas Stebila (Sun Microsystems Laboratories)*
9973 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9974 curve secp160r1 to the tests.
9976 *Douglas Stebila (Sun Microsystems Laboratories)*
9978 * Add the possibility to load symbols globally with DSO.
9980 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9982 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9983 control of the error stack.
9987 * Add support for STORE in ENGINE.
9991 * Add the STORE type. The intention is to provide a common interface
9992 to certificate and key stores, be they simple file-based stores, or
9993 HSM-type store, or LDAP stores, or...
9994 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9998 * Add a generic structure called OPENSSL_ITEM. This can be used to
9999 pass a list of arguments to any function as well as provide a way
10000 for a function to pass data back to the caller.
10004 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10005 works like BUF_strdup() but can be used to duplicate a portion of
10006 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10011 * Add the function sk_find_ex() which works like sk_find(), but will
10012 return an index to an element even if an exact match couldn't be
10013 found. The index is guaranteed to point at the element where the
10014 searched-for key would be inserted to preserve sorting order.
10018 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10019 takes an extra flags argument for optional functionality. Currently,
10020 the following flags are defined:
10022 OBJ_BSEARCH_VALUE_ON_NOMATCH
10023 This one gets OBJ_bsearch_ex() to return a pointer to the first
10024 element where the comparing function returns a negative or zero
10027 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10028 This one gets OBJ_bsearch_ex() to return a pointer to the first
10029 element where the comparing function returns zero. This is useful
10030 if there are more than one element where the comparing function
10035 * Make it possible to create self-signed certificates with 'openssl ca'
10036 in such a way that the self-signed certificate becomes part of the
10037 CA database and uses the same mechanisms for serial number generation
10038 as all other certificate signing. The new flag '-selfsign' enables
10039 this functionality. Adapt CA.sh and CA.pl.in.
10043 * Add functionality to check the public key of a certificate request
10044 against a given private. This is useful to check that a certificate
10045 request can be signed by that key (self-signing).
10049 * Make it possible to have multiple active certificates with the same
10050 subject in the CA index file. This is done only if the keyword
10051 'unique_subject' is set to 'no' in the main CA section (default
10052 if 'CA_default') of the configuration file. The value is saved
10053 with the database itself in a separate index attribute file,
10054 named like the index file with '.attr' appended to the name.
10058 * Generate multi-valued AVAs using '+' notation in config files for
10063 * Support for nameConstraints certificate extension.
10067 * Support for policyConstraints certificate extension.
10071 * Support for policyMappings certificate extension.
10075 * Make sure the default DSA_METHOD implementation only uses its
10076 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10077 and change its own handlers to be NULL so as to remove unnecessary
10078 indirection. This lets alternative implementations fallback to the
10079 default implementation more easily.
10083 * Support for directoryName in GeneralName related extensions
10088 * Make it possible to link applications using Makefile.shared.
10089 Make that possible even when linking against static libraries!
10093 * Support for single pass processing for S/MIME signing. This now
10094 means that S/MIME signing can be done from a pipe, in addition
10095 cleartext signing (multipart/signed type) is effectively streaming
10096 and the signed data does not need to be all held in memory.
10098 This is done with a new flag PKCS7_STREAM. When this flag is set
10099 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10100 is done after the data is output (and digests calculated) in
10101 SMIME_write_PKCS7().
10105 * Add full support for -rpath/-R, both in shared libraries and
10106 applications, at least on the platforms where it's known how
10111 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10112 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10113 will now compute a table of multiples of the generator that
10114 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10115 faster (notably in the case of a single point multiplication,
10116 scalar * generator).
10118 *Nils Larsch, Bodo Moeller*
10120 * IPv6 support for certificate extensions. The various extensions
10121 which use the IP:a.b.c.d can now take IPv6 addresses using the
10122 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10127 * Added an ENGINE that implements RSA by performing private key
10128 exponentiations with the GMP library. The conversions to and from
10129 GMP's mpz_t format aren't optimised nor are any montgomery forms
10130 cached, and on x86 it appears OpenSSL's own performance has caught up.
10131 However there are likely to be other architectures where GMP could
10132 provide a boost. This ENGINE is not built in by default, but it can be
10133 specified at Configure time and should be accompanied by the necessary
10134 linker additions, eg;
10135 ./config -DOPENSSL_USE_GMP -lgmp
10139 * "openssl engine" will not display ENGINE/DSO load failure errors when
10140 testing availability of engines with "-t" - the old behaviour is
10141 produced by increasing the feature's verbosity with "-tt".
10145 * ECDSA routines: under certain error conditions uninitialized BN objects
10146 could be freed. Solution: make sure initialization is performed early
10147 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10152 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10153 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10154 software implementations. For DSA and DH, parameter generation can
10155 also be overridden by providing the appropriate method callbacks.
10159 * Change the "progress" mechanism used in key-generation and
10160 primality testing to functions that take a new BN_GENCB pointer in
10161 place of callback/argument pairs. The new API functions have `_ex`
10162 postfixes and the older functions are reimplemented as wrappers for
10163 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10164 declarations of the old functions to help (graceful) attempts to
10165 migrate to the new functions. Also, the new key-generation API
10166 functions operate on a caller-supplied key-structure and return
10167 success/failure rather than returning a key or NULL - this is to
10168 help make "keygen" another member function of RSA_METHOD etc.
10170 Example for using the new callback interface:
10172 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10173 void *my_arg = ...;
10176 BN_GENCB_set(&my_cb, my_callback, my_arg);
10178 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10179 /* For the meaning of a, b in calls to my_callback(), see the
10180 * documentation of the function that calls the callback.
10181 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10182 * my_callback should return 1 if it wants BN_is_prime_ex()
10183 * to continue, or 0 to stop.
10188 * Change the ZLIB compression method to be stateful, and make it
10189 available to TLS with the number defined in
10190 draft-ietf-tls-compression-04.txt.
10194 * Add the ASN.1 structures and functions for CertificatePair, which
10195 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10197 CertificatePair ::= SEQUENCE {
10198 forward [0] Certificate OPTIONAL,
10199 reverse [1] Certificate OPTIONAL,
10200 -- at least one of the pair shall be present -- }
10202 Also implement the PEM functions to read and write certificate
10203 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10205 This needed to be defined, mostly for the sake of the LDAP
10206 attribute crossCertificatePair, but may prove useful elsewhere as
10211 * Make it possible to inhibit symlinking of shared libraries in
10212 Makefile.shared, for Cygwin's sake.
10216 * Extend the BIGNUM API by creating a function
10217 void BN_set_negative(BIGNUM *a, int neg);
10218 and a macro that behave like
10219 int BN_is_negative(const BIGNUM *a);
10221 to avoid the need to access 'a->neg' directly in applications.
10225 * Implement fast modular reduction for pseudo-Mersenne primes
10226 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10227 EC_GROUP_new_curve_GFp() will now automatically use this
10230 *Nils Larsch <nla@trustcenter.de>*
10232 * Add new lock type (CRYPTO_LOCK_BN).
10236 * Change the ENGINE framework to automatically load engines
10237 dynamically from specific directories unless they could be
10238 found to already be built in or loaded. Move all the
10239 current engines except for the cryptodev one to a new
10240 directory engines/.
10241 The engines in engines/ are built as shared libraries if
10242 the "shared" options was given to ./Configure or ./config.
10243 Otherwise, they are inserted in libcrypto.a.
10244 /usr/local/ssl/engines is the default directory for dynamic
10245 engines, but that can be overridden at configure time through
10246 the usual use of --prefix and/or --openssldir, and at run
10247 time with the environment variable OPENSSL_ENGINES.
10249 *Geoff Thorpe and Richard Levitte*
10251 * Add Makefile.shared, a helper makefile to build shared
10252 libraries. Adapt Makefile.org.
10256 * Add version info to Win32 DLLs.
10258 *Peter 'Luna' Runestig" <peter@runestig.com>*
10260 * Add new 'medium level' PKCS#12 API. Certificates and keys
10261 can be added using this API to created arbitrary PKCS#12
10262 files while avoiding the low level API.
10264 New options to PKCS12_create(), key or cert can be NULL and
10265 will then be omitted from the output file. The encryption
10266 algorithm NIDs can be set to -1 for no encryption, the mac
10267 iteration count can be set to 0 to omit the mac.
10269 Enhance pkcs12 utility by making the -nokeys and -nocerts
10270 options work when creating a PKCS#12 file. New option -nomac
10271 to omit the mac, NONE can be set for an encryption algorithm.
10272 New code is modified to use the enhanced PKCS12_create()
10273 instead of the low level API.
10277 * Extend ASN1 encoder to support indefinite length constructed
10278 encoding. This can output sequences tags and octet strings in
10279 this form. Modify pk7_asn1.c to support indefinite length
10280 encoding. This is experimental and needs additional code to
10281 be useful, such as an ASN1 bio and some enhanced streaming
10284 Extend template encode functionality so that tagging is passed
10285 down to the template encoder.
10289 * Let 'openssl req' fail if an argument to '-newkey' is not
10290 recognized instead of using RSA as a default.
10294 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10295 As these are not official, they are not included in "ALL";
10296 the "ECCdraft" ciphersuite group alias can be used to select them.
10298 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10300 * Add ECDH engine support.
10302 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10304 * Add ECDH in new directory crypto/ecdh/.
10306 *Douglas Stebila (Sun Microsystems Laboratories)*
10308 * Let BN_rand_range() abort with an error after 100 iterations
10309 without success (which indicates a broken PRNG).
10313 * Change BN_mod_sqrt() so that it verifies that the input value
10314 is really the square of the return value. (Previously,
10315 BN_mod_sqrt would show GIGO behaviour.)
10319 * Add named elliptic curves over binary fields from X9.62, SECG,
10320 and WAP/WTLS; add OIDs that were still missing.
10322 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10324 * Extend the EC library for elliptic curves over binary fields
10325 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10328 EC_GF2m_simple_method
10332 EC_GROUP_new_curve_GF2m
10333 EC_GROUP_set_curve_GF2m
10334 EC_GROUP_get_curve_GF2m
10335 EC_POINT_set_affine_coordinates_GF2m
10336 EC_POINT_get_affine_coordinates_GF2m
10337 EC_POINT_set_compressed_coordinates_GF2m
10339 Point compression for binary fields is disabled by default for
10340 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10343 As binary polynomials are represented as BIGNUMs, various members
10344 of the EC_GROUP and EC_POINT data structures can be shared
10345 between the implementations for prime fields and binary fields;
10346 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10347 are essentially identical to their `..._GFp` counterparts.
10348 (For simplicity, the `..._GFp` prefix has been dropped from
10349 various internal method names.)
10351 An internal 'field_div' method (similar to 'field_mul' and
10352 'field_sqr') has been added; this is used only for binary fields.
10354 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10356 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10357 through methods ('mul', 'precompute_mult').
10359 The generic implementations (now internally called 'ec_wNAF_mul'
10360 and 'ec_wNAF_precomputed_mult') remain the default if these
10361 methods are undefined.
10363 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10365 * New function EC_GROUP_get_degree, which is defined through
10366 EC_METHOD. For curves over prime fields, this returns the bit
10367 length of the modulus.
10369 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10371 * New functions EC_GROUP_dup, EC_POINT_dup.
10372 (These simply call ..._new and ..._copy).
10374 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10376 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10377 Polynomials are represented as BIGNUMs (where the sign bit is not
10378 used) in the following functions [macros]:
10381 BN_GF2m_sub [= BN_GF2m_add]
10382 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10383 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10384 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10386 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10387 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10388 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10389 BN_GF2m_cmp [= BN_ucmp]
10391 (Note that only the 'mod' functions are actually for fields GF(2^m).
10392 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10394 For some functions, an the irreducible polynomial defining a
10395 field can be given as an 'unsigned int[]' with strictly
10396 decreasing elements giving the indices of those bits that are set;
10397 i.e., p[] represents the polynomial
10398 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10400 p[0] > p[1] > ... > p[k] = 0.
10401 This applies to the following functions:
10404 BN_GF2m_mod_mul_arr
10405 BN_GF2m_mod_sqr_arr
10406 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10407 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10408 BN_GF2m_mod_exp_arr
10409 BN_GF2m_mod_sqrt_arr
10410 BN_GF2m_mod_solve_quad_arr
10414 Conversion can be performed by the following functions:
10419 bntest.c has additional tests for binary polynomial arithmetic.
10421 Two implementations for BN_GF2m_mod_div() are available.
10422 The default algorithm simply uses BN_GF2m_mod_inv() and
10423 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10424 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10425 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10427 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10429 * Add new error code 'ERR_R_DISABLED' that can be used when some
10430 functionality is disabled at compile-time.
10432 *Douglas Stebila <douglas.stebila@sun.com>*
10434 * Change default behaviour of 'openssl asn1parse' so that more
10435 information is visible when viewing, e.g., a certificate:
10437 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10438 mode the content of non-printable OCTET STRINGs is output in a
10439 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10440 avoid the appearance of a printable string.
10442 *Nils Larsch <nla@trustcenter.de>*
10444 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10446 EC_GROUP_set_asn1_flag()
10447 EC_GROUP_get_asn1_flag()
10448 EC_GROUP_set_point_conversion_form()
10449 EC_GROUP_get_point_conversion_form()
10450 These control ASN1 encoding details:
10451 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10452 has been set to OPENSSL_EC_NAMED_CURVE.
10453 - Points are encoded in uncompressed form by default; options for
10454 asn1_for are as for point2oct, namely
10455 POINT_CONVERSION_COMPRESSED
10456 POINT_CONVERSION_UNCOMPRESSED
10457 POINT_CONVERSION_HYBRID
10459 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10461 EC_GROUP_set_seed()
10462 EC_GROUP_get0_seed()
10463 EC_GROUP_get_seed_len()
10464 This is used only for ASN1 purposes (so far).
10466 *Nils Larsch <nla@trustcenter.de>*
10468 * Add 'field_type' member to EC_METHOD, which holds the NID
10469 of the appropriate field type OID. The new function
10470 EC_METHOD_get_field_type() returns this value.
10472 *Nils Larsch <nla@trustcenter.de>*
10475 EC_POINT_point2bn()
10476 EC_POINT_bn2point()
10477 EC_POINT_point2hex()
10478 EC_POINT_hex2point()
10479 providing useful interfaces to EC_POINT_point2oct() and
10480 EC_POINT_oct2point().
10482 *Nils Larsch <nla@trustcenter.de>*
10484 * Change internals of the EC library so that the functions
10485 EC_GROUP_set_generator()
10486 EC_GROUP_get_generator()
10487 EC_GROUP_get_order()
10488 EC_GROUP_get_cofactor()
10489 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10490 to methods, which would lead to unnecessary code duplication when
10491 adding different types of curves.
10493 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10495 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10496 arithmetic, and such that modified wNAFs are generated
10497 (which avoid length expansion in many cases).
10501 * Add a function EC_GROUP_check_discriminant() (defined via
10502 EC_METHOD) that verifies that the curve discriminant is non-zero.
10504 Add a function EC_GROUP_check() that makes some sanity tests
10505 on a EC_GROUP, its generator and order. This includes
10506 EC_GROUP_check_discriminant().
10508 *Nils Larsch <nla@trustcenter.de>*
10510 * Add ECDSA in new directory crypto/ecdsa/.
10512 Add applications 'openssl ecparam' and 'openssl ecdsa'
10513 (these are based on 'openssl dsaparam' and 'openssl dsa').
10515 ECDSA support is also included in various other files across the
10516 library. Most notably,
10517 - 'openssl req' now has a '-newkey ecdsa:file' option;
10518 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10519 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10520 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10521 them suitable for ECDSA where domain parameters must be
10522 extracted before the specific public key;
10523 - ECDSA engine support has been added.
10525 *Nils Larsch <nla@trustcenter.de>*
10527 * Include some named elliptic curves, and add OIDs from X9.62,
10528 SECG, and WAP/WTLS. Each curve can be obtained from the new
10530 EC_GROUP_new_by_curve_name(),
10531 and the list of available named curves can be obtained with
10532 EC_get_builtin_curves().
10533 Also add a 'curve_name' member to EC_GROUP objects, which can be
10535 EC_GROUP_set_curve_name()
10536 EC_GROUP_get_curve_name()
10538 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10540 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10541 was actually never needed) and in BN_mul(). The removal in BN_mul()
10542 required a small change in bn_mul_part_recursive() and the addition
10543 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10544 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10545 bn_sub_words() and bn_add_words() except they take arrays with
10550 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10552 * Cleanse PEM buffers before freeing them since they may contain
10555 *Benjamin Bennett <ben@psc.edu>*
10557 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10558 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10559 authentication-only ciphersuites.
10563 * Since AES128 and AES256 share a single mask bit in the logic of
10564 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10565 kludge to work properly if AES128 is available and AES256 isn't.
10569 * Expand security boundary to match 1.1.1 module.
10573 * Remove redundant features: hash file source, editing of test vectors
10574 modify fipsld to use external fips_premain.c signature.
10578 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10579 run algorithm test programs.
10583 * Make algorithm test programs more tolerant of whitespace.
10587 * Have SSL/TLS server implementation tolerate "mismatched" record
10588 protocol version while receiving ClientHello even if the
10589 ClientHello is fragmented. (The server can't insist on the
10590 particular protocol version it has chosen before the ServerHello
10591 message has informed the client about his choice.)
10595 * Load error codes if they are not already present instead of using a
10596 static variable. This allows them to be cleanly unloaded and reloaded.
10600 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10602 * Introduce limits to prevent malicious keys being able to
10603 cause a denial of service. [CVE-2006-2940][]
10605 *Steve Henson, Bodo Moeller*
10607 * Fix ASN.1 parsing of certain invalid structures that can result
10608 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10610 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10611 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10613 * Fix SSL client code which could crash if connecting to a
10614 malicious SSLv2 server. [CVE-2006-4343][]
10616 *Tavis Ormandy and Will Drewry, Google Security Team*
10618 * Change ciphersuite string processing so that an explicit
10619 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10620 will no longer include "AES128-SHA"), and any other similar
10621 ciphersuite (same bitmap) from *other* protocol versions (so that
10622 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10623 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10624 changes from 0.9.8b and 0.9.8d.
10628 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10630 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10631 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10633 * Change the Unix randomness entropy gathering to use poll() when
10634 possible instead of select(), since the latter has some
10635 undesirable limitations.
10637 *Darryl Miles via Richard Levitte and Bodo Moeller*
10639 * Disable rogue ciphersuites:
10641 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10642 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10643 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10645 The latter two were purportedly from
10646 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10649 Also deactivate the remaining ciphersuites from
10650 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10651 unofficial, and the ID has long expired.
10655 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10656 dual-core machines) and other potential thread-safety issues.
10660 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10662 * Adapt fipsld and the build system to link against the validated FIPS
10663 module in FIPS mode.
10667 * Fixes for VC++ 2005 build under Windows.
10671 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10672 from a Windows bash shell such as MSYS. It is autodetected from the
10673 "config" script when run from a VC++ environment. Modify standard VC++
10674 build to use fipscanister.o from the GNU make build.
10678 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10680 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10681 The value now differs depending on if you build for FIPS or not.
10682 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10683 safely run with a non-FIPSed libcrypto, as it may crash because of
10684 the difference induced by this change.
10688 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10690 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10691 (part of SSL_OP_ALL). This option used to disable the
10692 countermeasure against man-in-the-middle protocol-version
10693 rollback in the SSL 2.0 server implementation, which is a bad
10694 idea. [CVE-2005-2969][]
10696 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10697 for Information Security, National Institute of Advanced Industrial
10698 Science and Technology [AIST, Japan)]*
10700 * Minimal support for X9.31 signatures and PSS padding modes. This is
10701 mainly for FIPS compliance and not fully integrated at this stage.
10705 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10706 the exponentiation using a fixed-length exponent. (Otherwise,
10707 the information leaked through timing could expose the secret key
10708 after many signatures; cf. Bleichenbacher's attack on DSA with
10713 * Make a new fixed-window mod_exp implementation the default for
10714 RSA, DSA, and DH private-key operations so that the sequence of
10715 squares and multiplies and the memory access pattern are
10716 independent of the particular secret key. This will mitigate
10717 cache-timing and potential related attacks.
10719 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10720 and this is automatically used by BN_mod_exp_mont() if the new flag
10721 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10722 will use this BN flag for private exponents unless the flag
10723 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10724 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10726 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10728 * Change the client implementation for SSLv23_method() and
10729 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10730 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10731 (Previously, the SSL 2.0 backwards compatible Client Hello
10732 message format would be used even with SSL_OP_NO_SSLv2.)
10736 * Add support for smime-type MIME parameter in S/MIME messages which some
10741 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10742 a threadsafe manner. Modify rsa code to use new function and add calls
10743 to dsa and dh code (which had race conditions before).
10747 * Include the fixed error library code in the C error file definitions
10748 instead of fixing them up at runtime. This keeps the error code
10749 structures constant.
10753 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10755 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10758 * Fixes for newer kerberos headers. NB: the casts are needed because
10759 the 'length' field is signed on one version and unsigned on another
10760 with no (?) obvious way to tell the difference, without these VC++
10761 complains. Also the "definition" of FAR (blank) is no longer included
10762 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10763 some needed definitions.
10767 * Undo Cygwin change.
10771 * Added support for proxy certificates according to RFC 3820.
10772 Because they may be a security thread to unaware applications,
10773 they must be explicitly allowed in run-time. See
10774 docs/HOWTO/proxy_certificates.txt for further information.
10778 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10780 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10781 server and client random values. Previously
10782 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10783 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10785 This change has negligible security impact because:
10787 1. Server and client random values still have 24 bytes of pseudo random
10790 2. Server and client random values are sent in the clear in the initial
10793 3. The master secret is derived using the premaster secret (48 bytes in
10794 size for static RSA ciphersuites) as well as client server and random
10797 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10800 *Stephen Henson, reported by UK NISCC*
10802 * Use Windows randomness collection on Cygwin.
10806 * Fix hang in EGD/PRNGD query when communication socket is closed
10807 prematurely by EGD/PRNGD.
10809 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10811 * Prompt for pass phrases when appropriate for PKCS12 input format.
10815 * Back-port of selected performance improvements from development
10816 branch, as well as improved support for PowerPC platforms.
10820 * Add lots of checks for memory allocation failure, error codes to indicate
10821 failure and freeing up memory if a failure occurs.
10823 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10825 * Add new -passin argument to dgst.
10829 * Perform some character comparisons of different types in X509_NAME_cmp:
10830 this is needed for some certificates that re-encode DNs into UTF8Strings
10831 (in violation of RFC3280) and can't or won't issue name rollover
10836 * Make an explicit check during certificate validation to see that
10837 the CA setting in each certificate on the chain is correct. As a
10838 side effect always do the following basic checks on extensions,
10839 not just when there's an associated purpose to the check:
10841 - if there is an unhandled critical extension (unless the user
10842 has chosen to ignore this fault)
10843 - if the path length has been exceeded (if one is set at all)
10844 - that certain extensions fit the associated purpose (if one has
10849 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10851 * Avoid a race condition when CRLs are checked in a multi threaded
10852 environment. This would happen due to the reordering of the revoked
10853 entries during signature checking and serial number lookup. Now the
10854 encoding is cached and the serial number sort performed under a lock.
10855 Add new STACK function sk_is_sorted().
10859 * Add Delta CRL to the extension code.
10863 * Various fixes to s3_pkt.c so alerts are sent properly.
10865 *David Holmes <d.holmes@f5.com>*
10867 * Reduce the chances of duplicate issuer name and serial numbers (in
10868 violation of RFC3280) using the OpenSSL certificate creation utilities.
10869 This is done by creating a random 64 bit value for the initial serial
10870 number when a serial number file is created or when a self signed
10871 certificate is created using 'openssl req -x509'. The initial serial
10872 number file is created using 'openssl x509 -next_serial' in CA.pl
10873 rather than being initialized to 1.
10877 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10879 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10880 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10882 *Joe Orton, Steve Henson*
10884 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10887 *Joe Orton, Steve Henson*
10889 * Make it possible to have multiple active certificates with the same
10890 subject in the CA index file. This is done only if the keyword
10891 'unique_subject' is set to 'no' in the main CA section (default
10892 if 'CA_default') of the configuration file. The value is saved
10893 with the database itself in a separate index attribute file,
10894 named like the index file with '.attr' appended to the name.
10898 * X509 verify fixes. Disable broken certificate workarounds when
10899 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10900 keyUsage extension present. Don't accept CRLs with unhandled critical
10901 extensions: since verify currently doesn't process CRL extensions this
10902 rejects a CRL with *any* critical extensions. Add new verify error codes
10907 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10908 A clarification of RFC2560 will require the use of OCTET STRINGs and
10909 some implementations cannot handle the current raw format. Since OpenSSL
10910 copies and compares OCSP nonces as opaque blobs without any attempt at
10911 parsing them this should not create any compatibility issues.
10915 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10916 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10917 this HMAC (and other) operations are several times slower than OpenSSL
10922 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10924 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10926 * Use the correct content when signing type "other".
10930 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10932 * Fix various bugs revealed by running the NISCC test suite:
10934 Stop out of bounds reads in the ASN1 code when presented with
10935 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10937 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10939 If verify callback ignores invalid public key errors don't try to check
10940 certificate signature with the NULL public key.
10944 * New -ignore_err option in ocsp application to stop the server
10945 exiting on the first error in a request.
10949 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10950 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10955 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10956 extra data after the compression methods not only for TLS 1.0
10957 but also for SSL 3.0 (as required by the specification).
10959 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10961 * Change X509_certificate_type() to mark the key as exported/exportable
10962 when it's 512 *bits* long, not 512 bytes.
10966 * Change AES_cbc_encrypt() so it outputs exact multiple of
10967 blocks during encryption.
10971 * Various fixes to base64 BIO and non blocking I/O. On write
10972 flushes were not handled properly if the BIO retried. On read
10973 data was not being buffered properly and had various logic bugs.
10974 This also affects blocking I/O when the data being decoded is a
10979 * Various S/MIME bugfixes and compatibility changes:
10980 output correct application/pkcs7 MIME type if
10981 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10982 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10983 of files as .eml work). Correctly handle very long lines in MIME
10988 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10990 * Countermeasure against the Klima-Pokorny-Rosa extension of
10991 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10992 a protocol version number mismatch like a decryption error
10993 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10997 * Turn on RSA blinding by default in the default implementation
10998 to avoid a timing attack. Applications that don't want it can call
10999 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11000 They would be ill-advised to do so in most cases.
11002 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11004 * Change RSA blinding code so that it works when the PRNG is not
11005 seeded (in this case, the secret RSA exponent is abused as
11006 an unpredictable seed -- if it is not unpredictable, there
11007 is no point in blinding anyway). Make RSA blinding thread-safe
11008 by remembering the creator's thread ID in rsa->blinding and
11009 having all other threads use local one-time blinding factors
11010 (this requires more computation than sharing rsa->blinding, but
11011 avoids excessive locking; and if an RSA object is not shared
11012 between threads, blinding will still be very fast).
11016 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11017 ENGINE as defaults for all supported algorithms irrespective of
11018 the 'flags' parameter. 'flags' is now honoured, so applications
11019 should make sure they are passing it correctly.
11023 * Target "mingw" now allows native Windows code to be generated in
11024 the Cygwin environment as well as with the MinGW compiler.
11028 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11030 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11031 via timing by performing a MAC computation even if incorrect
11032 block cipher padding has been found. This is a countermeasure
11033 against active attacks where the attacker has to distinguish
11034 between bad padding and a MAC verification error. [CVE-2003-0078][]
11036 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11037 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11038 Martin Vuagnoux (EPFL, Ilion)*
11040 * Make the no-err option work as intended. The intention with no-err
11041 is not to have the whole error stack handling routines removed from
11042 libcrypto, it's only intended to remove all the function name and
11043 reason texts, thereby removing some of the footprint that may not
11044 be interesting if those errors aren't displayed anyway.
11046 NOTE: it's still possible for any application or module to have its
11047 own set of error texts inserted. The routines are there, just not
11048 used by default when no-err is given.
11052 * Add support for FreeBSD on IA64.
11054 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11056 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11057 Kerberos function mit_des_cbc_cksum(). Before this change,
11058 the value returned by DES_cbc_cksum() was like the one from
11059 mit_des_cbc_cksum(), except the bytes were swapped.
11061 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11063 * Allow an application to disable the automatic SSL chain building.
11064 Before this a rather primitive chain build was always performed in
11065 ssl3_output_cert_chain(): an application had no way to send the
11066 correct chain if the automatic operation produced an incorrect result.
11068 Now the chain builder is disabled if either:
11070 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11072 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11074 The reasoning behind this is that an application would not want the
11075 auto chain building to take place if extra chain certificates are
11076 present and it might also want a means of sending no additional
11077 certificates (for example the chain has two certificates and the
11082 * Add the possibility to build without the ENGINE framework.
11084 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11086 * Under Win32 gmtime() can return NULL: check return value in
11087 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11091 * DSA routines: under certain error conditions uninitialized BN objects
11092 could be freed. Solution: make sure initialization is performed early
11093 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11094 Nils Larsch <nla@trustcenter.de> via PR#459)
11098 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11099 checked on reconnect on the client side, therefore session resumption
11100 could still fail with a "ssl session id is different" error. This
11101 behaviour is masked when SSL_OP_ALL is used due to
11102 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11103 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11104 followup to PR #377.
11108 * IA-32 assembler support enhancements: unified ELF targets, support
11109 for SCO/Caldera platforms, fix for Cygwin shared build.
11113 * Add support for FreeBSD on sparc64. As a consequence, support for
11114 FreeBSD on non-x86 processors is separate from x86 processors on
11115 the config script, much like the NetBSD support.
11117 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11119 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11121 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11124 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11125 code (06) was taken as the first octet of the session ID and the last
11126 octet was ignored consequently. As a result SSLv2 client side session
11127 caching could not have worked due to the session ID mismatch between
11129 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11134 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11135 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11140 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11141 seems that in spite of existing for more than a year, many application
11142 author have done nothing to provide the necessary callbacks, which
11143 means that this particular engine will not work properly anywhere.
11144 This is a very unfortunate situation which forces us, in the name
11145 of usability, to give the hw_ncipher.c a static lock, which is part
11147 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11148 appear in 0.9.8 or later. We EXPECT application authors to have
11149 dealt properly with this when 0.9.8 is released (unless we actually
11150 make such changes in the libcrypto locking code that changes will
11151 have to be made anyway).
11155 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11156 octets have been read, EOF or an error occurs. Without this change
11157 some truncated ASN1 structures will not produce an error.
11161 * Disable Heimdal support, since it hasn't been fully implemented.
11162 Still give the possibility to force the use of Heimdal, but with
11163 warnings and a request that patches get sent to openssl-dev.
11167 * Add the VC-CE target, introduce the WINCE sysname, and add
11168 INSTALL.WCE and appropriate conditionals to make it build.
11170 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11172 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11173 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11174 edit numbers of the version.
11176 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11178 * Introduce safe string copy and catenation functions
11179 (BUF_strlcpy() and BUF_strlcat()).
11181 *Ben Laurie (CHATS) and Richard Levitte*
11183 * Avoid using fixed-size buffers for one-line DNs.
11185 *Ben Laurie (CHATS)*
11187 * Add BUF_MEM_grow_clean() to avoid information leakage when
11188 resizing buffers containing secrets, and use where appropriate.
11190 *Ben Laurie (CHATS)*
11192 * Avoid using fixed size buffers for configuration file location.
11194 *Ben Laurie (CHATS)*
11196 * Avoid filename truncation for various CA files.
11198 *Ben Laurie (CHATS)*
11200 * Use sizeof in preference to magic numbers.
11202 *Ben Laurie (CHATS)*
11204 * Avoid filename truncation in cert requests.
11206 *Ben Laurie (CHATS)*
11208 * Add assertions to check for (supposedly impossible) buffer
11211 *Ben Laurie (CHATS)*
11213 * Don't cache truncated DNS entries in the local cache (this could
11214 potentially lead to a spoofing attack).
11216 *Ben Laurie (CHATS)*
11218 * Fix various buffers to be large enough for hex/decimal
11219 representations in a platform independent manner.
11221 *Ben Laurie (CHATS)*
11223 * Add CRYPTO_realloc_clean() to avoid information leakage when
11224 resizing buffers containing secrets, and use where appropriate.
11226 *Ben Laurie (CHATS)*
11228 * Add BIO_indent() to avoid much slightly worrying code to do
11231 *Ben Laurie (CHATS)*
11233 * Convert sprintf()/BIO_puts() to BIO_printf().
11235 *Ben Laurie (CHATS)*
11237 * buffer_gets() could terminate with the buffer only half
11240 *Ben Laurie (CHATS)*
11242 * Add assertions to prevent user-supplied crypto functions from
11243 overflowing internal buffers by having large block sizes, etc.
11245 *Ben Laurie (CHATS)*
11247 * New OPENSSL_assert() macro (similar to assert(), but enabled
11250 *Ben Laurie (CHATS)*
11252 * Eliminate unused copy of key in RC4.
11254 *Ben Laurie (CHATS)*
11256 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11258 *Ben Laurie (CHATS)*
11260 * Fix off-by-one error in EGD path.
11262 *Ben Laurie (CHATS)*
11264 * If RANDFILE path is too long, ignore instead of truncating.
11266 *Ben Laurie (CHATS)*
11268 * Eliminate unused and incorrectly sized X.509 structure
11271 *Ben Laurie (CHATS)*
11273 * Eliminate unused and dangerous function knumber().
11275 *Ben Laurie (CHATS)*
11277 * Eliminate unused and dangerous structure, KSSL_ERR.
11279 *Ben Laurie (CHATS)*
11281 * Protect against overlong session ID context length in an encoded
11282 session object. Since these are local, this does not appear to be
11285 *Ben Laurie (CHATS)*
11287 * Change from security patch (see 0.9.6e below) that did not affect
11288 the 0.9.6 release series:
11290 Remote buffer overflow in SSL3 protocol - an attacker could
11291 supply an oversized master key in Kerberos-enabled versions.
11294 *Ben Laurie (CHATS)*
11296 * Change the SSL kerb5 codes to match RFC 2712.
11300 * Make -nameopt work fully for req and add -reqopt switch.
11302 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11304 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11306 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11308 * Make sure tests can be performed even if the corresponding algorithms
11309 have been removed entirely. This was also the last step to make
11310 OpenSSL compilable with DJGPP under all reasonable conditions.
11312 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11314 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11315 to allow version independent disabling of normally unselected ciphers,
11316 which may be activated as a side-effect of selecting a single cipher.
11318 (E.g., cipher list string "RSA" enables ciphersuites that are left
11319 out of "ALL" because they do not provide symmetric encryption.
11320 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11322 *Lutz Jaenicke, Bodo Moeller*
11324 * Add appropriate support for separate platform-dependent build
11325 directories. The recommended way to make a platform-dependent
11326 build directory is the following (tested on Linux), maybe with
11329 # Place yourself outside of the OpenSSL source tree. In
11330 # this example, the environment variable OPENSSL_SOURCE
11331 # is assumed to contain the absolute OpenSSL source directory.
11332 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11333 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11334 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11335 mkdir -p `dirname $F`
11336 ln -s $OPENSSL_SOURCE/$F $F
11339 To be absolutely sure not to disturb the source tree, a "make clean"
11340 is a good thing. If it isn't successful, don't worry about it,
11341 it probably means the source directory is very clean.
11345 * Make sure any ENGINE control commands make local copies of string
11346 pointers passed to them whenever necessary. Otherwise it is possible
11347 the caller may have overwritten (or deallocated) the original string
11348 data when a later ENGINE operation tries to use the stored values.
11350 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11352 * Improve diagnostics in file reading and command-line digests.
11354 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11356 * Add AES modes CFB and OFB to the object database. Correct an
11357 error in AES-CFB decryption.
11361 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11362 allows existing EVP_CIPHER_CTX structures to be reused after
11363 calling `EVP_*Final()`. This behaviour is used by encryption
11364 BIOs and some applications. This has the side effect that
11365 applications must explicitly clean up cipher contexts with
11366 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11370 * Check the values of dna and dnb in bn_mul_recursive before calling
11371 bn_mul_comba (a non zero value means the a or b arrays do not contain
11372 n2 elements) and fallback to bn_mul_normal if either is not zero.
11376 * Fix escaping of non-ASCII characters when using the -subj option
11377 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11381 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11382 form for "surname", serialNumber has no short form.
11383 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11384 therefore remove "mail" short name for "internet 7".
11385 The OID for unique identifiers in X509 certificates is
11386 x500UniqueIdentifier, not uniqueIdentifier.
11387 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11391 * Add an "init" command to the ENGINE config module and auto initialize
11392 ENGINEs. Without any "init" command the ENGINE will be initialized
11393 after all ctrl commands have been executed on it. If init=1 the
11394 ENGINE is initialized at that point (ctrls before that point are run
11395 on the uninitialized ENGINE and after on the initialized one). If
11396 init=0 then the ENGINE will not be initialized at all.
11400 * Fix the 'app_verify_callback' interface so that the user-defined
11401 argument is actually passed to the callback: In the
11402 SSL_CTX_set_cert_verify_callback() prototype, the callback
11403 declaration has been changed from
11406 int (*cb)(X509_STORE_CTX *,void *);
11407 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11408 i=s->ctx->app_verify_callback(&ctx)
11409 has been changed into
11410 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11412 To update applications using SSL_CTX_set_cert_verify_callback(),
11413 a dummy argument can be added to their callback functions.
11415 *D. K. Smetters <smetters@parc.xerox.com>*
11417 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11419 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11421 * Add and OPENSSL_LOAD_CONF define which will cause
11422 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11423 This allows older applications to transparently support certain
11424 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11425 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11426 load the config file and OPENSSL_add_all_algorithms_conf() which will
11427 always load it have also been added.
11431 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11432 Adjust NIDs and EVP layer.
11434 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11436 * Config modules support in openssl utility.
11438 Most commands now load modules from the config file,
11439 though in a few (such as version) this isn't done
11440 because it couldn't be used for anything.
11442 In the case of ca and req the config file used is
11443 the same as the utility itself: that is the -config
11444 command line option can be used to specify an
11449 * Move default behaviour from OPENSSL_config(). If appname is NULL
11450 use "openssl_conf" if filename is NULL use default openssl config file.
11454 * Add an argument to OPENSSL_config() to allow the use of an alternative
11455 config section name. Add a new flag to tolerate a missing config file
11456 and move code to CONF_modules_load_file().
11460 * Support for crypto accelerator cards from Accelerated Encryption
11461 Processing, www.aep.ie. (Use engine 'aep')
11462 The support was copied from 0.9.6c [engine] and adapted/corrected
11463 to work with the new engine framework.
11465 *AEP Inc. and Richard Levitte*
11467 * Support for SureWare crypto accelerator cards from Baltimore
11468 Technologies. (Use engine 'sureware')
11469 The support was copied from 0.9.6c [engine] and adapted
11470 to work with the new engine framework.
11474 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11475 make the newer ENGINE framework commands for the CHIL engine work.
11477 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11479 * Make it possible to produce shared libraries on ReliantUNIX.
11481 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11483 * Add the configuration target debug-linux-ppro.
11484 Make 'openssl rsa' use the general key loading routines
11485 implemented in apps.c, and make those routines able to
11486 handle the key format FORMAT_NETSCAPE and the variant
11489 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11491 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11493 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11495 * Add -keyform to rsautl, and document -engine.
11497 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11499 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11500 BIO_R_NO_SUCH_FILE error code rather than the generic
11501 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11505 * Add new functions
11506 ERR_peek_last_error
11507 ERR_peek_last_error_line
11508 ERR_peek_last_error_line_data.
11509 These are similar to
11511 ERR_peek_error_line
11512 ERR_peek_error_line_data,
11513 but report on the latest error recorded rather than the first one
11514 still in the error queue.
11516 *Ben Laurie, Bodo Moeller*
11518 * default_algorithms option in ENGINE config module. This allows things
11520 default_algorithms = ALL
11521 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11525 * Preliminary ENGINE config module.
11529 * New experimental application configuration code.
11533 * Change the AES code to follow the same name structure as all other
11534 symmetric ciphers, and behave the same way. Move everything to
11535 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11537 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11539 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11541 *Ben Laurie and Theo de Raadt*
11543 * Add option to output public keys in req command.
11545 *Massimiliano Pala madwolf@openca.org*
11547 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11548 (up to about 10% better than before for P-192 and P-224).
11552 * New functions/macros
11554 SSL_CTX_set_msg_callback(ctx, cb)
11555 SSL_CTX_set_msg_callback_arg(ctx, arg)
11556 SSL_set_msg_callback(ssl, cb)
11557 SSL_set_msg_callback_arg(ssl, arg)
11559 to request calling a callback function
11561 void cb(int write_p, int version, int content_type,
11562 const void *buf, size_t len, SSL *ssl, void *arg)
11564 whenever a protocol message has been completely received
11565 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11566 protocol version according to which the SSL library interprets
11567 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11568 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11569 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11570 specification (change_cipher_spec(20), alert(21), handshake(22)).
11571 'buf' and 'len' point to the actual message, 'ssl' to the
11572 SSL object, and 'arg' is the application-defined value set by
11573 SSL[_CTX]_set_msg_callback_arg().
11575 'openssl s_client' and 'openssl s_server' have new '-msg' options
11576 to enable a callback that displays all protocol messages.
11580 * Change the shared library support so shared libraries are built as
11581 soon as the corresponding static library is finished, and thereby get
11582 openssl and the test programs linked against the shared library.
11583 This still only happens when the keyword "shard" has been given to
11584 the configuration scripts.
11586 NOTE: shared library support is still an experimental thing, and
11587 backward binary compatibility is still not guaranteed.
11589 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11591 * Add support for Subject Information Access extension.
11593 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11595 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11596 additional bytes when new memory had to be allocated, not just
11597 when reusing an existing buffer.
11601 * New command line and configuration option 'utf8' for the req command.
11602 This allows field values to be specified as UTF8 strings.
11606 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11607 runs for the former and machine-readable output for the latter.
11611 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11612 of the e-mail address in the DN (i.e., it will go into a certificate
11613 extension only). The new configuration file option 'email_in_dn = no'
11614 has the same effect.
11616 *Massimiliano Pala madwolf@openca.org*
11618 * Change all functions with names starting with `des_` to be starting
11619 with `DES_` instead. Add wrappers that are compatible with libdes,
11620 but are named `_ossl_old_des_*`. Finally, add macros that map the
11621 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11622 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11623 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11626 Since we provide two compatibility mappings, the user needs to
11627 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11628 compatibility is desired. The default (i.e., when that macro
11629 isn't defined) is OpenSSL 0.9.6c compatibility.
11631 There are also macros that enable and disable the support of old
11632 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11633 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11634 are defined, the default will apply: to support the old des routines.
11636 In either case, one must include openssl/des.h to get the correct
11637 definitions. Do not try to just include openssl/des_old.h, that
11640 NOTE: This is a major break of an old API into a new one. Software
11641 authors are encouraged to switch to the `DES_` style functions. Some
11642 time in the future, des_old.h and the libdes compatibility functions
11643 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11644 default), and then completely removed.
11648 * Test for certificates which contain unsupported critical extensions.
11649 If such a certificate is found during a verify operation it is
11650 rejected by default: this behaviour can be overridden by either
11651 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11652 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11653 X509_supported_extension() has also been added which returns 1 if a
11654 particular extension is supported.
11658 * Modify the behaviour of EVP cipher functions in similar way to digests
11659 to retain compatibility with existing code.
11663 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11664 compatibility with existing code. In particular the 'ctx' parameter does
11665 not have to be to be initialized before the call to EVP_DigestInit() and
11666 it is tidied up after a call to EVP_DigestFinal(). New function
11667 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11668 EVP_MD_CTX_copy() changed to not require the destination to be
11669 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11670 requires the destination to be valid.
11672 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11673 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11677 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11678 so that complete 'Handshake' protocol structures are kept in memory
11679 instead of overwriting 'msg_type' and 'length' with 'body' data.
11683 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11685 *Massimo Santin via Richard Levitte*
11687 * Major restructuring to the underlying ENGINE code. This includes
11688 reduction of linker bloat, separation of pure "ENGINE" manipulation
11689 (initialisation, etc) from functionality dealing with implementations
11690 of specific crypto interfaces. This change also introduces integrated
11691 support for symmetric ciphers and digest implementations - so ENGINEs
11692 can now accelerate these by providing EVP_CIPHER and EVP_MD
11693 implementations of their own. This is detailed in
11694 [crypto/engine/README.md](crypto/engine/README.md)
11695 as it couldn't be adequately described here. However, there are a few
11696 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11697 were changed in the original introduction of ENGINE code have now
11698 reverted back - the hooking from this code to ENGINE is now a good
11699 deal more passive and at run-time, operations deal directly with
11700 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11701 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11702 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11703 they were not being used by the framework as there is no concept of a
11704 BIGNUM_METHOD and they could not be generalised to the new
11705 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11706 ENGINE_cpy() has been removed as it cannot be consistently defined in
11711 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11715 * Change mkdef.pl to sort symbols that get the same entry number,
11716 and make sure the automatically generated functions `ERR_load_*`
11717 become part of libeay.num as well.
11721 * New function SSL_renegotiate_pending(). This returns true once
11722 renegotiation has been requested (either SSL_renegotiate() call
11723 or HelloRequest/ClientHello received from the peer) and becomes
11724 false once a handshake has been completed.
11725 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11726 sends a HelloRequest, but does not ensure that a handshake takes
11727 place. SSL_renegotiate_pending() is useful for checking if the
11728 client has followed the request.)
11732 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11733 By default, clients may request session resumption even during
11734 renegotiation (if session ID contexts permit); with this option,
11735 session resumption is possible only in the first handshake.
11737 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11738 more bits available for options that should not be part of
11739 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11743 * Add some demos for certificate and certificate request creation.
11747 * Make maximum certificate chain size accepted from the peer application
11748 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11749 "Douglas E. Engert" <deengert@anl.gov>.
11753 * Add support for shared libraries for Unixware-7
11754 (Boyd Lynn Gerber <gerberb@zenez.com>).
11758 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11759 be done prior to destruction. Use this to unload error strings from
11760 ENGINEs that load their own error strings. NB: This adds two new API
11761 functions to "get" and "set" this destroy handler in an ENGINE.
11765 * Alter all existing ENGINE implementations (except "openssl" and
11766 "openbsd") to dynamically instantiate their own error strings. This
11767 makes them more flexible to be built both as statically-linked ENGINEs
11768 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11769 Also, add stub code to each that makes building them as self-contained
11770 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11774 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11775 implementations into applications that are completely implemented in
11776 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11777 commands that can be used to configure what shared-library to load and
11778 to control aspects of the way it is handled. Also, made an update to
11779 the [README-Engine.md](README-Engine.md) file
11780 that brings its information up-to-date and
11781 provides some information and instructions on the "dynamic" ENGINE
11782 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11786 * Make it possible to unload ranges of ERR strings with a new
11787 "ERR_unload_strings" function.
11791 * Add a copy() function to EVP_MD.
11795 * Make EVP_MD routines take a context pointer instead of just the
11796 md_data void pointer.
11800 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11801 that the digest can only process a single chunk of data
11802 (typically because it is provided by a piece of
11803 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11804 is only going to provide a single chunk of data, and hence the
11805 framework needn't accumulate the data for oneshot drivers.
11809 * As with "ERR", make it possible to replace the underlying "ex_data"
11810 functions. This change also alters the storage and management of global
11811 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11812 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11813 index counters. The API functions that use this state have been changed
11814 to take a "class_index" rather than pointers to the class's local STACK
11815 and counter, and there is now an API function to dynamically create new
11816 classes. This centralisation allows us to (a) plug a lot of the
11817 thread-safety problems that existed, and (b) makes it possible to clean
11818 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11819 such data would previously have always leaked in application code and
11820 workarounds were in place to make the memory debugging turn a blind eye
11821 to it. Application code that doesn't use this new function will still
11822 leak as before, but their memory debugging output will announce it now
11823 rather than letting it slide.
11825 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11826 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11827 has a return value to indicate success or failure.
11831 * Make it possible to replace the underlying "ERR" functions such that the
11832 global state (2 LHASH tables and 2 locks) is only used by the "default"
11833 implementation. This change also adds two functions to "get" and "set"
11834 the implementation prior to it being automatically set the first time
11835 any other ERR function takes place. Ie. an application can call "get",
11836 pass the return value to a module it has just loaded, and that module
11837 can call its own "set" function using that value. This means the
11838 module's "ERR" operations will use (and modify) the error state in the
11839 application and not in its own statically linked copy of OpenSSL code.
11843 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11844 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11845 the operation, and provides a more encapsulated way for external code
11846 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11847 to use these functions rather than manually incrementing the counts.
11849 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11853 * Add EVP test program.
11857 * Add symmetric cipher support to ENGINE. Expect the API to change!
11861 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11862 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11863 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11864 These allow a CRL to be built without having to access X509_CRL fields
11865 directly. Modify 'ca' application to use new functions.
11869 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11870 bug workarounds. Rollback attack detection is a security feature.
11871 The problem will only arise on OpenSSL servers when TLSv1 is not
11872 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11873 Software authors not wanting to support TLSv1 will have special reasons
11874 for their choice and can explicitly enable this option.
11876 *Bodo Moeller, Lutz Jaenicke*
11878 * Rationalise EVP so it can be extended: don't include a union of
11879 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11880 (similar to those existing for EVP_CIPHER_CTX).
11885 EVP_MD_CTX_init(&md); /* new function call */
11886 EVP_DigestInit(&md, EVP_sha1());
11887 EVP_DigestUpdate(&md, in, len);
11888 EVP_DigestFinal(&md, out, NULL);
11889 EVP_MD_CTX_cleanup(&md); /* new function call */
11893 * Make DES key schedule conform to the usual scheme, as well as
11894 correcting its structure. This means that calls to DES functions
11895 now have to pass a pointer to a des_key_schedule instead of a
11896 plain des_key_schedule (which was actually always a pointer
11899 des_key_schedule ks;
11901 des_set_key_checked(..., &ks);
11902 des_ncbc_encrypt(..., &ks, ...);
11904 (Note that a later change renames 'des_...' into 'DES_...'.)
11908 * Initial reduction of linker bloat: the use of some functions, such as
11909 PEM causes large amounts of unused functions to be linked in due to
11910 poor organisation. For example pem_all.c contains every PEM function
11911 which has a knock on effect of linking in large amounts of (unused)
11912 ASN1 code. Grouping together similar functions and splitting unrelated
11913 functions prevents this.
11917 * Cleanup of EVP macros.
11921 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11922 correct `_ecb suffix`.
11926 * Add initial OCSP responder support to ocsp application. The
11927 revocation information is handled using the text based index
11928 use by the ca application. The responder can either handle
11929 requests generated internally, supplied in files (for example
11930 via a CGI script) or using an internal minimal server.
11934 * Add configuration choices to get zlib compression for TLS.
11938 * Changes to Kerberos SSL for RFC 2712 compliance:
11939 1. Implemented real KerberosWrapper, instead of just using
11940 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11941 2. Implemented optional authenticator field of KerberosWrapper.
11943 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11944 and authenticator structs; see crypto/krb5/.
11946 Generalized Kerberos calls to support multiple Kerberos libraries.
11947 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11948 via Richard Levitte*
11950 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11951 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11952 values for each of the key sizes rather than having just
11953 parameters (and 'speed' generating keys each time).
11957 * Speed up EVP routines.
11960 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11961 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11962 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11963 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11965 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11966 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11967 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11970 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11972 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11976 * Added the OS2-EMX target.
11978 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11980 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11981 to support NCONF routines in extension code. New function CONF_set_nconf()
11982 to allow functions which take an NCONF to also handle the old LHASH
11983 structure: this means that the old CONF compatible routines can be
11984 retained (in particular wrt extensions) without having to duplicate the
11985 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11989 * Enhance the general user interface with mechanisms for inner control
11990 and with possibilities to have yes/no kind of prompts.
11994 * Change all calls to low level digest routines in the library and
11995 applications to use EVP. Add missing calls to HMAC_cleanup() and
11996 don't assume HMAC_CTX can be copied using memcpy().
11998 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12000 * Add the possibility to control engines through control names but with
12001 arbitrary arguments instead of just a string.
12002 Change the key loaders to take a UI_METHOD instead of a callback
12003 function pointer. NOTE: this breaks binary compatibility with earlier
12004 versions of OpenSSL [engine].
12005 Adapt the nCipher code for these new conditions and add a card insertion
12010 * Enhance the general user interface with mechanisms to better support
12011 dialog box interfaces, application-defined prompts, the possibility
12012 to use defaults (for example default passwords from somewhere else)
12013 and interrupts/cancellations.
12017 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12018 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12022 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12023 tidy up some unnecessarily weird code in 'sk_new()').
12025 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12027 * Change the key loading routines for ENGINEs to use the same kind
12028 callback (pem_password_cb) as all other routines that need this
12033 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12034 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12035 than this minimum value is recommended.
12039 * New random seeder for OpenVMS, using the system process statistics
12040 that are easily reachable.
12044 * Windows apparently can't transparently handle global
12045 variables defined in DLLs. Initialisations such as:
12047 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12049 won't compile. This is used by the any applications that need to
12050 declare their own ASN1 modules. This was fixed by adding the option
12051 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12052 needed for static libraries under Win32.
12056 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12057 setting of purpose and trust fields. New X509_STORE trust and
12058 purpose functions and tidy up setting in other SSL functions.
12062 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12063 structure. These are inherited by X509_STORE_CTX when it is
12064 initialised. This allows various defaults to be set in the
12065 X509_STORE structure (such as flags for CRL checking and custom
12066 purpose or trust settings) for functions which only use X509_STORE_CTX
12067 internally such as S/MIME.
12069 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12070 trust settings if they are not set in X509_STORE. This allows X509_STORE
12071 purposes and trust (in S/MIME for example) to override any set by default.
12073 Add command line options for CRL checking to smime, s_client and s_server
12078 * Initial CRL based revocation checking. If the CRL checking flag(s)
12079 are set then the CRL is looked up in the X509_STORE structure and
12080 its validity and signature checked, then if the certificate is found
12081 in the CRL the verify fails with a revoked error.
12083 Various new CRL related callbacks added to X509_STORE_CTX structure.
12085 Command line options added to 'verify' application to support this.
12087 This needs some additional work, such as being able to handle multiple
12088 CRLs with different times, extension based lookup (rather than just
12089 by subject name) and ultimately more complete V2 CRL extension
12094 * Add a general user interface API (crypto/ui/). This is designed
12095 to replace things like des_read_password and friends (backward
12096 compatibility functions using this new API are provided).
12097 The purpose is to remove prompting functions from the DES code
12098 section as well as provide for prompting through dialog boxes in
12099 a window system and the like.
12103 * Add "ex_data" support to ENGINE so implementations can add state at a
12104 per-structure level rather than having to store it globally.
12108 * Make it possible for ENGINE structures to be copied when retrieved by
12109 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12110 This causes the "original" ENGINE structure to act like a template,
12111 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12112 operational state can be localised to each ENGINE structure, despite the
12113 fact they all share the same "methods". New ENGINE structures returned in
12114 this case have no functional references and the return value is the single
12115 structural reference. This matches the single structural reference returned
12116 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12121 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12122 needs to match any other type at all we need to manually clear the
12127 * Changes to the "openssl engine" utility to include;
12128 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12129 about an ENGINE's available control commands.
12130 - executing control commands from command line arguments using the
12131 '-pre' and '-post' switches. '-post' is only used if '-t' is
12132 specified and the ENGINE is successfully initialised. The syntax for
12133 the individual commands are colon-separated, for example;
12134 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12138 * New dynamic control command support for ENGINEs. ENGINEs can now
12139 declare their own commands (numbers), names (strings), descriptions,
12140 and input types for run-time discovery by calling applications. A
12141 subset of these commands are implicitly classed as "executable"
12142 depending on their input type, and only these can be invoked through
12143 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12144 can be based on user input, config files, etc). The distinction is
12145 that "executable" commands cannot return anything other than a boolean
12146 result and can only support numeric or string input, whereas some
12147 discoverable commands may only be for direct use through
12148 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12149 pointers, or other custom uses. The "executable" commands are to
12150 support parameterisations of ENGINE behaviour that can be
12151 unambiguously defined by ENGINEs and used consistently across any
12152 OpenSSL-based application. Commands have been added to all the
12153 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12154 control over shared-library paths without source code alterations.
12158 * Changed all ENGINE implementations to dynamically allocate their
12159 ENGINEs rather than declaring them statically. Apart from this being
12160 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12161 this also allows the implementations to compile without using the
12162 internal engine_int.h header.
12166 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12167 'const' value. Any code that should be able to modify a RAND_METHOD
12168 should already have non-const pointers to it (ie. they should only
12169 modify their own ones).
12173 * Made a variety of little tweaks to the ENGINE code.
12174 - "atalla" and "ubsec" string definitions were moved from header files
12175 to C code. "nuron" string definitions were placed in variables
12176 rather than hard-coded - allowing parameterisation of these values
12177 later on via ctrl() commands.
12178 - Removed unused "#if 0"'d code.
12179 - Fixed engine list iteration code so it uses ENGINE_free() to release
12180 structural references.
12181 - Constified the RAND_METHOD element of ENGINE structures.
12182 - Constified various get/set functions as appropriate and added
12183 missing functions (including a catch-all ENGINE_cpy that duplicates
12184 all ENGINE values onto a new ENGINE except reference counts/state).
12185 - Removed NULL parameter checks in get/set functions. Setting a method
12186 or function to NULL is a way of cancelling out a previously set
12187 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12188 and doesn't justify the extra error symbols and code.
12189 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12190 flags from engine_int.h to engine.h.
12191 - Changed prototypes for ENGINE handler functions (init(), finish(),
12192 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12196 * Implement binary inversion algorithm for BN_mod_inverse in addition
12197 to the algorithm using long division. The binary algorithm can be
12198 used only if the modulus is odd. On 32-bit systems, it is faster
12199 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12200 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12201 up to 450 bits. In 64-bit environments, the binary algorithm
12202 appears to be advantageous for much longer moduli; here we use it
12203 for moduli up to 2048 bits.
12207 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12208 could not support the combine flag in choice fields.
12212 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12213 extensions from a certificate request to the certificate.
12217 * Allow multiple 'certopt' and 'nameopt' options to be separated
12218 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12219 file: this allows the display of the certificate about to be
12220 signed to be customised, to allow certain fields to be included
12221 or excluded and extension details. The old system didn't display
12222 multicharacter strings properly, omitted fields not in the policy
12223 and couldn't display additional details such as extensions.
12227 * Function EC_POINTs_mul for multiple scalar multiplication
12228 of an arbitrary number of elliptic curve points
12229 \sum scalars[i]*points[i],
12230 optionally including the generator defined for the EC_GROUP:
12231 scalar*generator + \sum scalars[i]*points[i].
12233 EC_POINT_mul is a simple wrapper function for the typical case
12234 that the point list has just one item (besides the optional
12239 * First EC_METHODs for curves over GF(p):
12241 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12242 operations and provides various method functions that can also
12243 operate with faster implementations of modular arithmetic.
12245 EC_GFp_mont_method() reuses most functions that are part of
12246 EC_GFp_simple_method, but uses Montgomery arithmetic.
12248 *Bodo Moeller; point addition and point doubling
12249 implementation directly derived from source code provided by
12250 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12252 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12253 crypto/ec/ec_lib.c):
12255 Curves are EC_GROUP objects (with an optional group generator)
12256 based on EC_METHODs that are built into the library.
12258 Points are EC_POINT objects based on EC_GROUP objects.
12260 Most of the framework would be able to handle curves over arbitrary
12261 finite fields, but as there are no obvious types for fields other
12262 than GF(p), some functions are limited to that for now.
12266 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12267 that the file contains a complete HTTP response.
12271 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12272 change the def and num file printf format specifier from "%-40sXXX"
12273 to "%-39s XXX". The latter will always guarantee a space after the
12274 field while the former will cause them to run together if the field
12275 is 40 of more characters long.
12279 * Constify the cipher and digest 'method' functions and structures
12280 and modify related functions to take constant EVP_MD and EVP_CIPHER
12285 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12286 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12290 * Modify `EVP_Digest*()` routines so they now return values. Although the
12291 internal software routines can never fail additional hardware versions
12296 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12298 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12299 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12302 ERR_R_NESTED_ASN1_ERROR
12304 ERR_R_MISSING_ASN1_EOS
12305 were 4 .. 9, conflicting with
12306 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12308 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12309 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12311 Add new error code 'ERR_R_INTERNAL_ERROR'.
12315 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12320 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12321 sets the subject name for a new request or supersedes the
12322 subject name in a given request. Formats that can be parsed are
12323 'CN=Some Name, OU=myOU, C=IT'
12325 'CN=Some Name/OU=myOU/C=IT'.
12327 Add options '-batch' and '-verbose' to 'openssl req'.
12329 *Massimiliano Pala <madwolf@hackmasters.net>*
12331 * Introduce the possibility to access global variables through
12332 functions on platform were that's the best way to handle exporting
12333 global variables in shared libraries. To enable this functionality,
12334 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12335 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12336 is normally done by Configure or something similar).
12338 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12339 in the source file (foo.c) like this:
12341 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12342 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12344 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12345 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12347 OPENSSL_DECLARE_GLOBAL(int,foo);
12348 #define foo OPENSSL_GLOBAL_REF(foo)
12349 OPENSSL_DECLARE_GLOBAL(double,bar);
12350 #define bar OPENSSL_GLOBAL_REF(bar)
12352 The #defines are very important, and therefore so is including the
12353 header file everywhere where the defined globals are used.
12355 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12356 of ASN.1 items, but that structure is a bit different.
12358 The largest change is in util/mkdef.pl which has been enhanced with
12359 better and easier to understand logic to choose which symbols should
12360 go into the Windows .def files as well as a number of fixes and code
12361 cleanup (among others, algorithm keywords are now sorted
12362 lexicographically to avoid constant rewrites).
12366 * In BN_div() keep a copy of the sign of 'num' before writing the
12367 result to 'rm' because if rm==num the value will be overwritten
12368 and produce the wrong result if 'num' is negative: this caused
12369 problems with BN_mod() and BN_nnmod().
12373 * Function OCSP_request_verify(). This checks the signature on an
12374 OCSP request and verifies the signer certificate. The signer
12375 certificate is just checked for a generic purpose and OCSP request
12380 * Add OCSP_check_validity() function to check the validity of OCSP
12381 responses. OCSP responses are prepared in real time and may only
12382 be a few seconds old. Simply checking that the current time lies
12383 between thisUpdate and nextUpdate max reject otherwise valid responses
12384 caused by either OCSP responder or client clock inaccuracy. Instead
12385 we allow thisUpdate and nextUpdate to fall within a certain period of
12386 the current time. The age of the response can also optionally be
12387 checked. Two new options -validity_period and -status_age added to
12392 * If signature or public key algorithm is unrecognized print out its
12393 OID rather that just UNKNOWN.
12397 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12398 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12399 ID to be generated from the issuer certificate alone which can then be
12400 passed to OCSP_id_issuer_cmp().
12404 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12405 ASN1 modules to export functions returning ASN1_ITEM pointers
12406 instead of the ASN1_ITEM structures themselves. This adds several
12407 new macros which allow the underlying ASN1 function/structure to
12408 be accessed transparently. As a result code should not use ASN1_ITEM
12409 references directly (such as &X509_it) but instead use the relevant
12410 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12411 use of the new ASN1 code on platforms where exporting structures
12412 is problematical (for example in shared libraries) but exporting
12413 functions returning pointers to structures is not.
12417 * Add support for overriding the generation of SSL/TLS session IDs.
12418 These callbacks can be registered either in an SSL_CTX or per SSL.
12419 The purpose of this is to allow applications to control, if they wish,
12420 the arbitrary values chosen for use as session IDs, particularly as it
12421 can be useful for session caching in multiple-server environments. A
12422 command-line switch for testing this (and any client code that wishes
12423 to use such a feature) has been added to "s_server".
12425 *Geoff Thorpe, Lutz Jaenicke*
12427 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12428 of the form `#if defined(...) || defined(...) || ...` and
12429 `#if !defined(...) && !defined(...) && ...`. This also avoids
12430 the growing number of special cases it was previously handling.
12434 * Make all configuration macros available for application by making
12435 sure they are available in opensslconf.h, by giving them names starting
12436 with `OPENSSL_` to avoid conflicts with other packages and by making
12437 sure e_os2.h will cover all platform-specific cases together with
12439 Additionally, it is now possible to define configuration/platform-
12440 specific names (called "system identities"). In the C code, these
12441 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12442 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12443 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12448 * New option -set_serial to 'req' and 'x509' this allows the serial
12449 number to use to be specified on the command line. Previously self
12450 signed certificates were hard coded with serial number 0 and the
12451 CA options of 'x509' had to use a serial number in a file which was
12456 * New options to 'ca' utility to support V2 CRL entry extensions.
12457 Currently CRL reason, invalidity date and hold instruction are
12458 supported. Add new CRL extensions to V3 code and some new objects.
12462 * New function EVP_CIPHER_CTX_set_padding() this is used to
12463 disable standard block padding (aka PKCS#5 padding) in the EVP
12464 API, which was previously mandatory. This means that the data is
12465 not padded in any way and so the total length much be a multiple
12466 of the block size, otherwise an error occurs.
12470 * Initial (incomplete) OCSP SSL support.
12474 * New function OCSP_parse_url(). This splits up a URL into its host,
12475 port and path components: primarily to parse OCSP URLs. New -url
12476 option to ocsp utility.
12480 * New nonce behavior. The return value of OCSP_check_nonce() now
12481 reflects the various checks performed. Applications can decide
12482 whether to tolerate certain situations such as an absent nonce
12483 in a response when one was present in a request: the ocsp application
12484 just prints out a warning. New function OCSP_add1_basic_nonce()
12485 this is to allow responders to include a nonce in a response even if
12486 the request is nonce-less.
12490 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12491 skipped when using openssl x509 multiple times on a single input file,
12492 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12496 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12497 set string type: to handle setting ASN1_TIME structures. Fix ca
12498 utility to correctly initialize revocation date of CRLs.
12502 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12503 the clients preferred ciphersuites and rather use its own preferences.
12504 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12505 Internet Explorer by ensuring unchanged hash method during stepup.
12506 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12510 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12511 to aes and add a new 'exist' option to print out symbols that don't
12516 * Additional options to ocsp utility to allow flags to be set and
12517 additional certificates supplied.
12521 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12522 OCSP client a number of certificate to only verify the response
12527 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12528 handle the new API. Currently only ECB, CBC modes supported. Add new
12531 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12532 Encryption Standard (AES) Ciphersuites for Transport Layer
12533 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12534 not enabled by default and were not part of the "ALL" ciphersuite
12535 alias because they were not yet official; they could be
12536 explicitly requested by specifying the "AESdraft" ciphersuite
12537 group alias. In the final release of OpenSSL 0.9.7, the group
12538 alias is called "AES" and is part of "ALL".)
12540 *Ben Laurie, Steve Henson, Bodo Moeller*
12542 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12543 request to response.
12547 * Functions for OCSP responders. OCSP_request_onereq_count(),
12548 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12549 extract information from a certificate request. OCSP_response_create()
12550 creates a response and optionally adds a basic response structure.
12551 OCSP_basic_add1_status() adds a complete single response to a basic
12552 response and returns the OCSP_SINGLERESP structure just added (to allow
12553 extensions to be included for example). OCSP_basic_add1_cert() adds a
12554 certificate to a basic response and OCSP_basic_sign() signs a basic
12555 response with various flags. New helper functions ASN1_TIME_check()
12556 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12557 (converts ASN1_TIME to GeneralizedTime).
12561 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12562 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12563 structure from a certificate. X509_pubkey_digest() digests the public_key
12564 contents: this is used in various key identifiers.
12568 * Make sk_sort() tolerate a NULL argument.
12570 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12572 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12573 passed by the function are trusted implicitly. If any of them signed the
12574 response then it is assumed to be valid and is not verified.
12578 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12579 to data. This was previously part of the PKCS7 ASN1 code. This
12580 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12581 *Steve Henson, reported by Kenneth R. Robinette
12582 <support@securenetterm.com>*
12584 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12585 routines: without these tracing memory leaks is very painful.
12586 Fix leaks in PKCS12 and PKCS7 routines.
12590 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12591 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12592 effectively meant GeneralizedTime would never be used. Now it
12593 is initialised to -1 but X509_time_adj() now has to check the value
12594 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12595 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12596 *Steve Henson, reported by Kenneth R. Robinette
12597 <support@securenetterm.com>*
12599 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12600 result in a zero length in the ASN1_INTEGER structure which was
12601 not consistent with the structure when d2i_ASN1_INTEGER() was used
12602 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12603 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12604 where it did not print out a minus for negative ASN1_INTEGER.
12608 * Add summary printout to ocsp utility. The various functions which
12609 convert status values to strings have been renamed to:
12610 OCSP_response_status_str(), OCSP_cert_status_str() and
12611 OCSP_crl_reason_str() and are no longer static. New options
12612 to verify nonce values and to disable verification. OCSP response
12613 printout format cleaned up.
12617 * Add additional OCSP certificate checks. These are those specified
12618 in RFC2560. This consists of two separate checks: the CA of the
12619 certificate being checked must either be the OCSP signer certificate
12620 or the issuer of the OCSP signer certificate. In the latter case the
12621 OCSP signer certificate must contain the OCSP signing extended key
12622 usage. This check is performed by attempting to match the OCSP
12623 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12624 in the OCSP_CERTID structures of the response.
12628 * Initial OCSP certificate verification added to OCSP_basic_verify()
12629 and related routines. This uses the standard OpenSSL certificate
12630 verify routines to perform initial checks (just CA validity) and
12631 to obtain the certificate chain. Then additional checks will be
12632 performed on the chain. Currently the root CA is checked to see
12633 if it is explicitly trusted for OCSP signing. This is used to set
12634 a root CA as a global signing root: that is any certificate that
12635 chains to that CA is an acceptable OCSP signing certificate.
12639 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12640 extensions from a separate configuration file.
12641 As when reading extensions from the main configuration file,
12642 the '-extensions ...' option may be used for specifying the
12645 *Massimiliano Pala <madwolf@comune.modena.it>*
12647 * New OCSP utility. Allows OCSP requests to be generated or
12648 read. The request can be sent to a responder and the output
12649 parsed, outputted or printed in text form. Not complete yet:
12650 still needs to check the OCSP response validity.
12654 * New subcommands for 'openssl ca':
12655 `openssl ca -status <serial>` prints the status of the cert with
12656 the given serial number (according to the index file).
12657 `openssl ca -updatedb` updates the expiry status of certificates
12660 *Massimiliano Pala <madwolf@comune.modena.it>*
12662 * New '-newreq-nodes' command option to CA.pl. This is like
12663 '-newreq', but calls 'openssl req' with the '-nodes' option
12664 so that the resulting key is not encrypted.
12666 *Damien Miller <djm@mindrot.org>*
12668 * New configuration for the GNU Hurd.
12670 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12672 * Initial code to implement OCSP basic response verify. This
12673 is currently incomplete. Currently just finds the signer's
12674 certificate and verifies the signature on the response.
12678 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12679 value of OPENSSLDIR. This is available via the new '-d' option
12680 to 'openssl version', and is also included in 'openssl version -a'.
12684 * Allowing defining memory allocation callbacks that will be given
12685 file name and line number information in additional arguments
12686 (a `const char*` and an int). The basic functionality remains, as
12687 well as the original possibility to just replace malloc(),
12688 realloc() and free() by functions that do not know about these
12689 additional arguments. To register and find out the current
12690 settings for extended allocation functions, the following
12691 functions are provided:
12693 CRYPTO_set_mem_ex_functions
12694 CRYPTO_set_locked_mem_ex_functions
12695 CRYPTO_get_mem_ex_functions
12696 CRYPTO_get_locked_mem_ex_functions
12698 These work the same way as CRYPTO_set_mem_functions and friends.
12699 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12700 extended allocation function is enabled.
12701 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12702 a conventional allocation function is enabled.
12704 *Richard Levitte, Bodo Moeller*
12706 * Finish off removing the remaining LHASH function pointer casts.
12707 There should no longer be any prototype-casting required when using
12708 the LHASH abstraction, and any casts that remain are "bugs". See
12709 the callback types and macros at the head of lhash.h for details
12710 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12714 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12715 If /dev/[u]random devices are not available or do not return enough
12716 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12718 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12719 /etc/entropy will be queried once each in this sequence, querying stops
12720 when enough entropy was collected without querying more sockets.
12724 * Change the Unix RAND_poll() variant to be able to poll several
12725 random devices, as specified by DEVRANDOM, until a sufficient amount
12726 of data has been collected. We spend at most 10 ms on each file
12727 (select timeout) and read in non-blocking mode. DEVRANDOM now
12728 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12729 (previously it was just the string "/dev/urandom"), so on typical
12730 platforms the 10 ms delay will never occur.
12731 Also separate out the Unix variant to its own file, rand_unix.c.
12732 For VMS, there's a currently-empty rand_vms.c.
12736 * Move OCSP client related routines to ocsp_cl.c. These
12737 provide utility functions which an application needing
12738 to issue a request to an OCSP responder and analyse the
12739 response will typically need: as opposed to those which an
12740 OCSP responder itself would need which will be added later.
12742 OCSP_request_sign() signs an OCSP request with an API similar
12743 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12744 response. OCSP_response_get1_basic() extracts basic response
12745 from response. OCSP_resp_find_status(): finds and extracts status
12746 information from an OCSP_CERTID structure (which will be created
12747 when the request structure is built). These are built from lower
12748 level functions which work on OCSP_SINGLERESP structures but
12749 won't normally be used unless the application wishes to examine
12750 extensions in the OCSP response for example.
12752 Replace nonce routines with a pair of functions.
12753 OCSP_request_add1_nonce() adds a nonce value and optionally
12754 generates a random value. OCSP_check_nonce() checks the
12755 validity of the nonce in an OCSP response.
12759 * Change function OCSP_request_add() to OCSP_request_add0_id().
12760 This doesn't copy the supplied OCSP_CERTID and avoids the
12761 need to free up the newly created id. Change return type
12762 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12763 This can then be used to add extensions to the request.
12764 Deleted OCSP_request_new(), since most of its functionality
12765 is now in OCSP_REQUEST_new() (and the case insensitive name
12766 clash) apart from the ability to set the request name which
12767 will be added elsewhere.
12771 * Update OCSP API. Remove obsolete extensions argument from
12772 various functions. Extensions are now handled using the new
12773 OCSP extension code. New simple OCSP HTTP function which
12774 can be used to send requests and parse the response.
12778 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12779 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12780 uses the special reorder version of SET OF to sort the attributes
12781 and reorder them to match the encoded order. This resolves a long
12782 standing problem: a verify on a PKCS7 structure just after signing
12783 it used to fail because the attribute order did not match the
12784 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12785 it uses the received order. This is necessary to tolerate some broken
12786 software that does not order SET OF. This is handled by encoding
12787 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12788 to produce the required SET OF.
12792 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12793 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12794 files to get correct declarations of the ASN.1 item variables.
12798 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12799 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12800 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12801 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12802 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12803 ASN1_ITEM and no wrapper functions.
12807 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12808 replace the old function pointer based I/O routines. Change most of
12809 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12813 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12814 lines, recognize more "algorithms" that can be deselected, and make
12815 it complain about algorithm deselection that isn't recognised.
12819 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12820 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12821 to use new functions. Add NO_ASN1_OLD which can be set to remove
12822 some old style ASN1 functions: this can be used to determine if old
12823 code will still work when these eventually go away.
12827 * New extension functions for OCSP structures, these follow the
12828 same conventions as certificates and CRLs.
12832 * New function X509V3_add1_i2d(). This automatically encodes and
12833 adds an extension. Its behaviour can be customised with various
12834 flags to append, replace or delete. Various wrappers added for
12835 certificates and CRLs.
12839 * Fix to avoid calling the underlying ASN1 print routine when
12840 an extension cannot be parsed. Correct a typo in the
12841 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12845 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12846 entries for variables.
12850 * Add functionality to apps/openssl.c for detecting locking
12851 problems: As the program is single-threaded, all we have
12852 to do is register a locking callback using an array for
12853 storing which locks are currently held by the program.
12857 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12858 SSL_get_ex_data_X509_STORE_idx(), which is used in
12859 ssl_verify_cert_chain() and thus can be called at any time
12860 during TLS/SSL handshakes so that thread-safety is essential.
12861 Unfortunately, the ex_data design is not at all suited
12862 for multi-threaded use, so it probably should be abolished.
12866 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12868 *Broadcom, tweaked and integrated by Geoff Thorpe*
12870 * Move common extension printing code to new function
12871 X509V3_print_extensions(). Reorganise OCSP print routines and
12872 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12876 * New function X509_signature_print() to remove duplication in some
12881 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12882 set (this was treated exactly the same as SET OF previously). This
12883 is used to reorder the STACK representing the structure to match the
12884 encoding. This will be used to get round a problem where a PKCS7
12885 structure which was signed could not be verified because the STACK
12886 order did not reflect the encoded order.
12890 * Reimplement the OCSP ASN1 module using the new code.
12894 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12895 for its ASN1 operations. The old style function pointers still exist
12896 for now but they will eventually go away.
12900 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12901 completely replaces the old ASN1 functionality with a table driven
12902 encoder and decoder which interprets an ASN1_ITEM structure describing
12903 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12904 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12905 has also been converted to the new form.
12909 * Change BN_mod_exp_recp so that negative moduli are tolerated
12910 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12911 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12912 for negative moduli.
12916 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12917 of not touching the result's sign bit.
12921 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12926 * Changed the LHASH code to use prototypes for callbacks, and created
12927 macros to declare and implement thin (optionally static) functions
12928 that provide type-safety and avoid function pointer casting for the
12929 type-specific callbacks.
12933 * Added Kerberos Cipher Suites to be used with TLS, as written in
12935 *Veers Staats <staatsvr@asc.hpc.mil>,
12936 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12938 * Reformat the FAQ so the different questions and answers can be divided
12939 in sections depending on the subject.
12943 * Have the zlib compression code load ZLIB.DLL dynamically under
12948 * New function BN_mod_sqrt for computing square roots modulo a prime
12949 (using the probabilistic Tonelli-Shanks algorithm unless
12950 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12951 be handled deterministically).
12953 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12955 * Make BN_mod_inverse faster by explicitly handling small quotients
12956 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12957 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12961 * New function BN_kronecker.
12965 * Fix BN_gcd so that it works on negative inputs; the result is
12966 positive unless both parameters are zero.
12967 Previously something reasonably close to an infinite loop was
12968 possible because numbers could be growing instead of shrinking
12969 in the implementation of Euclid's algorithm.
12973 * Fix BN_is_word() and BN_is_one() macros to take into account the
12974 sign of the number in question.
12976 Fix BN_is_word(a,w) to work correctly for w == 0.
12978 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12979 because its test if the absolute value of 'a' equals 'w'.
12980 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12981 it exists mostly for use in the implementations of BN_is_zero(),
12982 BN_is_one(), and BN_is_word().
12986 * New function BN_swap.
12990 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12991 the exponentiation functions are more likely to produce reasonable
12992 results on negative inputs.
12996 * Change BN_mod_mul so that the result is always non-negative.
12997 Previously, it could be negative if one of the factors was negative;
12998 I don't think anyone really wanted that behaviour.
13002 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13003 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13004 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13005 and add new functions:
13014 BN_mod_lshift1_quick
13016 BN_mod_lshift_quick
13018 These functions always generate non-negative results.
13020 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13021 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13023 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13024 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13025 be reduced modulo `m`.
13027 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13030 The following entry accidentally appeared in the CHANGES file
13031 distributed with OpenSSL 0.9.7. The modifications described in
13032 it do *not* apply to OpenSSL 0.9.7.
13034 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13035 was actually never needed) and in BN_mul(). The removal in BN_mul()
13036 required a small change in bn_mul_part_recursive() and the addition
13037 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13038 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13039 bn_sub_words() and bn_add_words() except they take arrays with
13045 * In 'openssl passwd', verify passwords read from the terminal
13046 unless the '-salt' option is used (which usually means that
13047 verification would just waste user's time since the resulting
13048 hash is going to be compared with some given password hash)
13049 or the new '-noverify' option is used.
13051 This is an incompatible change, but it does not affect
13052 non-interactive use of 'openssl passwd' (passwords on the command
13053 line, '-stdin' option, '-in ...' option) and thus should not
13054 cause any problems.
13058 * Remove all references to RSAref, since there's no more need for it.
13062 * Make DSO load along a path given through an environment variable
13063 (SHLIB_PATH) with shl_load().
13067 * Constify the ENGINE code as a result of BIGNUM constification.
13068 Also constify the RSA code and most things related to it. In a
13069 few places, most notable in the depth of the ASN.1 code, ugly
13070 casts back to non-const were required (to be solved at a later
13075 * Make it so the openssl application has all engines loaded by default.
13079 * Constify the BIGNUM routines a little more.
13083 * Add the following functions:
13085 ENGINE_load_cswift()
13087 ENGINE_load_atalla()
13088 ENGINE_load_nuron()
13089 ENGINE_load_builtin_engines()
13091 That way, an application can itself choose if external engines that
13092 are built-in in OpenSSL shall ever be used or not. The benefit is
13093 that applications won't have to be linked with libdl or other dso
13094 libraries unless it's really needed.
13096 Changed 'openssl engine' to load all engines on demand.
13097 Changed the engine header files to avoid the duplication of some
13098 declarations (they differed!).
13102 * 'openssl engine' can now list capabilities.
13106 * Better error reporting in 'openssl engine'.
13110 * Never call load_dh_param(NULL) in s_server.
13114 * Add engine application. It can currently list engines by name and
13115 identity, and test if they are actually available.
13119 * Improve RPM specification file by forcing symbolic linking and making
13120 sure the installed documentation is also owned by root.root.
13122 *Damien Miller <djm@mindrot.org>*
13124 * Give the OpenSSL applications more possibilities to make use of
13125 keys (public as well as private) handled by engines.
13129 * Add OCSP code that comes from CertCo.
13133 * Add VMS support for the Rijndael code.
13137 * Added untested support for Nuron crypto accelerator.
13141 * Add support for external cryptographic devices. This code was
13142 previously distributed separately as the "engine" branch.
13144 *Geoff Thorpe, Richard Levitte*
13146 * Rework the filename-translation in the DSO code. It is now possible to
13147 have far greater control over how a "name" is turned into a filename
13148 depending on the operating environment and any oddities about the
13149 different shared library filenames on each system.
13153 * Support threads on FreeBSD-elf in Configure.
13157 * Fix for SHA1 assembly problem with MASM: it produces
13158 warnings about corrupt line number information when assembling
13159 with debugging information. This is caused by the overlapping
13162 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13165 NCONF_get_number() has no error checking at all. As a replacement,
13166 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13167 promoted strongly. The old NCONF_get_number is kept around for
13168 binary backward compatibility.
13169 Make it possible for methods to load from something other than a BIO,
13170 by providing a function pointer that is given a name instead of a BIO.
13171 For example, this could be used to load configuration data from an
13176 * Fix for non blocking accept BIOs. Added new I/O special reason
13177 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13178 with non blocking I/O was not possible because no retry code was
13179 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13184 * Added the beginnings of Rijndael support.
13188 * Fix for bug in DirectoryString mask setting. Add support for
13189 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13190 to allow certificate printing to more controllable, additional
13191 'certopt' option to 'x509' to allow new printing options to be
13196 * Clean old EAY MD5 hack from e_os.h.
13200 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13202 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13203 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13205 *Joe Orton, Steve Henson*
13207 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13209 * Fix additional bug revealed by the NISCC test suite:
13211 Stop bug triggering large recursion when presented with
13212 certain ASN.1 tags [CVE-2003-0851][]
13216 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13218 * Fix various bugs revealed by running the NISCC test suite:
13220 Stop out of bounds reads in the ASN1 code when presented with
13221 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13223 If verify callback ignores invalid public key errors don't try to check
13224 certificate signature with the NULL public key.
13228 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13229 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13234 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13235 extra data after the compression methods not only for TLS 1.0
13236 but also for SSL 3.0 (as required by the specification).
13238 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13240 * Change X509_certificate_type() to mark the key as exported/exportable
13241 when it's 512 *bits* long, not 512 bytes.
13245 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13247 * Countermeasure against the Klima-Pokorny-Rosa extension of
13248 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13249 a protocol version number mismatch like a decryption error
13250 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13254 * Turn on RSA blinding by default in the default implementation
13255 to avoid a timing attack. Applications that don't want it can call
13256 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13257 They would be ill-advised to do so in most cases.
13259 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13261 * Change RSA blinding code so that it works when the PRNG is not
13262 seeded (in this case, the secret RSA exponent is abused as
13263 an unpredictable seed -- if it is not unpredictable, there
13264 is no point in blinding anyway). Make RSA blinding thread-safe
13265 by remembering the creator's thread ID in rsa->blinding and
13266 having all other threads use local one-time blinding factors
13267 (this requires more computation than sharing rsa->blinding, but
13268 avoids excessive locking; and if an RSA object is not shared
13269 between threads, blinding will still be very fast).
13273 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13275 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13276 via timing by performing a MAC computation even if incorrect
13277 block cipher padding has been found. This is a countermeasure
13278 against active attacks where the attacker has to distinguish
13279 between bad padding and a MAC verification error. [CVE-2003-0078][]
13281 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13282 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13283 Martin Vuagnoux (EPFL, Ilion)*
13285 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13287 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13288 memory from its contents. This is done with a counter that will
13289 place alternating values in each byte. This can be used to solve
13290 two issues: 1) the removal of calls to memset() by highly optimizing
13291 compilers, and 2) cleansing with other values than 0, since those can
13292 be read through on certain media, for example a swap space on disk.
13296 * Bugfix: client side session caching did not work with external caching,
13297 because the session->cipher setting was not restored when reloading
13298 from the external cache. This problem was masked, when
13299 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13300 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13304 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13305 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13307 *Zeev Lieber <zeev-l@yahoo.com>*
13309 * Undo an undocumented change introduced in 0.9.6e which caused
13310 repeated calls to OpenSSL_add_all_ciphers() and
13311 OpenSSL_add_all_digests() to be ignored, even after calling
13316 * Change the default configuration reader to deal with last line not
13317 being properly terminated.
13321 * Change X509_NAME_cmp() so it applies the special rules on handling
13322 DN values that are of type PrintableString, as well as RDNs of type
13323 emailAddress where the value has the type ia5String.
13325 *stefank@valicert.com via Richard Levitte*
13327 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13328 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13329 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13330 the bitwise-OR of the two for use by the majority of applications
13331 wanting this behaviour, and update the docs. The documented
13332 behaviour and actual behaviour were inconsistent and had been
13333 changing anyway, so this is more a bug-fix than a behavioural
13336 *Geoff Thorpe, diagnosed by Nadav Har'El*
13338 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13339 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13343 * Fix initialization code race conditions in
13344 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13345 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13346 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13347 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13348 ssl2_get_cipher_by_char(),
13349 ssl3_get_cipher_by_char().
13351 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13353 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13354 the cached sessions are flushed, as the remove_cb() might use ex_data
13355 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13356 (see [openssl.org #212]).
13358 *Geoff Thorpe, Lutz Jaenicke*
13360 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13361 length, instead of the encoding length to d2i_ASN1_OBJECT.
13365 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13367 * [In 0.9.6g-engine release:]
13368 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13370 *Lynn Gazis <lgazis@rainbow.com>*
13372 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13374 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13375 and get fix the header length calculation.
13376 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13377 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13379 * Use proper error handling instead of 'assertions' in buffer
13380 overflow checks added in 0.9.6e. This prevents DoS (the
13381 assertions could call abort()).
13383 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13385 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13387 * Add various sanity checks to asn1_get_length() to reject
13388 the ASN1 length bytes if they exceed sizeof(long), will appear
13389 negative or the content length exceeds the length of the
13392 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13394 * Fix cipher selection routines: ciphers without encryption had no flags
13395 for the cipher strength set and where therefore not handled correctly
13396 by the selection routines (PR #130).
13400 * Fix EVP_dsa_sha macro.
13405 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13406 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13407 that was added in OpenSSL 0.9.6d.
13409 As the countermeasure turned out to be incompatible with some
13410 broken SSL implementations, the new option is part of SSL_OP_ALL.
13411 SSL_OP_ALL is usually employed when compatibility with weird SSL
13412 implementations is desired (e.g. '-bugs' option to 's_client' and
13413 's_server'), so the new option is automatically set in many
13418 * Changes in security patch:
13420 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13421 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13422 Air Force Materiel Command, USAF, under agreement number
13425 * Add various sanity checks to asn1_get_length() to reject
13426 the ASN1 length bytes if they exceed sizeof(long), will appear
13427 negative or the content length exceeds the length of the
13428 supplied buffer. [CVE-2002-0659][]
13430 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13432 * Assertions for various potential buffer overflows, not known to
13433 happen in practice.
13435 *Ben Laurie (CHATS)*
13437 * Various temporary buffers to hold ASCII versions of integers were
13438 too small for 64 bit platforms. [CVE-2002-0655][]
13439 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13441 * Remote buffer overflow in SSL3 protocol - an attacker could
13442 supply an oversized session ID to a client. [CVE-2002-0656][]
13444 *Ben Laurie (CHATS)*
13446 * Remote buffer overflow in SSL2 protocol - an attacker could
13447 supply an oversized client master key. [CVE-2002-0656][]
13449 *Ben Laurie (CHATS)*
13451 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13453 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13454 encoded as NULL) with id-dsa-with-sha1.
13456 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13458 * Check various `X509_...()` return values in apps/req.c.
13460 *Nils Larsch <nla@trustcenter.de>*
13462 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13463 an end-of-file condition would erroneously be flagged, when the CRLF
13464 was just at the end of a processed block. The bug was discovered when
13465 processing data through a buffering memory BIO handing the data to a
13466 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13467 <ptsekov@syntrex.com> and Nedelcho Stanev.
13471 * Implement a countermeasure against a vulnerability recently found
13472 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13473 before application data chunks to avoid the use of known IVs
13474 with data potentially chosen by the attacker.
13478 * Fix length checks in ssl3_get_client_hello().
13482 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13483 to prevent ssl3_read_internal() from incorrectly assuming that
13484 ssl3_read_bytes() found application data while handshake
13485 processing was enabled when in fact s->s3->in_read_app_data was
13486 merely automatically cleared during the initial handshake.
13488 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13490 * Fix object definitions for Private and Enterprise: they were not
13491 recognized in their shortname (=lowercase) representation. Extend
13492 obj_dat.pl to issue an error when using undefined keywords instead
13493 of silently ignoring the problem (Svenning Sorensen
13494 <sss@sss.dnsalias.net>).
13498 * Fix DH_generate_parameters() so that it works for 'non-standard'
13499 generators, i.e. generators other than 2 and 5. (Previously, the
13500 code did not properly initialise the 'add' and 'rem' values to
13501 BN_generate_prime().)
13503 In the new general case, we do not insist that 'generator' is
13504 actually a primitive root: This requirement is rather pointless;
13505 a generator of the order-q subgroup is just as good, if not
13510 * Map new X509 verification errors to alerts. Discovered and submitted by
13511 Tom Wu <tom@arcot.com>.
13515 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13516 returning non-zero before the data has been completely received
13517 when using non-blocking I/O.
13519 *Bodo Moeller; problem pointed out by John Hughes*
13521 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13523 *Ben Laurie, Lutz Jaenicke*
13525 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13526 Yoram Zahavi <YoramZ@gilian.com>).
13530 * Add information about CygWin 1.3 and on, and preserve proper
13531 configuration for the versions before that.
13533 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13535 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13536 check whether we deal with a copy of a session and do not delete from
13537 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13538 <izhar@checkpoint.com>.
13542 * Do not store session data into the internal session cache, if it
13543 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13544 flag is set). Proposed by Aslam <aslam@funk.com>.
13548 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13553 * [In 0.9.6d-engine release:]
13554 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13556 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13558 * Add the configuration target linux-s390x.
13560 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13562 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13563 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13564 variable as an indication that a ClientHello message has been
13565 received. As the flag value will be lost between multiple
13566 invocations of ssl3_accept when using non-blocking I/O, the
13567 function may not be aware that a handshake has actually taken
13568 place, thus preventing a new session from being added to the
13571 To avoid this problem, we now set s->new_session to 2 instead of
13572 using a local variable.
13574 *Lutz Jaenicke, Bodo Moeller*
13576 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13577 if the SSL_R_LENGTH_MISMATCH error is detected.
13579 *Geoff Thorpe, Bodo Moeller*
13581 * New 'shared_ldflag' column in Configure platform table.
13585 * Fix EVP_CIPHER_mode macro.
13587 *"Dan S. Camper" <dan@bti.net>*
13589 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13590 type, we must throw them away by setting rr->length to 0.
13592 *D P Chang <dpc@qualys.com>*
13594 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13596 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13597 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13598 worked incorrectly for those cases where range = `10..._2` and
13599 `3*range` is two bits longer than range.)
13603 * Only add signing time to PKCS7 structures if it is not already
13608 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13609 OBJ_ld_ce should be OBJ_id_ce.
13610 Also some ip-pda OIDs in crypto/objects/objects.txt were
13611 incorrect (cf. RFC 3039).
13613 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13615 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13616 returns early because it has nothing to do.
13618 *Andy Schneider <andy.schneider@bjss.co.uk>*
13620 * [In 0.9.6c-engine release:]
13621 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13623 *Andy Schneider <andy.schneider@bjss.co.uk>*
13625 * [In 0.9.6c-engine release:]
13626 Add support for Cryptographic Appliance's keyserver technology.
13627 (Use engine 'keyclient')
13629 *Cryptographic Appliances and Geoff Thorpe*
13631 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13632 is called via tools/c89.sh because arguments have to be
13633 rearranged (all '-L' options must appear before the first object
13636 *Richard Shapiro <rshapiro@abinitio.com>*
13638 * [In 0.9.6c-engine release:]
13639 Add support for Broadcom crypto accelerator cards, backported
13642 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13644 * [In 0.9.6c-engine release:]
13645 Add support for SureWare crypto accelerator cards from
13646 Baltimore Technologies. (Use engine 'sureware')
13648 *Baltimore Technologies and Mark Cox*
13650 * [In 0.9.6c-engine release:]
13651 Add support for crypto accelerator cards from Accelerated
13652 Encryption Processing, www.aep.ie. (Use engine 'aep')
13654 *AEP Inc. and Mark Cox*
13656 * Add a configuration entry for gcc on UnixWare.
13658 *Gary Benson <gbenson@redhat.com>*
13660 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13661 messages are stored in a single piece (fixed-length part and
13662 variable-length part combined) and fix various bugs found on the way.
13666 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13667 instead. BIO_gethostbyname() does not know what timeouts are
13668 appropriate, so entries would stay in cache even when they have
13670 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13672 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13673 faced with a pathologically small ClientHello fragment that does
13674 not contain client_version: Instead of aborting with an error,
13675 simply choose the highest available protocol version (i.e.,
13676 TLS 1.0 unless it is disabled). In practice, ClientHello
13677 messages are never sent like this, but this change gives us
13678 strictly correct behaviour at least for TLS.
13682 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13683 never resets s->method to s->ctx->method when called from within
13684 one of the SSL handshake functions.
13686 *Bodo Moeller; problem pointed out by Niko Baric*
13688 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13689 (sent using the client's version number) if client_version is
13690 smaller than the protocol version in use. Also change
13691 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13692 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13693 the client will at least see that alert.
13697 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13702 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13703 client receives HelloRequest while in a handshake.
13705 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13707 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13708 should end in 'break', not 'goto end' which circumvents various
13709 cleanups done in state SSL_ST_OK. But session related stuff
13710 must be disabled for SSL_ST_OK in the case that we just sent a
13713 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13714 before just sending a HelloRequest.
13716 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13718 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13719 reveal whether illegal block cipher padding was found or a MAC
13720 verification error occurred. (Neither SSLerr() codes nor alerts
13721 are directly visible to potential attackers, but the information
13722 may leak via logfiles.)
13724 Similar changes are not required for the SSL 2.0 implementation
13725 because the number of padding bytes is sent in clear for SSL 2.0,
13726 and the extra bytes are just ignored. However ssl/s2_pkt.c
13727 failed to verify that the purported number of padding bytes is in
13732 * Add OpenUNIX-8 support including shared libraries
13733 (Boyd Lynn Gerber <gerberb@zenez.com>).
13737 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13738 'wristwatch attack' using huge encoding parameters (cf.
13739 James H. Manger's CRYPTO 2001 paper). Note that the
13740 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13741 encoding parameters and hence was not vulnerable.
13745 * BN_sqr() bug fix.
13747 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13749 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13750 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13751 followed by modular reduction.
13753 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13755 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13756 equivalent based on BN_pseudo_rand() instead of BN_rand().
13760 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13761 This function was broken, as the check for a new client hello message
13762 to handle SGC did not allow these large messages.
13763 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13767 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13771 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13772 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13776 * Rework the configuration and shared library support for Tru64 Unix.
13777 The configuration part makes use of modern compiler features and
13778 still retains old compiler behavior for those that run older versions
13779 of the OS. The shared library support part includes a variant that
13780 uses the RPATH feature, and is available through the special
13781 configuration target "alpha-cc-rpath", which will never be selected
13784 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13786 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13787 with the same message size as in ssl3_get_certificate_request().
13788 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13789 messages might inadvertently be reject as too long.
13791 *Petr Lampa <lampa@fee.vutbr.cz>*
13793 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13797 * Modified SSL library such that the verify_callback that has been set
13798 specifically for an SSL object with SSL_set_verify() is actually being
13799 used. Before the change, a verify_callback set with this function was
13800 ignored and the verify_callback() set in the SSL_CTX at the time of
13801 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13802 to allow the necessary settings.
13806 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13807 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13808 done automatically (in contradiction to the requirements of the C
13809 standard). This made problems when used from OpenSSH.
13813 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13814 dh->length and always used
13816 BN_rand_range(priv_key, dh->p).
13818 BN_rand_range() is not necessary for Diffie-Hellman, and this
13819 specific range makes Diffie-Hellman unnecessarily inefficient if
13820 dh->length (recommended exponent length) is much smaller than the
13821 length of dh->p. We could use BN_rand_range() if the order of
13822 the subgroup was stored in the DH structure, but we only have
13827 BN_rand(priv_key, l, ...)
13829 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13836 RSA_eay_public_encrypt
13837 RSA_eay_private_decrypt
13838 RSA_eay_private_encrypt (signing)
13839 RSA_eay_public_decrypt (signature verification)
13841 (default implementations for RSA_public_encrypt,
13842 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13843 always reject numbers >= n.
13847 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13848 to synchronize access to 'locking_thread'. This is necessary on
13849 systems where access to 'locking_thread' (an 'unsigned long'
13850 variable) is not atomic.
13854 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13855 *before* setting the 'crypto_lock_rand' flag. The previous code had
13856 a race condition if 0 is a valid thread ID.
13858 *Travis Vitek <vitek@roguewave.com>*
13860 * Add support for shared libraries under Irix.
13862 *Albert Chin-A-Young <china@thewrittenword.com>*
13864 * Add configuration option to build on Linux on both big-endian and
13865 little-endian MIPS.
13867 *Ralf Baechle <ralf@uni-koblenz.de>*
13869 * Add the possibility to create shared libraries on HP-UX.
13873 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13875 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13876 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13877 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13878 PRNG state recovery was possible based on the output of
13879 one PRNG request appropriately sized to gain knowledge on
13880 'md' followed by enough consecutive 1-byte PRNG requests
13881 to traverse all of 'state'.
13883 1. When updating 'md_local' (the current thread's copy of 'md')
13884 during PRNG output generation, hash all of the previous
13885 'md_local' value, not just the half used for PRNG output.
13887 2. Make the number of bytes from 'state' included into the hash
13888 independent from the number of PRNG bytes requested.
13890 The first measure alone would be sufficient to avoid
13891 Markku-Juhani's attack. (Actually it had never occurred
13892 to me that the half of 'md_local' used for chaining was the
13893 half from which PRNG output bytes were taken -- I had always
13894 assumed that the secret half would be used.) The second
13895 measure makes sure that additional data from 'state' is never
13896 mixed into 'md_local' in small portions; this heuristically
13897 further strengthens the PRNG.
13901 * Fix crypto/bn/asm/mips3.s.
13905 * When only the key is given to "enc", the IV is undefined. Print out
13906 an error message in this case.
13910 * Handle special case when X509_NAME is empty in X509 printing routines.
13914 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13915 positive and less than q.
13919 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13920 used: it isn't thread safe and the add_lock_callback should handle
13923 *Paul Rose <Paul.Rose@bridge.com>*
13925 * Verify that incoming data obeys the block size in
13926 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13932 *Ulf Möller, Bodo Möller*
13934 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13935 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13936 when fixing the server behaviour for backwards-compatible 'client
13937 hello' messages. (Note that the attack is impractical against
13938 SSL 3.0 and TLS 1.0 anyway because length and version checking
13939 means that the probability of guessing a valid ciphertext is
13940 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13943 Before 0.9.5, the countermeasure (hide the error by generating a
13944 random 'decryption result') did not work properly because
13945 ERR_clear_error() was missing, meaning that SSL_get_error() would
13946 detect the supposedly ignored error.
13948 Both problems are now fixed.
13952 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13953 (previously it was 1024).
13957 * Fix for compatibility mode trust settings: ignore trust settings
13958 unless some valid trust or reject settings are present.
13962 * Fix for blowfish EVP: its a variable length cipher.
13966 * Fix various bugs related to DSA S/MIME verification. Handle missing
13967 parameters in DSA public key structures and return an error in the
13968 DSA routines if parameters are absent.
13972 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13973 in the current directory if neither $RANDFILE nor $HOME was set.
13974 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13975 caused some confusion to Windows users who haven't defined $HOME.
13976 Thus RAND_file_name() is changed again: e_os.h can define a
13977 DEFAULT_HOME, which will be used if $HOME is not set.
13978 For Windows, we use "C:"; on other platforms, we still require
13979 environment variables.
13981 * Move 'if (!initialized) RAND_poll()' into regions protected by
13982 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13983 having multiple threads call RAND_poll() concurrently.
13987 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13988 combination of a flag and a thread ID variable.
13989 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13990 flag), *other* threads can enter ssleay_add_bytes without obeying
13991 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13992 that they do not hold after the first thread unsets add_do_not_lock).
13996 * Change bctest again: '-x' expressions are not available in all
13997 versions of 'test'.
14001 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14003 * Fix a couple of memory leaks in PKCS7_dataDecode()
14005 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14007 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14008 the default extension for executables, if any. Also, make the perl
14009 scripts that use symlink() to test if it really exists and use "cp"
14010 if it doesn't. All this made OpenSSL compilable and installable in
14015 * Fix for asn1_GetSequence() for indefinite length constructed data.
14016 If SEQUENCE is length is indefinite just set c->slen to the total
14017 amount of data available.
14019 *Steve Henson, reported by shige@FreeBSD.org*
14021 *This change does not apply to 0.9.7.*
14023 * Change bctest to avoid here-documents inside command substitution
14024 (workaround for FreeBSD /bin/sh bug).
14025 For compatibility with Ultrix, avoid shell functions (introduced
14026 in the bctest version that searches along $PATH).
14030 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14031 with des_encrypt() defined on some operating systems, like Solaris
14036 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14037 On the Importance of Eliminating Errors in Cryptographic
14038 Computations, J. Cryptology 14 (2001) 2, 101-119,
14039 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14043 * MIPS assembler BIGNUM division bug fix.
14047 * Disabled incorrect Alpha assembler code.
14051 * Fix PKCS#7 decode routines so they correctly update the length
14052 after reading an EOC for the EXPLICIT tag.
14056 *This change does not apply to 0.9.7.*
14058 * Fix bug in PKCS#12 key generation routines. This was triggered
14059 if a 3DES key was generated with a 0 initial byte. Include
14060 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14061 (but broken) behaviour.
14065 * Enhance bctest to search for a working bc along $PATH and print
14068 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14070 * Fix memory leaks in err.c: free err_data string if necessary;
14071 don't write to the wrong index in ERR_set_error_data.
14075 * Implement ssl23_peek (analogous to ssl23_read), which previously
14080 * Replace rdtsc with `_emit` statements for VC++ version 5.
14082 *Jeremy Cooper <jeremy@baymoo.org>*
14084 * Make it possible to reuse SSLv2 sessions.
14088 * In copy_email() check for >= 0 as a return value for
14089 X509_NAME_get_index_by_NID() since 0 is a valid index.
14091 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14093 * Avoid coredump with unsupported or invalid public keys by checking if
14094 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14095 PKCS7_verify() fails with non detached data.
14099 * Don't use getenv in library functions when run as setuid/setgid.
14100 New function OPENSSL_issetugid().
14104 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14105 due to incorrect handling of multi-threading:
14107 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14109 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14111 3. Count how many times MemCheck_off() has been called so that
14112 nested use can be treated correctly. This also avoids
14113 inband-signalling in the previous code (which relied on the
14114 assumption that thread ID 0 is impossible).
14118 * Add "-rand" option also to s_client and s_server.
14122 * Fix CPU detection on Irix 6.x.
14123 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14124 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14126 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14131 *This change does not apply to 0.9.7.*
14133 * Use the cached encoding of an X509_NAME structure rather than
14134 copying it. This is apparently the reason for the libsafe "errors"
14135 but the code is actually correct.
14139 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14140 Bleichenbacher's DSA attack.
14141 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14142 to be set and top=0 forces the highest bit to be set; top=-1 is new
14143 and leaves the highest bit random.
14145 *Ulf Moeller, Bodo Moeller*
14147 * In the `NCONF_...`-based implementations for `CONF_...` queries
14148 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14149 a temporary CONF structure with the data component set to NULL
14150 (which gives segmentation faults in lh_retrieve).
14151 Instead, use NULL for the CONF pointer in CONF_get_string and
14152 CONF_get_number (which may use environment variables) and directly
14153 return NULL from CONF_get_section.
14157 * Fix potential buffer overrun for EBCDIC.
14161 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14162 keyUsage if basicConstraints absent for a CA.
14166 * Make SMIME_write_PKCS7() write mail header values with a format that
14167 is more generally accepted (no spaces before the semicolon), since
14168 some programs can't parse those values properly otherwise. Also make
14169 sure BIO's that break lines after each write do not create invalid
14174 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14175 macros previously used would not encode an empty SEQUENCE OF
14176 and break the signature.
14180 *This change does not apply to 0.9.7.*
14182 * Zero the premaster secret after deriving the master secret in
14187 * Add some EVP_add_digest_alias registrations (as found in
14188 OpenSSL_add_all_digests()) to SSL_library_init()
14189 aka OpenSSL_add_ssl_algorithms(). This provides improved
14190 compatibility with peers using X.509 certificates
14191 with unconventional AlgorithmIdentifier OIDs.
14195 * Fix for Irix with NO_ASM.
14197 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14199 * ./config script fixes.
14201 *Ulf Moeller, Richard Levitte*
14203 * Fix 'openssl passwd -1'.
14207 * Change PKCS12_key_gen_asc() so it can cope with non null
14208 terminated strings whose length is passed in the passlen
14209 parameter, for example from PEM callbacks. This was done
14210 by adding an extra length parameter to asc2uni().
14212 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14214 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14215 call failed, free the DSA structure.
14219 * Fix to uni2asc() to cope with zero length Unicode strings.
14220 These are present in some PKCS#12 files.
14224 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14225 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14226 when writing a 32767 byte record.
14228 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14230 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14231 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14233 (RSA objects have a reference count access to which is protected
14234 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14235 so they are meant to be shared between threads.)
14236 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14237 "Reddie, Steven" <Steven.Reddie@ca.com>*
14239 * Fix a deadlock in CRYPTO_mem_leaks().
14243 * Use better test patterns in bntest.
14247 * rand_win.c fix for Borland C.
14251 * BN_rshift bugfix for n == 0.
14255 * Add a 'bctest' script that checks for some known 'bc' bugs
14256 so that 'make test' does not abort just because 'bc' is broken.
14260 * Store verify_result within SSL_SESSION also for client side to
14261 avoid potential security hole. (Re-used sessions on the client side
14262 always resulted in verify_result==X509_V_OK, not using the original
14263 result of the server certificate verification.)
14267 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14268 SSL3_RT_APPLICATION_DATA, return 0.
14269 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14274 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14275 releases, have been re-implemented by renaming the previous
14276 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14277 and ssl3_read_internal, respectively, and adding 'peek' parameters
14278 to them. The new ssl[23]_{read,peek} functions are calls to
14279 ssl[23]_read_internal with the 'peek' flag set appropriately.
14280 A 'peek' parameter has also been added to ssl3_read_bytes, which
14281 does the actual work for ssl3_read_internal.
14285 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14286 the method-specific "init()" handler. Also clean up ex_data after
14287 calling the method-specific "finish()" handler. Previously, this was
14288 happening the other way round.
14292 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14293 The previous value, 12, was not always sufficient for BN_mod_exp().
14297 * Make sure that shared libraries get the internal name engine with
14298 the full version number and not just 0. This should mark the
14299 shared libraries as not backward compatible. Of course, this should
14300 be changed again when we can guarantee backward binary compatibility.
14304 * Fix typo in get_cert_by_subject() in by_dir.c
14306 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14308 * Rework the system to generate shared libraries:
14310 - Make note of the expected extension for the shared libraries and
14311 if there is a need for symbolic links from for example libcrypto.so.0
14312 to libcrypto.so.0.9.7. There is extended info in Configure for
14315 - Make as few rebuilds of the shared libraries as possible.
14317 - Still avoid linking the OpenSSL programs with the shared libraries.
14319 - When installing, install the shared libraries separately from the
14324 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14326 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14327 and not in SSL_clear because the latter is also used by the
14328 accept/connect functions; previously, the settings made by
14329 SSL_set_read_ahead would be lost during the handshake.
14331 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14333 * Correct util/mkdef.pl to be selective about disabled algorithms.
14334 Previously, it would create entries for disabled algorithms no
14339 * Added several new manual pages for SSL_* function.
14343 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14345 * In ssl23_get_client_hello, generate an error message when faced
14346 with an initial SSL 3.0/TLS record that is too small to contain the
14347 first two bytes of the ClientHello message, i.e. client_version.
14348 (Note that this is a pathologic case that probably has never happened
14349 in real life.) The previous approach was to use the version number
14350 from the record header as a substitute; but our protocol choice
14351 should not depend on that one because it is not authenticated
14352 by the Finished messages.
14356 * More robust randomness gathering functions for Windows.
14358 *Jeffrey Altman <jaltman@columbia.edu>*
14360 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14361 not set then we don't setup the error code for issuer check errors
14362 to avoid possibly overwriting other errors which the callback does
14363 handle. If an application does set the flag then we assume it knows
14364 what it is doing and can handle the new informational codes
14369 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14370 a general "ANY" type, as such it should be able to decode anything
14371 including tagged types. However it didn't check the class so it would
14372 wrongly interpret tagged types in the same way as their universal
14373 counterpart and unknown types were just rejected. Changed so that the
14374 tagged and unknown types are handled in the same way as a SEQUENCE:
14375 that is the encoding is stored intact. There is also a new type
14376 "V_ASN1_OTHER" which is used when the class is not universal, in this
14377 case we have no idea what the actual type is so we just lump them all
14382 * On VMS, stdout may very well lead to a file that is written to
14383 in a record-oriented fashion. That means that every write() will
14384 write a separate record, which will be read separately by the
14385 programs trying to read from it. This can be very confusing.
14387 The solution is to put a BIO filter in the way that will buffer
14388 text until a linefeed is reached, and then write everything a
14389 line at a time, so every record written will be an actual line,
14390 not chunks of lines and not (usually doesn't happen, but I've
14391 seen it once) several lines in one record. BIO_f_linebuffer() is
14394 Currently, it's a VMS-only method, because that's where it has
14395 been tested well enough.
14399 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14400 it can return incorrect results.
14401 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14402 but it was in 0.9.6-beta[12].)
14406 * Disable the check for content being present when verifying detached
14407 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14408 include zero length content when signing messages.
14412 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14413 BIO_ctrl (for BIO pairs).
14417 * Add DSO method for VMS.
14421 * Bug fix: Montgomery multiplication could produce results with the
14426 * Add RPM specification openssl.spec and modify it to build three
14427 packages. The default package contains applications, application
14428 documentation and run-time libraries. The devel package contains
14429 include files, static libraries and function documentation. The
14430 doc package contains the contents of the doc directory. The original
14431 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14435 * Add a large number of documentation files for many SSL routines.
14437 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14439 * Add a configuration entry for Sony News 4.
14441 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14443 * Don't set the two most significant bits to one when generating a
14444 random number < q in the DSA library.
14448 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14449 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14450 the underlying transport is blocking) if a handshake took place.
14451 (The default behaviour is needed by applications such as s_client
14452 and s_server that use select() to determine when to use SSL_read;
14453 but for applications that know in advance when to expect data, it
14454 just makes things more complicated.)
14458 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14463 * Add a few more EBCDIC conditionals that make `req` and `x509`
14464 work better on such systems.
14466 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14468 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14469 Update PKCS12_parse() so it copies the friendlyName and the
14470 keyid to the certificates aux info.
14474 * Fix bug in PKCS7_verify() which caused an infinite loop
14475 if there was more than one signature.
14477 *Sven Uszpelkat <su@celocom.de>*
14479 * Major change in util/mkdef.pl to include extra information
14480 about each symbol, as well as presenting variables as well
14481 as functions. This change means that there's n more need
14482 to rebuild the .num files when some algorithms are excluded.
14486 * Allow the verify time to be set by an application,
14487 rather than always using the current time.
14491 * Phase 2 verify code reorganisation. The certificate
14492 verify code now looks up an issuer certificate by a
14493 number of criteria: subject name, authority key id
14494 and key usage. It also verifies self signed certificates
14495 by the same criteria. The main comparison function is
14496 X509_check_issued() which performs these checks.
14498 Lot of changes were necessary in order to support this
14499 without completely rewriting the lookup code.
14501 Authority and subject key identifier are now cached.
14503 The LHASH 'certs' is X509_STORE has now been replaced
14504 by a STACK_OF(X509_OBJECT). This is mainly because an
14505 LHASH can't store or retrieve multiple objects with
14506 the same hash value.
14508 As a result various functions (which were all internal
14509 use only) have changed to handle the new X509_STORE
14510 structure. This will break anything that messed round
14511 with X509_STORE internally.
14513 The functions X509_STORE_add_cert() now checks for an
14514 exact match, rather than just subject name.
14516 The X509_STORE API doesn't directly support the retrieval
14517 of multiple certificates matching a given criteria, however
14518 this can be worked round by performing a lookup first
14519 (which will fill the cache with candidate certificates)
14520 and then examining the cache for matches. This is probably
14521 the best we can do without throwing out X509_LOOKUP
14522 entirely (maybe later...).
14524 The X509_VERIFY_CTX structure has been enhanced considerably.
14526 All certificate lookup operations now go via a get_issuer()
14527 callback. Although this currently uses an X509_STORE it
14528 can be replaced by custom lookups. This is a simple way
14529 to bypass the X509_STORE hackery necessary to make this
14530 work and makes it possible to use more efficient techniques
14531 in future. A very simple version which uses a simple
14532 STACK for its trusted certificate store is also provided
14533 using X509_STORE_CTX_trusted_stack().
14535 The verify_cb() and verify() callbacks now have equivalents
14536 in the X509_STORE_CTX structure.
14538 X509_STORE_CTX also has a 'flags' field which can be used
14539 to customise the verify behaviour.
14543 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14544 excludes S/MIME capabilities.
14548 * When a certificate request is read in keep a copy of the
14549 original encoding of the signed data and use it when outputting
14550 again. Signatures then use the original encoding rather than
14551 a decoded, encoded version which may cause problems if the
14552 request is improperly encoded.
14556 * For consistency with other BIO_puts implementations, call
14557 buffer_write(b, ...) directly in buffer_puts instead of calling
14560 In BIO_puts, increment b->num_write as in BIO_write.
14562 *Peter.Sylvester@EdelWeb.fr*
14564 * Fix BN_mul_word for the case where the word is 0. (We have to use
14565 BN_zero, we may not return a BIGNUM with an array consisting of
14566 words set to zero.)
14570 * Avoid calling abort() from within the library when problems are
14571 detected, except if preprocessor symbols have been defined
14572 (such as REF_CHECK, BN_DEBUG etc.).
14576 * New openssl application 'rsautl'. This utility can be
14577 used for low level RSA operations. DER public key
14578 BIO/fp routines also added.
14582 * New Configure entry and patches for compiling on QNX 4.
14584 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14586 * A demo state-machine implementation was sponsored by
14587 Nuron (<http://www.nuron.com/>) and is now available in
14588 demos/state_machine.
14592 * New options added to the 'dgst' utility for signature
14593 generation and verification.
14597 * Unrecognized PKCS#7 content types are now handled via a
14598 catch all ASN1_TYPE structure. This allows unsupported
14599 types to be stored as a "blob" and an application can
14600 encode and decode it manually.
14604 * Fix various signed/unsigned issues to make a_strex.c
14605 compile under VC++.
14607 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14609 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14610 length if passed a buffer. ASN1_INTEGER_to_BN failed
14611 if passed a NULL BN and its argument was negative.
14613 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14615 * Modification to PKCS#7 encoding routines to output definite
14616 length encoding. Since currently the whole structures are in
14617 memory there's not real point in using indefinite length
14618 constructed encoding. However if OpenSSL is compiled with
14619 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14623 * Added BIO_vprintf() and BIO_vsnprintf().
14627 * Added more prefixes to parse for in the strings written
14628 through a logging bio, to cover all the levels that are available
14629 through syslog. The prefixes are now:
14631 PANIC, EMERG, EMR => LOG_EMERG
14632 ALERT, ALR => LOG_ALERT
14633 CRIT, CRI => LOG_CRIT
14634 ERROR, ERR => LOG_ERR
14635 WARNING, WARN, WAR => LOG_WARNING
14636 NOTICE, NOTE, NOT => LOG_NOTICE
14637 INFO, INF => LOG_INFO
14638 DEBUG, DBG => LOG_DEBUG
14640 and as before, if none of those prefixes are present at the
14641 beginning of the string, LOG_ERR is chosen.
14643 On Win32, the `LOG_*` levels are mapped according to this:
14645 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14646 LOG_WARNING => EVENTLOG_WARNING_TYPE
14647 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14651 * Made it possible to reconfigure with just the configuration
14652 argument "reconf" or "reconfigure". The command line arguments
14653 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14654 and are retrieved from there when reconfiguring.
14660 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14662 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14666 * The obj_dat.pl script was messing up the sorting of object
14667 names. The reason was that it compared the quoted version
14668 of strings as a result "OCSP" > "OCSP Signing" because
14669 " > SPACE. Changed script to store unquoted versions of
14670 names and add quotes on output. It was also omitting some
14671 names from the lookup table if they were given a default
14672 value (that is if SN is missing it is given the same
14673 value as LN and vice versa), these are now added on the
14674 grounds that if an object has a name we should be able to
14675 look it up. Finally added warning output when duplicate
14676 short or long names are found.
14680 * Changes needed for Tandem NSK.
14682 *Scott Uroff <scott@xypro.com>*
14684 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14685 RSA_padding_check_SSLv23(), special padding was never detected
14686 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14687 version rollback attacks was not effective.
14689 In s23_clnt.c, don't use special rollback-attack detection padding
14690 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14691 client; similarly, in s23_srvr.c, don't do the rollback check if
14692 SSL 2.0 is the only protocol enabled in the server.
14696 * Make it possible to get hexdumps of unprintable data with 'openssl
14697 asn1parse'. By implication, the functions ASN1_parse_dump() and
14698 BIO_dump_indent() are added.
14702 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14703 these print out strings and name structures based on various
14704 flags including RFC2253 support and proper handling of
14705 multibyte characters. Added options to the 'x509' utility
14706 to allow the various flags to be set.
14710 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14711 Also change the functions X509_cmp_current_time() and
14712 X509_gmtime_adj() work with an ASN1_TIME structure,
14713 this will enable certificates using GeneralizedTime in validity
14714 dates to be checked.
14718 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14719 negative public key encodings) on by default,
14720 NO_NEG_PUBKEY_BUG can be set to disable it.
14724 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14725 content octets. An i2c_ASN1_OBJECT is unnecessary because
14726 the encoding can be trivially obtained from the structure.
14730 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14731 not read locks (`CRYPTO_r_[un]lock`).
14735 * A first attempt at creating official support for shared
14736 libraries through configuration. I've kept it so the
14737 default is static libraries only, and the OpenSSL programs
14738 are always statically linked for now, but there are
14739 preparations for dynamic linking in place.
14740 This has been tested on Linux and Tru64.
14744 * Randomness polling function for Win9x, as described in:
14745 Peter Gutmann, Software Generation of Practically Strong
14750 * Fix so PRNG is seeded in req if using an already existing
14755 * New options to smime application. -inform and -outform
14756 allow alternative formats for the S/MIME message including
14757 PEM and DER. The -content option allows the content to be
14758 specified separately. This should allow things like Netscape
14759 form signing output easier to verify.
14763 * Fix the ASN1 encoding of tags using the 'long form'.
14767 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14768 STRING types. These convert content octets to and from the
14769 underlying type. The actual tag and length octets are
14770 already assumed to have been read in and checked. These
14771 are needed because all other string types have virtually
14772 identical handling apart from the tag. By having versions
14773 of the ASN1 functions that just operate on content octets
14774 IMPLICIT tagging can be handled properly. It also allows
14775 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14776 and ASN1_INTEGER are identical apart from the tag.
14780 * Change the handling of OID objects as follows:
14782 - New object identifiers are inserted in objects.txt, following
14783 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14784 - objects.pl is used to process obj_mac.num and create a new
14786 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14789 This is currently kind of a hack, and the perl code in objects.pl
14790 isn't very elegant, but it works as I intended. The simplest way
14791 to check that it worked correctly is to look in obj_dat.h and
14792 check the array nid_objs and make sure the objects haven't moved
14793 around (this is important!). Additions are OK, as well as
14794 consistent name changes.
14798 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14802 * Addition of the command line parameter '-rand file' to 'openssl req'.
14803 The given file adds to whatever has already been seeded into the
14804 random pool through the RANDFILE configuration file option or
14805 environment variable, or the default random state file.
14809 * mkstack.pl now sorts each macro group into lexical order.
14810 Previously the output order depended on the order the files
14811 appeared in the directory, resulting in needless rewriting
14816 * Patches to make OpenSSL compile under Win32 again. Mostly
14817 work arounds for the VC++ problem that it treats func() as
14818 func(void). Also stripped out the parts of mkdef.pl that
14819 added extra typesafe functions: these no longer exist.
14823 * Reorganisation of the stack code. The macros are now all
14824 collected in safestack.h . Each macro is defined in terms of
14825 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14826 DEBUG_SAFESTACK is now handled in terms of function casts,
14827 this has the advantage of retaining type safety without the
14828 use of additional functions. If DEBUG_SAFESTACK is not defined
14829 then the non typesafe macros are used instead. Also modified the
14830 mkstack.pl script to handle the new form. Needs testing to see
14831 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14832 the default if no major problems. Similar behaviour for ASN1_SET_OF
14833 and PKCS12_STACK_OF.
14837 * When some versions of IIS use the 'NET' form of private key the
14838 key derivation algorithm is different. Normally MD5(password) is
14839 used as a 128 bit RC4 key. In the modified case
14840 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14841 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14842 as the old Netscape_RSA functions except they have an additional
14843 'sgckey' parameter which uses the modified algorithm. Also added
14844 an -sgckey command line option to the rsa utility. Thanks to
14845 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14846 algorithm to openssl-dev.
14850 * The evp_local.h macros were using 'c.##kname' which resulted in
14851 invalid expansion on some systems (SCO 5.0.5 for example).
14852 Corrected to 'c.kname'.
14854 *Phillip Porch <root@theporch.com>*
14856 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14857 a STACK of email addresses from a certificate or request, these look
14858 in the subject name and the subject alternative name extensions and
14859 omit any duplicate addresses.
14863 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14864 This makes DSA verification about 2 % faster.
14868 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14869 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14870 plus overhead for 1024 bit moduli).
14871 This makes exponentiations about 0.5 % faster for 1024 bit
14872 exponents (as measured by "openssl speed rsa2048").
14876 * Rename memory handling macros to avoid conflicts with other
14878 Malloc => OPENSSL_malloc
14879 Malloc_locked => OPENSSL_malloc_locked
14880 Realloc => OPENSSL_realloc
14881 Free => OPENSSL_free
14885 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14886 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14890 * CygWin32 support.
14892 *John Jarvie <jjarvie@newsguy.com>*
14894 * The type-safe stack code has been rejigged. It is now only compiled
14895 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14896 by default all type-specific stack functions are "#define"d back to
14897 standard stack functions. This results in more streamlined output
14898 but retains the type-safety checking possibilities of the original
14903 * The STACK code has been cleaned up, and certain type declarations
14904 that didn't make a lot of sense have been brought in line. This has
14905 also involved a cleanup of sorts in safestack.h to more correctly
14906 map type-safe stack functions onto their plain stack counterparts.
14907 This work has also resulted in a variety of "const"ifications of
14908 lots of the code, especially `_cmp` operations which should normally
14909 be prototyped with "const" parameters anyway.
14913 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14914 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14915 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14916 where all of 'md' is used each time the PRNG is used, but 'state'
14917 is used only indexed by a cyclic counter. As entropy may not be
14918 well distributed from the beginning, 'md' is important as a
14919 chaining variable. However, the output function chains only half
14920 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14921 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14922 in all of 'state' being rewritten, with the new values depending
14923 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14927 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14928 the handshake is continued after ssl_verify_cert_chain();
14929 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14930 can lead to 'unexplainable' connection aborts later.
14932 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14934 * Major EVP API cipher revision.
14935 Add hooks for extra EVP features. This allows various cipher
14936 parameters to be set in the EVP interface. Support added for variable
14937 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14938 setting of RC2 and RC5 parameters.
14940 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14943 Remove lots of duplicated code from the EVP library. For example *every*
14944 cipher init() function handles the 'iv' in the same way according to the
14945 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14946 for CFB and OFB modes they zero ctx->num.
14948 New functionality allows removal of S/MIME code RC2 hack.
14950 Most of the routines have the same form and so can be declared in terms
14953 By shifting this to the top level EVP_CipherInit() it can be removed from
14954 all individual ciphers. If the cipher wants to handle IVs or keys
14955 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14958 Change lots of functions like EVP_EncryptUpdate() to now return a
14959 value: although software versions of the algorithms cannot fail
14960 any installed hardware versions can.
14964 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14965 this option is set, tolerate broken clients that send the negotiated
14966 protocol version number instead of the requested protocol version
14971 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14972 i.e. non-zero for export ciphersuites, zero otherwise.
14973 Previous versions had this flag inverted, inconsistent with
14974 rsa_tmp_cb (..._TMP_RSA_CB).
14976 *Bodo Moeller; problem reported by Amit Chopra*
14978 * Add missing DSA library text string. Work around for some IIS
14979 key files with invalid SEQUENCE encoding.
14983 * Add a document (doc/standards.txt) that list all kinds of standards
14984 and so on that are implemented in OpenSSL.
14988 * Enhance c_rehash script. Old version would mishandle certificates
14989 with the same subject name hash and wouldn't handle CRLs at all.
14990 Added -fingerprint option to crl utility, to support new c_rehash
14995 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14999 * Fix for SSL server purpose checking. Server checking was
15000 rejecting certificates which had extended key usage present
15001 but no ssl client purpose.
15003 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15005 * Make PKCS#12 code work with no password. The PKCS#12 spec
15006 is a little unclear about how a blank password is handled.
15007 Since the password in encoded as a BMPString with terminating
15008 double NULL a zero length password would end up as just the
15009 double NULL. However no password at all is different and is
15010 handled differently in the PKCS#12 key generation code. NS
15011 treats a blank password as zero length. MSIE treats it as no
15012 password on export: but it will try both on import. We now do
15013 the same: PKCS12_parse() tries zero length and no password if
15014 the password is set to "" or NULL (NULL is now a valid password:
15015 it wasn't before) as does the pkcs12 application.
15019 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15020 perror when PEM_read_bio_X509_REQ fails, the error message must
15021 be obtained from the error queue.
15025 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15026 it in ERR_remove_state if appropriate, and change ERR_get_state
15027 accordingly to avoid race conditions (this is necessary because
15028 thread_hash is no longer constant once set).
15032 * Bugfix for linux-elf makefile.one.
15036 * RSA_get_default_method() will now cause a default
15037 RSA_METHOD to be chosen if one doesn't exist already.
15038 Previously this was only set during a call to RSA_new()
15039 or RSA_new_method(NULL) meaning it was possible for
15040 RSA_get_default_method() to return NULL.
15044 * Added native name translation to the existing DSO code
15045 that will convert (if the flag to do so is set) filenames
15046 that are sufficiently small and have no path information
15047 into a canonical native form. Eg. "blah" converted to
15048 "libblah.so" or "blah.dll" etc.
15052 * New function ERR_error_string_n(e, buf, len) which is like
15053 ERR_error_string(e, buf), but writes at most 'len' bytes
15054 including the 0 terminator. For ERR_error_string_n, 'buf'
15057 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15059 * CONF library reworked to become more general. A new CONF
15060 configuration file reader "class" is implemented as well as a
15061 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15062 old `CONF_*` functions are still there, but are reimplemented to
15063 work in terms of the new functions. Also, a set of functions
15064 to handle the internal storage of the configuration data is
15065 provided to make it easier to write new configuration file
15066 reader "classes" (I can definitely see something reading a
15067 configuration file in XML format, for example), called `_CONF_*`,
15068 or "the configuration storage API"...
15070 The new configuration file reading functions are:
15072 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15073 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15075 NCONF_default, NCONF_WIN32
15077 NCONF_dump_fp, NCONF_dump_bio
15079 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15080 NCONF_new creates a new CONF object. This works in the same way
15081 as other interfaces in OpenSSL, like the BIO interface.
15082 `NCONF_dump_*` dump the internal storage of the configuration file,
15083 which is useful for debugging. All other functions take the same
15084 arguments as the old `CONF_*` functions with the exception of the
15085 first that must be a `CONF *` instead of a `LHASH *`.
15087 To make it easier to use the new classes with the old `CONF_*` functions,
15088 the function CONF_set_default_method is provided.
15092 * Add '-tls1' option to 'openssl ciphers', which was already
15093 mentioned in the documentation but had not been implemented.
15094 (This option is not yet really useful because even the additional
15095 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15099 * Initial DSO code added into libcrypto for letting OpenSSL (and
15100 OpenSSL-based applications) load shared libraries and bind to
15101 them in a portable way.
15103 *Geoff Thorpe, with contributions from Richard Levitte*
15105 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15107 * Make sure _lrotl and _lrotr are only used with MSVC.
15109 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15110 (the default implementation of RAND_status).
15112 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15113 to '-clrext' (= clear extensions), as intended and documented.
15114 *Bodo Moeller; inconsistency pointed out by Michael Attili
15115 <attili@amaxo.com>*
15117 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15118 was larger than the MD block size.
15120 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15122 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15123 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15124 using the passed key: if the passed key was a private key the result
15125 of X509_print(), for example, would be to print out all the private key
15130 * des_quad_cksum() byte order bug fix.
15131 *Ulf Möller, using the problem description in krb4-0.9.7, where
15132 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15134 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15137 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15139 * For easily testing in shell scripts whether some command
15140 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15141 returns with exit code 0 iff no command of the given name is available.
15142 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15143 the output goes to stdout and nothing is printed to stderr.
15144 Additional arguments are always ignored.
15146 Since for each cipher there is a command of the same name,
15147 the 'no-cipher' compilation switches can be tested this way.
15149 ('openssl no-XXX' is not able to detect pseudo-commands such
15150 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15154 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15158 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15159 is set; it will be thrown away anyway because each handshake creates
15161 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15162 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15163 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15164 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15168 * New s_client option -ign_eof: EOF at stdin is ignored, and
15169 'Q' and 'R' lose their special meanings (quit/renegotiate).
15170 This is part of what -quiet does; unlike -quiet, -ign_eof
15171 does not suppress any output.
15175 * Add compatibility options to the purpose and trust code. The
15176 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15177 accepts a certificate or CA, this was the previous behaviour,
15178 with all the associated security issues.
15180 X509_TRUST_COMPAT is the old trust behaviour: only and
15181 automatically trust self signed roots in certificate store. A
15182 new trust setting X509_TRUST_DEFAULT is used to specify that
15183 a purpose has no associated trust setting and it should instead
15184 use the value in the default purpose.
15188 * Fix the PKCS#8 DSA private key code so it decodes keys again
15189 and fix a memory leak.
15193 * In util/mkerr.pl (which implements 'make errors'), preserve
15194 reason strings from the previous version of the .c file, as
15195 the default to have only downcase letters (and digits) in
15196 automatically generated reasons codes is not always appropriate.
15200 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15201 using strerror. Previously, ERR_reason_error_string() returned
15202 library names as reason strings for SYSerr; but SYSerr is a special
15203 case where small numbers are errno values, not library numbers.
15207 * Add '-dsaparam' option to 'openssl dhparam' application. This
15208 converts DSA parameters into DH parameters. (When creating parameters,
15209 DSA_generate_parameters is used.)
15213 * Include 'length' (recommended exponent length) in C code generated
15214 by 'openssl dhparam -C'.
15218 * The second argument to set_label in perlasm was already being used
15219 so couldn't be used as a "file scope" flag. Moved to third argument
15224 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15225 instead of RAND_bytes for encryption IVs and salts.
15229 * Include RAND_status() into RAND_METHOD instead of implementing
15230 it only for md_rand.c Otherwise replacing the PRNG by calling
15231 RAND_set_rand_method would be impossible.
15235 * Don't let DSA_generate_key() enter an infinite loop if the random
15236 number generation fails.
15240 * New 'rand' application for creating pseudo-random output.
15244 * Added configuration support for Linux/IA64
15246 *Rolf Haberrecker <rolf@suse.de>*
15248 * Assembler module support for Mingw32.
15252 * Shared library support for HPUX (in shlib/).
15254 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15256 * Shared library support for Solaris gcc.
15258 *Lutz Behnke <behnke@trustcenter.de>*
15260 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15262 * PKCS7_encrypt() was adding text MIME headers twice because they
15263 were added manually and by SMIME_crlf_copy().
15267 * In bntest.c don't call BN_rand with zero bits argument.
15269 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15271 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15272 case was implemented. This caused BN_div_recp() to fail occasionally.
15276 * Add an optional second argument to the set_label() in the perl
15277 assembly language builder. If this argument exists and is set
15278 to 1 it signals that the assembler should use a symbol whose
15279 scope is the entire file, not just the current function. This
15280 is needed with MASM which uses the format label:: for this scope.
15282 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15284 * Change the ASN1 types so they are typedefs by default. Before
15285 almost all types were #define'd to ASN1_STRING which was causing
15286 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15291 * Change names of new functions to the new get1/get0 naming
15292 convention: After 'get1', the caller owns a reference count
15293 and has to call `..._free`; 'get0' returns a pointer to some
15294 data structure without incrementing reference counters.
15295 (Some of the existing 'get' functions increment a reference
15296 counter, some don't.)
15297 Similarly, 'set1' and 'add1' functions increase reference
15298 counters or duplicate objects.
15302 * Allow for the possibility of temp RSA key generation failure:
15303 the code used to assume it always worked and crashed on failure.
15307 * Fix potential buffer overrun problem in BIO_printf().
15308 *Ulf Möller, using public domain code by Patrick Powell; problem
15309 pointed out by David Sacerdote <das33@cornell.edu>*
15311 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15312 RAND_egd() and RAND_status(). In the command line application,
15313 the EGD socket can be specified like a seed file using RANDFILE
15318 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15319 Some CAs (e.g. Verisign) distribute certificates in this form.
15323 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15324 list to exclude them. This means that no special compilation option
15325 is needed to use anonymous DH: it just needs to be included in the
15330 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15331 EVP_MD_type. The old functionality is available in a new macro called
15332 EVP_MD_md(). Change code that uses it and update docs.
15336 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15337 where the `void *` argument is replaced by a function pointer argument.
15338 Previously `void *` was abused to point to functions, which works on
15339 many platforms, but is not correct. As these functions are usually
15340 called by macros defined in OpenSSL header files, most source code
15341 should work without changes.
15345 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15346 sections with information on -D... compiler switches used for
15347 compiling the library so that applications can see them. To enable
15348 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15349 must be defined. E.g.,
15350 #define OPENSSL_ALGORITHM_DEFINES
15351 #include <openssl/opensslconf.h>
15352 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15354 *Richard Levitte, Ulf and Bodo Möller*
15356 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15361 * Change the 'other' type in certificate aux info to a STACK_OF
15362 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15363 the required ASN1 format: arbitrary types determined by an OID.
15367 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15368 argument to 'req'. This is not because the function is newer or
15369 better than others it just uses the work 'NEW' in the certificate
15370 request header lines. Some software needs this.
15374 * Reorganise password command line arguments: now passwords can be
15375 obtained from various sources. Delete the PEM_cb function and make
15376 it the default behaviour: i.e. if the callback is NULL and the
15377 usrdata argument is not NULL interpret it as a null terminated pass
15378 phrase. If usrdata and the callback are NULL then the pass phrase
15379 is prompted for as usual.
15383 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15384 the support is automatically enabled. The resulting binaries will
15385 autodetect the card and use it if present.
15387 *Ben Laurie and Compaq Inc.*
15389 * Work around for Netscape hang bug. This sends certificate request
15390 and server done in one record. Since this is perfectly legal in the
15391 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15392 the bugs/SSLv3 entry for more info.
15396 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15400 * Add -rand argument to smime and pkcs12 applications and read/write
15405 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15409 * Add command line password options to the remaining applications.
15413 * Bug fix for BN_div_recp() for numerators with an even number of
15418 * More tests in bntest.c, and changed test_bn output.
15422 * ./config recognizes MacOS X now.
15426 * Bug fix for BN_div() when the first words of num and divisor are
15427 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15431 * Add support for various broken PKCS#8 formats, and command line
15432 options to produce them.
15436 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15437 get temporary BIGNUMs from a BN_CTX.
15441 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15446 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15447 include a #define from the old name to the new. The original intent
15448 was that statically linked binaries could for example just call
15449 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15450 link with digests. This never worked because SSLeay_add_all_digests()
15451 and SSLeay_add_all_ciphers() were in the same source file so calling
15452 one would link with the other. They are now in separate source files.
15456 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15460 * Use a less unusual form of the Miller-Rabin primality test (it used
15461 a binary algorithm for exponentiation integrated into the Miller-Rabin
15462 loop, our standard modexp algorithms are faster).
15466 * Support for the EBCDIC character set completed.
15468 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15470 * Source code cleanups: use const where appropriate, eliminate casts,
15471 use `void *` instead of `char *` in lhash.
15475 * Bugfix: ssl3_send_server_key_exchange was not restartable
15476 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15477 this the server could overwrite ephemeral keys that the client
15482 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15483 using 50 iterations of the Rabin-Miller test.
15485 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15486 iterations of the Rabin-Miller test as required by the appendix
15487 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15488 As BN_is_prime_fasttest includes trial division, DSA parameter
15489 generation becomes much faster.
15491 This implies a change for the callback functions in DSA_is_prime
15492 and DSA_generate_parameters: The callback function is called once
15493 for each positive witness in the Rabin-Miller test, not just
15494 occasionally in the inner loop; and the parameters to the
15495 callback function now provide an iteration count for the outer
15496 loop rather than for the current invocation of the inner loop.
15497 DSA_generate_parameters additionally can call the callback
15498 function with an 'iteration count' of -1, meaning that a
15499 candidate has passed the trial division test (when q is generated
15500 from an application-provided seed, trial division is skipped).
15504 * New function BN_is_prime_fasttest that optionally does trial
15505 division before starting the Rabin-Miller test and has
15506 an additional BN_CTX * argument (whereas BN_is_prime always
15507 has to allocate at least one BN_CTX).
15508 'callback(1, -1, cb_arg)' is called when a number has passed the
15509 trial division stage.
15513 * Fix for bug in CRL encoding. The validity dates weren't being handled
15518 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15522 * New function BN_pseudo_rand().
15526 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15527 bignum version of BN_from_montgomery() with the working code from
15528 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15533 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15534 made it impossible to use the same SSL_SESSION data structure in
15535 SSL2 clients in multiple threads.
15539 * The return value of RAND_load_file() no longer counts bytes obtained
15540 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15541 to seed the PRNG (previously an explicit byte count was required).
15543 *Ulf Möller, Bodo Möller*
15545 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15546 used `char *` instead of `void *` and had casts all over the place.
15550 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15554 * Retain source code compatibility for BN_prime_checks macro:
15555 BN_is_prime(..., BN_prime_checks, ...) now uses
15556 BN_prime_checks_for_size to determine the appropriate number of
15557 Rabin-Miller iterations.
15561 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15562 DH_CHECK_P_NOT_SAFE_PRIME.
15563 (Check if this is true? OpenPGP calls them "strong".)
15567 * Merge the functionality of "dh" and "gendh" programs into a new program
15568 "dhparam". The old programs are retained for now but will handle DH keys
15569 (instead of parameters) in future.
15573 * Make the ciphers, s_server and s_client programs check the return values
15574 when a new cipher list is set.
15578 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15579 ciphers. Before when the 56bit ciphers were enabled the sorting was
15582 The syntax for the cipher sorting has been extended to support sorting by
15583 cipher-strength (using the strength_bits hard coded in the tables).
15584 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15586 Fix a bug in the cipher-command parser: when supplying a cipher command
15587 string with an "undefined" symbol (neither command nor alphanumeric
15588 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15589 an error is flagged.
15591 Due to the strength-sorting extension, the code of the
15592 ssl_create_cipher_list() function was completely rearranged. I hope that
15593 the readability was also increased :-)
15595 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15597 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15598 for the first serial number and places 2 in the serial number file. This
15599 avoids problems when the root CA is created with serial number zero and
15600 the first user certificate has the same issuer name and serial number
15605 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15606 the new code. Add documentation for this stuff.
15610 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15611 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15612 structures and behave in an analogous way to the X509v3 functions:
15613 they shouldn't be called directly but wrapper functions should be used
15616 So we also now have some wrapper functions that call the X509at functions
15617 when passed certificate requests. (TO DO: similar things can be done with
15618 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15619 things. Some of these need some d2i or i2d and print functionality
15620 because they handle more complex structures.)
15624 * Add missing #ifndefs that caused missing symbols when building libssl
15625 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15626 NO_RSA in `ssl/s2*.c`.
15628 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15630 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15631 has a return value which indicates the quality of the random data
15632 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15633 error queue. New function RAND_pseudo_bytes() generates output that is
15634 guaranteed to be unique but not unpredictable. RAND_add is like
15635 RAND_seed, but takes an extra argument for an entropy estimate
15636 (RAND_seed always assumes full entropy).
15640 * Do more iterations of Rabin-Miller probable prime test (specifically,
15641 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15642 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15643 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15644 false-positive rate of at most 2^-80 for random input.
15648 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15652 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15653 in the 0.9.5 release), this returns the chain
15654 from an X509_CTX structure with a dup of the stack and all
15655 the X509 reference counts upped: so the stack will exist
15656 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15659 Also make SSL_SESSION_print() print out the verify return
15664 * Add manpage for the pkcs12 command. Also change the default
15665 behaviour so MAC iteration counts are used unless the new
15666 -nomaciter option is used. This improves file security and
15667 only older versions of MSIE (4.0 for example) need it.
15671 * Honor the no-xxx Configure options when creating .DEF files.
15675 * Add PKCS#10 attributes to field table: challengePassword,
15676 unstructuredName and unstructuredAddress. These are taken from
15677 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15678 international characters are used.
15680 More changes to X509_ATTRIBUTE code: allow the setting of types
15681 based on strings. Remove the 'loc' parameter when adding
15682 attributes because these will be a SET OF encoding which is sorted
15687 * Initial changes to the 'req' utility to allow request generation
15688 automation. This will allow an application to just generate a template
15689 file containing all the field values and have req construct the
15692 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15693 used all over the place including certificate requests and PKCS#7
15694 structures. They are currently handled manually where necessary with
15695 some primitive wrappers for PKCS#7. The new functions behave in a
15696 manner analogous to the X509 extension functions: they allow
15697 attributes to be looked up by NID and added.
15699 Later something similar to the X509V3 code would be desirable to
15700 automatically handle the encoding, decoding and printing of the
15701 more complex types. The string types like challengePassword can
15702 be handled by the string table functions.
15704 Also modified the multi byte string table handling. Now there is
15705 a 'global mask' which masks out certain types. The table itself
15706 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15707 is useful when for example there is only one permissible type
15708 (as in countryName) and using the mask might result in no valid
15713 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15714 SSL_get_peer_finished to allow applications to obtain the latest
15715 Finished messages sent to the peer or expected from the peer,
15716 respectively. (SSL_get_peer_finished is usually the Finished message
15717 actually received from the peer, otherwise the protocol will be aborted.)
15719 As the Finished message are message digests of the complete handshake
15720 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15721 be used for external authentication procedures when the authentication
15722 provided by SSL/TLS is not desired or is not enough.
15726 * Enhanced support for Alpha Linux is added. Now ./config checks if
15727 the host supports BWX extension and if Compaq C is present on the
15728 $PATH. Just exploiting of the BWX extension results in 20-30%
15729 performance kick for some algorithms, e.g. DES and RC4 to mention
15730 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15735 * Add support for MS "fast SGC". This is arguably a violation of the
15736 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15737 weak crypto and after checking the certificate is SGC a second one
15738 with strong crypto. MS SGC stops the first handshake after receiving
15739 the server certificate message and sends a second client hello. Since
15740 a server will typically do all the time consuming operations before
15741 expecting any further messages from the client (server key exchange
15742 is the most expensive) there is little difference between the two.
15744 To get OpenSSL to support MS SGC we have to permit a second client
15745 hello message after we have sent server done. In addition we have to
15746 reset the MAC if we do get this second client hello.
15750 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15751 if a DER encoded private key is RSA or DSA traditional format. Changed
15752 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15753 format DER encoded private key. Newer code should use PKCS#8 format which
15754 has the key type encoded in the ASN1 structure. Added DER private key
15755 support to pkcs8 application.
15759 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15760 ciphersuites has been selected (as required by the SSL 3/TLS 1
15761 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15762 is set, we interpret this as a request to violate the specification
15763 (the worst that can happen is a handshake failure, and 'correct'
15764 behaviour would result in a handshake failure anyway).
15768 * In SSL_CTX_add_session, take into account that there might be multiple
15769 SSL_SESSION structures with the same session ID (e.g. when two threads
15770 concurrently obtain them from an external cache).
15771 The internal cache can handle only one SSL_SESSION with a given ID,
15772 so if there's a conflict, we now throw out the old one to achieve
15777 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15778 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15779 some routines that use cipher OIDs: some ciphers do not have OIDs
15780 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15785 * Simplify the trust setting structure and code. Now we just have
15786 two sequences of OIDs for trusted and rejected settings. These will
15787 typically have values the same as the extended key usage extension
15788 and any application specific purposes.
15790 The trust checking code now has a default behaviour: it will just
15791 check for an object with the same NID as the passed id. Functions can
15792 be provided to override either the default behaviour or the behaviour
15793 for a given id. SSL client, server and email already have functions
15794 in place for compatibility: they check the NID and also return "trusted"
15795 if the certificate is self signed.
15799 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15800 traditional format into an EVP_PKEY structure.
15804 * Add a password callback function PEM_cb() which either prompts for
15805 a password if usr_data is NULL or otherwise assumes it is a null
15806 terminated password. Allow passwords to be passed on command line
15807 environment or config files in a few more utilities.
15811 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15812 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15813 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15814 Update documentation.
15818 * Support for ASN1 "NULL" type. This could be handled before by using
15819 ASN1_TYPE but there wasn't any function that would try to read a NULL
15820 and produce an error if it couldn't. For compatibility we also have
15821 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15822 don't allocate anything because they don't need to.
15826 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15829 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15831 * Rebuild of the memory allocation routines used by OpenSSL code and
15832 possibly others as well. The purpose is to make an interface that
15833 provide hooks so anyone can build a separate set of allocation and
15834 deallocation routines to be used by OpenSSL, for example memory
15835 pool implementations, or something else, which was previously hard
15836 since Malloc(), Realloc() and Free() were defined as macros having
15837 the values malloc, realloc and free, respectively (except for Win32
15838 compilations). The same is provided for memory debugging code.
15839 OpenSSL already comes with functionality to find memory leaks, but
15840 this gives people a chance to debug other memory problems.
15842 With these changes, a new set of functions and macros have appeared:
15844 CRYPTO_set_mem_debug_functions() [F]
15845 CRYPTO_get_mem_debug_functions() [F]
15846 CRYPTO_dbg_set_options() [F]
15847 CRYPTO_dbg_get_options() [F]
15848 CRYPTO_malloc_debug_init() [M]
15850 The memory debug functions are NULL by default, unless the library
15851 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15852 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15853 gives the standard debugging functions that come with OpenSSL) or
15854 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15855 provided by the library user) must be used. When the standard
15856 debugging functions are used, CRYPTO_dbg_set_options can be used to
15857 request additional information:
15858 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15859 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15861 Also, things like CRYPTO_set_mem_functions will always give the
15862 expected result (the new set of functions is used for allocation
15863 and deallocation) at all times, regardless of platform and compiler
15866 To finish it up, some functions that were never use in any other
15867 way than through macros have a new API and new semantic:
15869 CRYPTO_dbg_malloc()
15870 CRYPTO_dbg_realloc()
15873 All macros of value have retained their old syntax.
15875 *Richard Levitte and Bodo Moeller*
15877 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15878 ordering of SMIMECapabilities wasn't in "strength order" and there
15879 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15884 * Some ASN1 types with illegal zero length encoding (INTEGER,
15885 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15887 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15889 * Merge in my S/MIME library for OpenSSL. This provides a simple
15890 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15891 functionality to handle multipart/signed properly) and a utility
15892 called 'smime' to call all this stuff. This is based on code I
15893 originally wrote for Celo who have kindly allowed it to be
15894 included in OpenSSL.
15898 * Add variants des_set_key_checked and des_set_key_unchecked of
15899 des_set_key (aka des_key_sched). Global variable des_check_key
15900 decides which of these is called by des_set_key; this way
15901 des_check_key behaves as it always did, but applications and
15902 the library itself, which was buggy for des_check_key == 1,
15903 have a cleaner way to pick the version they need.
15907 * New function PKCS12_newpass() which changes the password of a
15912 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15913 dynamic mix. In both cases the ids can be used as an index into the
15914 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15915 functions so they accept a list of the field values and the
15916 application doesn't need to directly manipulate the X509_TRUST
15921 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15926 * Modify the way the V3 extension code looks up extensions. This now
15927 works in a similar way to the object code: we have some "standard"
15928 extensions in a static table which is searched with OBJ_bsearch()
15929 and the application can add dynamic ones if needed. The file
15930 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15931 updated whenever a new extension is added to the core code and kept
15932 in ext_nid order. There is a simple program 'tabtest.c' which checks
15933 this. New extensions are not added too often so this file can readily
15934 be maintained manually.
15936 There are two big advantages in doing things this way. The extensions
15937 can be looked up immediately and no longer need to be "added" using
15938 X509V3_add_standard_extensions(): this function now does nothing.
15939 Side note: I get *lots* of email saying the extension code doesn't
15940 work because people forget to call this function.
15941 Also no dynamic allocation is done unless new extensions are added:
15942 so if we don't add custom extensions there is no need to call
15943 X509V3_EXT_cleanup().
15947 * Modify enc utility's salting as follows: make salting the default. Add a
15948 magic header, so unsalted files fail gracefully instead of just decrypting
15949 to garbage. This is because not salting is a big security hole, so people
15950 should be discouraged from doing it.
15954 * Fixes and enhancements to the 'x509' utility. It allowed a message
15955 digest to be passed on the command line but it only used this
15956 parameter when signing a certificate. Modified so all relevant
15957 operations are affected by the digest parameter including the
15958 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15959 DSA key was used because it didn't fix the digest.
15963 * Initial certificate chain verify code. Currently tests the untrusted
15964 certificates for consistency with the verify purpose (which is set
15965 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15967 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15968 this is because it will reject chains with invalid extensions whereas
15969 every previous version of OpenSSL and SSLeay made no checks at all.
15971 Trust code: checks the root CA for the relevant trust settings. Trust
15972 settings have an initial value consistent with the verify purpose: e.g.
15973 if the verify purpose is for SSL client use it expects the CA to be
15974 trusted for SSL client use. However the default value can be changed to
15975 permit custom trust settings: one example of this would be to only trust
15976 certificates from a specific "secure" set of CAs.
15978 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15979 which should be used for version portability: especially since the
15980 verify structure is likely to change more often now.
15982 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15983 to set them. If not set then assume SSL clients will verify SSL servers
15986 Two new options to the verify program: -untrusted allows a set of
15987 untrusted certificates to be passed in and -purpose which sets the
15988 intended purpose of the certificate. If a purpose is set then the
15989 new chain verify code is used to check extension consistency.
15993 * Support for the authority information access extension.
15997 * Modify RSA and DSA PEM read routines to transparently handle
15998 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15999 public keys in a format compatible with certificate
16000 SubjectPublicKeyInfo structures. Unfortunately there were already
16001 functions called *_PublicKey_* which used various odd formats so
16002 these are retained for compatibility: however the DSA variants were
16003 never in a public release so they have been deleted. Changed dsa/rsa
16004 utilities to handle the new format: note no releases ever handled public
16005 keys so we should be OK.
16007 The primary motivation for this change is to avoid the same fiasco
16008 that dogs private keys: there are several incompatible private key
16009 formats some of which are standard and some OpenSSL specific and
16010 require various evil hacks to allow partial transparent handling and
16011 even then it doesn't work with DER formats. Given the option anything
16012 other than PKCS#8 should be dumped: but the other formats have to
16013 stay in the name of compatibility.
16015 With public keys and the benefit of hindsight one standard format
16016 is used which works with EVP_PKEY, RSA or DSA structures: though
16017 it clearly returns an error if you try to read the wrong kind of key.
16019 Added a -pubkey option to the 'x509' utility to output the public key.
16020 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16021 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16022 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16023 that do the same as the `EVP_PKEY_assign_*()` except they up the
16024 reference count of the added key (they don't "swallow" the
16029 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16030 CRLs would fail if the file contained no certificates or no CRLs:
16031 added a new function to read in both types and return the number
16032 read: this means that if none are read it will be an error. The
16033 DER versions of the certificate and CRL reader would always fail
16034 because it isn't possible to mix certificates and CRLs in DER format
16035 without choking one or the other routine. Changed this to just read
16036 a certificate: this is the best we can do. Also modified the code
16037 in apps/verify.c to take notice of return codes: it was previously
16038 attempting to read in certificates from NULL pointers and ignoring
16039 any errors: this is one reason why the cert and CRL reader seemed
16040 to work. It doesn't check return codes from the default certificate
16041 routines: these may well fail if the certificates aren't installed.
16045 * Code to support otherName option in GeneralName.
16049 * First update to verify code. Change the verify utility
16050 so it warns if it is passed a self signed certificate:
16051 for consistency with the normal behaviour. X509_verify
16052 has been modified to it will now verify a self signed
16053 certificate if *exactly* the same certificate appears
16054 in the store: it was previously impossible to trust a
16055 single self signed certificate. This means that:
16056 openssl verify ss.pem
16057 now gives a warning about a self signed certificate but
16058 openssl verify -CAfile ss.pem ss.pem
16063 * For servers, store verify_result in SSL_SESSION data structure
16064 (and add it to external session representation).
16065 This is needed when client certificate verifications fails,
16066 but an application-provided verification callback (set by
16067 SSL_CTX_set_cert_verify_callback) allows accepting the session
16068 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16069 but returns 1): When the session is reused, we have to set
16070 ssl->verify_result to the appropriate error code to avoid
16073 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16075 * Fix a bug in the new PKCS#7 code: it didn't consider the
16076 case in PKCS7_dataInit() where the signed PKCS7 structure
16077 didn't contain any existing data because it was being created.
16079 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16081 * Add a salt to the key derivation routines in enc.c. This
16082 forms the first 8 bytes of the encrypted file. Also add a
16083 -S option to allow a salt to be input on the command line.
16087 * New function X509_cmp(). Oddly enough there wasn't a function
16088 to compare two certificates. We do this by working out the SHA1
16089 hash and comparing that. X509_cmp() will be needed by the trust
16094 * SSL_get1_session() is like SSL_get_session(), but increments
16095 the reference count in the SSL_SESSION returned.
16097 *Geoff Thorpe <geoff@eu.c2.net>*
16099 * Fix for 'req': it was adding a null to request attributes.
16100 Also change the X509_LOOKUP and X509_INFO code to handle
16101 certificate auxiliary information.
16105 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16110 * Add the possibility to add extra information to the memory leak
16111 detecting output, to form tracebacks, showing from where each
16112 allocation was originated: CRYPTO_push_info("constant string") adds
16113 the string plus current file name and line number to a per-thread
16114 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16115 is like calling CYRPTO_pop_info() until the stack is empty.
16116 Also updated memory leak detection code to be multi-thread-safe.
16120 * Add options -text and -noout to pkcs7 utility and delete the
16121 encryption options which never did anything. Update docs.
16125 * Add options to some of the utilities to allow the pass phrase
16126 to be included on either the command line (not recommended on
16127 OSes like Unix) or read from the environment. Update the
16128 manpages and fix a few bugs.
16132 * Add a few manpages for some of the openssl commands.
16136 * Fix the -revoke option in ca. It was freeing up memory twice,
16137 leaking and not finding already revoked certificates.
16141 * Extensive changes to support certificate auxiliary information.
16142 This involves the use of X509_CERT_AUX structure and X509_AUX
16143 functions. An X509_AUX function such as PEM_read_X509_AUX()
16144 can still read in a certificate file in the usual way but it
16145 will also read in any additional "auxiliary information". By
16146 doing things this way a fair degree of compatibility can be
16147 retained: existing certificates can have this information added
16148 using the new 'x509' options.
16150 Current auxiliary information includes an "alias" and some trust
16151 settings. The trust settings will ultimately be used in enhanced
16152 certificate chain verification routines: currently a certificate
16153 can only be trusted if it is self signed and then it is trusted
16158 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16159 The problem was that one of the replacement routines had not been working
16160 since SSLeay releases. For now the offending routine has been replaced
16161 with non-optimised assembler. Even so, this now gives around 95%
16162 performance improvement for 1024 bit RSA signs.
16166 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16167 handling. Most clients have the effective key size in bits equal to
16168 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16169 A few however don't do this and instead use the size of the decrypted key
16170 to determine the RC2 key length and the AlgorithmIdentifier to determine
16171 the effective key length. In this case the effective key length can still
16172 be 40 bits but the key length can be 168 bits for example. This is fixed
16173 by manually forcing an RC2 key into the EVP_PKEY structure because the
16174 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16175 the key length and effective key length are equal.
16179 * Add a bunch of functions that should simplify the creation of
16180 X509_NAME structures. Now you should be able to do:
16181 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16182 and have it automatically work out the correct field type and fill in
16183 the structures. The more adventurous can try:
16184 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16185 and it will (hopefully) work out the correct multibyte encoding.
16189 * Change the 'req' utility to use the new field handling and multibyte
16190 copy routines. Before the DN field creation was handled in an ad hoc
16191 way in req, ca, and x509 which was rather broken and didn't support
16192 BMPStrings or UTF8Strings. Since some software doesn't implement
16193 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16194 using the dirstring_type option. See the new comment in the default
16195 openssl.cnf for more info.
16199 * Make crypto/rand/md_rand.c more robust:
16200 - Assure unique random numbers after fork().
16201 - Make sure that concurrent threads access the global counter and
16202 md serializably so that we never lose entropy in them
16203 or use exactly the same state in multiple threads.
16204 Access to the large state is not always serializable because
16205 the additional locking could be a performance killer, and
16206 md should be large enough anyway.
16210 * New file apps/app_rand.c with commonly needed functionality
16211 for handling the random seed file.
16213 Use the random seed file in some applications that previously did not:
16215 dsaparam -genkey (which also ignored its '-rand' option),
16218 x509 (when signing).
16219 Except on systems with /dev/urandom, it is crucial to have a random
16220 seed file at least for key creation, DSA signing, and for DH exchanges;
16221 for RSA signatures we could do without one.
16223 gendh and gendsa (unlike genrsa) used to read only the first byte
16224 of each file listed in the '-rand' option. The function as previously
16225 found in genrsa is now in app_rand.c and is used by all programs
16226 that support '-rand'.
16230 * In RAND_write_file, use mode 0600 for creating files;
16231 don't just chmod when it may be too late.
16235 * Report an error from X509_STORE_load_locations
16236 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16240 * New function ASN1_mbstring_copy() this copies a string in either
16241 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16242 into an ASN1_STRING type. A mask of permissible types is passed
16243 and it chooses the "minimal" type to use or an error if not type
16248 * Add function equivalents to the various macros in asn1.h. The old
16249 macros are retained with an `M_` prefix. Code inside the library can
16250 use the `M_` macros. External code (including the openssl utility)
16251 should *NOT* in order to be "shared library friendly".
16255 * Add various functions that can check a certificate's extensions
16256 to see if it usable for various purposes such as SSL client,
16257 server or S/MIME and CAs of these types. This is currently
16258 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16259 verification. Also added a -purpose flag to x509 utility to
16260 print out all the purposes.
16264 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16269 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16270 for, obtain and decode and extension and obtain its critical flag.
16271 This allows all the necessary extension code to be handled in a
16272 single function call.
16276 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16277 platforms. See crypto/rc4/rc4_enc.c for further details.
16281 * New -noout option to asn1parse. This causes no output to be produced
16282 its main use is when combined with -strparse and -out to extract data
16283 from a file (which may not be in ASN.1 format).
16287 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16288 when producing the local key id.
16290 *Richard Levitte <levitte@stacken.kth.se>*
16292 * New option -dhparam in s_server. This allows a DH parameter file to be
16293 stated explicitly. If it is not stated then it tries the first server
16294 certificate file. The previous behaviour hard coded the filename
16299 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16300 a public key to be input or output. For example:
16301 openssl rsa -in key.pem -pubout -out pubkey.pem
16302 Also added necessary DSA public key functions to handle this.
16306 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16307 in the message. This was handled by allowing
16308 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16310 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16312 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16313 to the end of the strings whereas this didn't. This would cause problems
16314 if strings read with d2i_ASN1_bytes() were later modified.
16316 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16318 * Fix for base64 decode bug. When a base64 bio reads only one line of
16319 data and it contains EOF it will end up returning an error. This is
16320 caused by input 46 bytes long. The cause is due to the way base64
16321 BIOs find the start of base64 encoded data. They do this by trying a
16322 trial decode on each line until they find one that works. When they
16323 do a flag is set and it starts again knowing it can pass all the
16324 data directly through the decoder. Unfortunately it doesn't reset
16325 the context it uses. This means that if EOF is reached an attempt
16326 is made to pass two EOFs through the context and this causes the
16327 resulting error. This can also cause other problems as well. As is
16328 usual with these problems it takes *ages* to find and the fix is
16329 trivial: move one line.
16331 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16333 * Ugly workaround to get s_client and s_server working under Windows. The
16334 old code wouldn't work because it needed to select() on sockets and the
16335 tty (for keypresses and to see if data could be written). Win32 only
16336 supports select() on sockets so we select() with a 1s timeout on the
16337 sockets and then see if any characters are waiting to be read, if none
16338 are present then we retry, we also assume we can always write data to
16339 the tty. This isn't nice because the code then blocks until we've
16340 received a complete line of data and it is effectively polling the
16341 keyboard at 1s intervals: however it's quite a bit better than not
16342 working at all :-) A dedicated Windows application might handle this
16343 with an event loop for example.
16347 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16348 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16349 will be called when RSA_sign() and RSA_verify() are used. This is useful
16350 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16351 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16352 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16353 This necessitated the support of an extra signature type NID_md5_sha1
16354 for SSL signatures and modifications to the SSL library to use it instead
16355 of calling RSA_public_decrypt() and RSA_private_encrypt().
16359 * Add new -verify -CAfile and -CApath options to the crl program, these
16360 will lookup a CRL issuers certificate and verify the signature in a
16361 similar way to the verify program. Tidy up the crl program so it
16362 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16363 less strict. It will now permit CRL extensions even if it is not
16364 a V2 CRL: this will allow it to tolerate some broken CRLs.
16368 * Initialize all non-automatic variables each time one of the openssl
16369 sub-programs is started (this is necessary as they may be started
16370 multiple times from the "OpenSSL>" prompt).
16372 *Lennart Bang, Bodo Moeller*
16374 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16375 removing all other RSA functionality (this is what NO_RSA does). This
16376 is so (for example) those in the US can disable those operations covered
16377 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16382 * Non-copying interface to BIO pairs.
16383 (still largely untested)
16387 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16388 ASCII string. This was handled independently in various places before.
16392 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16393 UTF8 strings a character at a time.
16397 * Use client_version from client hello to select the protocol
16398 (s23_srvr.c) and for RSA client key exchange verification
16399 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16403 * Add various utility functions to handle SPKACs, these were previously
16404 handled by poking round in the structure internals. Added new function
16405 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16406 print, verify and generate SPKACs. Based on an original idea from
16407 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16411 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16415 * Allow the config file extension section to be overwritten on the
16416 command line. Based on an original idea from Massimiliano Pala
16417 <madwolf@comune.modena.it>. The new option is called -extensions
16418 and can be applied to ca, req and x509. Also -reqexts to override
16419 the request extensions in req and -crlexts to override the crl extensions
16424 * Add new feature to the SPKAC handling in ca. Now you can include
16425 the same field multiple times by preceding it by "XXXX." for example:
16428 this is the same syntax as used in the req config file.
16432 * Allow certificate extensions to be added to certificate requests. These
16433 are specified in a 'req_extensions' option of the req section of the
16434 config file. They can be printed out with the -text option to req but
16435 are otherwise ignored at present.
16439 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16440 data read consists of only the final block it would not decrypted because
16441 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16442 A misplaced 'break' also meant the decrypted final block might not be
16443 copied until the next read.
16447 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16448 a few extra parameters to the DH structure: these will be useful if
16449 for example we want the value of 'q' or implement X9.42 DH.
16453 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16454 provides hooks that allow the default DSA functions or functions on a
16455 "per key" basis to be replaced. This allows hardware acceleration and
16456 hardware key storage to be handled without major modification to the
16457 library. Also added low level modexp hooks and CRYPTO_EX structure and
16458 associated functions.
16462 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16463 as "read only": it can't be written to and the buffer it points to will
16464 not be freed. Reading from a read only BIO is much more efficient than
16465 a normal memory BIO. This was added because there are several times when
16466 an area of memory needs to be read from a BIO. The previous method was
16467 to create a memory BIO and write the data to it, this results in two
16468 copies of the data and an O(n^2) reading algorithm. There is a new
16469 function BIO_new_mem_buf() which creates a read only memory BIO from
16470 an area of memory. Also modified the PKCS#7 routines to use read only
16475 * Bugfix: ssl23_get_client_hello did not work properly when called in
16476 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16477 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16478 but a retry condition occurred while trying to read the rest.
16482 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16483 NID_pkcs7_encrypted by default: this was wrong since this should almost
16484 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16485 the encrypted data type: this is a more sensible place to put it and it
16486 allows the PKCS#12 code to be tidied up that duplicated this
16491 * Changed obj_dat.pl script so it takes its input and output files on
16492 the command line. This should avoid shell escape redirection problems
16497 * Initial support for certificate extension requests, these are included
16498 in things like Xenroll certificate requests. Included functions to allow
16499 extensions to be obtained and added.
16503 * -crlf option to s_client and s_server for sending newlines as
16504 CRLF (as required by many protocols).
16508 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16510 * Install libRSAglue.a when OpenSSL is built with RSAref.
16512 *Ralf S. Engelschall*
16514 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16516 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16518 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16523 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16524 DH parameters/keys (q is lost during that conversion, but the resulting
16525 DH parameters contain its length).
16527 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16528 much faster than DH_generate_parameters (which creates parameters
16529 where `p = 2*q + 1`), and also the smaller q makes DH computations
16530 much more efficient (160-bit exponentiation instead of 1024-bit
16531 exponentiation); so this provides a convenient way to support DHE
16532 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16533 utter importance to use
16534 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16536 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16537 when such DH parameters are used, because otherwise small subgroup
16538 attacks may become possible!
16542 * Avoid memory leak in i2d_DHparams.
16546 * Allow the -k option to be used more than once in the enc program:
16547 this allows the same encrypted message to be read by multiple recipients.
16551 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16552 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16553 it will always use the numerical form of the OID, even if it has a short
16558 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16559 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16560 otherwise bn_mod_exp was called. In the case of hardware keys for example
16561 no private key components need be present and it might store extra data
16562 in the RSA structure, which cannot be accessed from bn_mod_exp.
16563 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16564 private key operations.
16568 * Added support for SPARC Linux.
16572 * pem_password_cb function type incompatibly changed from
16573 typedef int pem_password_cb(char *buf, int size, int rwflag);
16575 ....(char *buf, int size, int rwflag, void *userdata);
16576 so that applications can pass data to their callbacks:
16577 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16578 additional void * argument, which is just handed through whenever
16579 the password callback is called.
16581 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16583 New function SSL_CTX_set_default_passwd_cb_userdata.
16585 Compatibility note: As many C implementations push function arguments
16586 onto the stack in reverse order, the new library version is likely to
16587 interoperate with programs that have been compiled with the old
16588 pem_password_cb definition (PEM_whatever takes some data that
16589 happens to be on the stack as its last argument, and the callback
16590 just ignores this garbage); but there is no guarantee whatsoever that
16593 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16594 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16595 problems not only on Windows, but also on some Unix platforms.
16596 To avoid problematic command lines, these definitions are now in an
16597 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16598 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16602 * MIPS III/IV assembler module is reimplemented.
16606 * More DES library cleanups: remove references to srand/rand and
16607 delete an unused file.
16611 * Add support for the free Netwide assembler (NASM) under Win32,
16612 since not many people have MASM (ml) and it can be hard to obtain.
16613 This is currently experimental but it seems to work OK and pass all
16614 the tests. Check out INSTALL.W32 for info.
16618 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16619 without temporary keys kept an extra copy of the server key,
16620 and connections with temporary keys did not free everything in case
16625 * New function RSA_check_key and new openssl rsa option -check
16626 for verifying the consistency of RSA keys.
16628 *Ulf Moeller, Bodo Moeller*
16630 * Various changes to make Win32 compile work:
16631 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16632 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16633 comparison" warnings.
16634 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16638 * Add a debugging option to PKCS#5 v2 key generation function: when
16639 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16640 derived keys are printed to stderr.
16644 * Copy the flags in ASN1_STRING_dup().
16646 *Roman E. Pavlov <pre@mo.msk.ru>*
16648 * The x509 application mishandled signing requests containing DSA
16649 keys when the signing key was also DSA and the parameters didn't match.
16651 It was supposed to omit the parameters when they matched the signing key:
16652 the verifying software was then supposed to automatically use the CA's
16653 parameters if they were absent from the end user certificate.
16655 Omitting parameters is no longer recommended. The test was also
16656 the wrong way round! This was probably due to unusual behaviour in
16657 EVP_cmp_parameters() which returns 1 if the parameters match.
16658 This meant that parameters were omitted when they *didn't* match and
16659 the certificate was useless. Certificates signed with 'ca' didn't have
16662 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16664 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16665 The interface is as follows:
16666 Applications can use
16667 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16668 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16669 "off" is now the default.
16670 The library internally uses
16671 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16672 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16673 to disable memory-checking temporarily.
16675 Some inconsistent states that previously were possible (and were
16676 even the default) are now avoided.
16678 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16679 with each memory chunk allocated; this is occasionally more helpful
16680 than just having a counter.
16682 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16684 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16689 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16690 which largely parallels "options", but is for changing API behaviour,
16691 whereas "options" are about protocol behaviour.
16692 Initial "mode" flags are:
16694 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16695 a single record has been written.
16696 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16697 retries use the same buffer location.
16698 (But all of the contents must be
16703 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16706 * Fix problems with no-hmac etc.
16708 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16710 * New functions RSA_get_default_method(), RSA_set_method() and
16711 RSA_get_method(). These allows replacement of RSA_METHODs without having
16712 to mess around with the internals of an RSA structure.
16716 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16717 Also really enable memory leak checks in openssl.c and in some
16720 *Chad C. Mulligan, Bodo Moeller*
16722 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16723 up the length of negative integers. This has now been simplified to just
16724 store the length when it is first determined and use it later, rather
16725 than trying to keep track of where data is copied and updating it to
16727 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16729 * Add a new function PKCS7_signatureVerify. This allows the verification
16730 of a PKCS#7 signature but with the signing certificate passed to the
16731 function itself. This contrasts with PKCS7_dataVerify which assumes the
16732 certificate is present in the PKCS#7 structure. This isn't always the
16733 case: certificates can be omitted from a PKCS#7 structure and be
16734 distributed by "out of band" means (such as a certificate database).
16738 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16739 function prototypes in pem.h, also change util/mkdef.pl to add the
16740 necessary function names.
16744 * mk1mf.pl (used by Windows builds) did not properly read the
16745 options set by Configure in the top level Makefile, and Configure
16746 was not even able to write more than one option correctly.
16747 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16751 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16752 file to be loaded from a BIO or FILE pointer. The BIO version will
16753 for example allow memory BIOs to contain config info.
16757 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16758 Whoever hopes to achieve shared-library compatibility across versions
16759 must use this, not the compile-time macro.
16760 (Exercise 0.9.4: Which is the minimum library version required by
16762 Note: All this applies only to multi-threaded programs, others don't
16767 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16768 through a BIO pair triggered the default case, i.e.
16769 SSLerr(...,SSL_R_UNKNOWN_STATE).
16773 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16774 can use the SSL library even if none of the specific BIOs is
16779 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16780 for the encoded length.
16782 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16784 * Add initial documentation of the X509V3 functions.
16788 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16789 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16790 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16791 secure PKCS#8 private key format with a high iteration count.
16795 * Fix determination of Perl interpreter: A perl or perl5
16796 *directory* in $PATH was also accepted as the interpreter.
16798 *Ralf S. Engelschall*
16800 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16801 wrong with it but it was very old and did things like calling
16802 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16803 unusual formatting.
16807 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16808 to use the new extension code.
16812 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16813 with macros. This should make it easier to change their form, add extra
16814 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16819 * Add to configuration table a new entry that can specify an alternative
16820 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16821 according to Mark Crispin <MRC@Panda.COM>.
16825 * DES CBC did not update the IV. Weird.
16829 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16830 Changing the behaviour of the former might break existing programs --
16831 where IV updating is needed, des_ncbc_encrypt can be used.
16834 * When bntest is run from "make test" it drives bc to check its
16835 calculations, as well as internally checking them. If an internal check
16836 fails, it needs to cause bc to give a non-zero result or make test carries
16837 on without noticing the failure. Fixed.
16841 * DES library cleanups.
16845 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16846 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16847 ciphers. NOTE: although the key derivation function has been verified
16848 against some published test vectors it has not been extensively tested
16849 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16854 * Instead of "mkdir -p", which is not fully portable, use new
16855 Perl script "util/mkdir-p.pl".
16859 * Rewrite the way password based encryption (PBE) is handled. It used to
16860 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16861 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16862 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16863 the 'parameter' field of the AlgorithmIdentifier is passed to the
16864 underlying key generation function so it must do its own ASN1 parsing.
16865 This has also changed the EVP_PBE_CipherInit() function which now has a
16866 'parameter' argument instead of literal salt and iteration count values
16867 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16871 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16872 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16873 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16874 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16875 value was just used as a "magic string" and not used directly its
16876 value doesn't matter.
16880 * Introduce some semblance of const correctness to BN. Shame C doesn't
16885 * "linux-sparc64" configuration (ultrapenguin).
16887 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16888 "linux-sparc" configuration.
16890 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16892 * config now generates no-xxx options for missing ciphers.
16896 * Support the EBCDIC character set (work in progress).
16897 File ebcdic.c not yet included because it has a different license.
16899 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16901 * Support BS2000/OSD-POSIX.
16903 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16905 * Make callbacks for key generation use `void *` instead of `char *`.
16909 * Make S/MIME samples compile (not yet tested).
16913 * Additional typesafe stacks.
16917 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16921 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16923 * New configuration variant "sco5-gcc".
16925 * Updated some demos.
16927 *Sean O Riordain, Wade Scholine*
16929 * Add missing BIO_free at exit of pkcs12 application.
16933 * Fix memory leak in conf.c.
16937 * Updates for Win32 to assembler version of MD5.
16941 * Set #! path to perl in apps/der_chop to where we found it
16942 instead of using a fixed path.
16946 * SHA library changes for irix64-mips4-cc.
16950 * Improvements for VMS support.
16954 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16956 * Bignum library bug fix. IRIX 6 passes "make test" now!
16957 This also avoids the problems with SC4.2 and unpatched SC5.
16959 *Andy Polyakov <appro@fy.chalmers.se>*
16961 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16962 These are required because of the typesafe stack would otherwise break
16963 existing code. If old code used a structure member which used to be STACK
16964 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16965 sk_num or sk_value it would produce an error because the num, data members
16966 are not present in STACK_OF. Now it just produces a warning. sk_set
16967 replaces the old method of assigning a value to sk_value
16968 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16969 that does this will no longer work (and should use sk_set instead) but
16970 this could be regarded as a "questionable" behaviour anyway.
16974 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16975 correctly handle encrypted S/MIME data.
16979 * Change type of various DES function arguments from des_cblock
16980 (which means, in function argument declarations, pointer to char)
16981 to des_cblock * (meaning pointer to array with 8 char elements),
16982 which allows the compiler to do more typechecking; it was like
16983 that back in SSLeay, but with lots of ugly casts.
16985 Introduce new type const_des_cblock.
16989 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16990 problems: find RecipientInfo structure that matches recipient certificate
16991 and initialise the ASN1 structures properly based on passed cipher.
16995 * Belatedly make the BN tests actually check the results.
16999 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17000 to and from BNs: it was completely broken. New compilation option
17001 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17002 key elements as negative integers.
17006 * Reorganize and speed up MD5.
17008 *Andy Polyakov <appro@fy.chalmers.se>*
17012 *Richard Levitte <richard@levitte.org>*
17014 * New option -out to asn1parse to allow the parsed structure to be
17015 output to a file. This is most useful when combined with the -strparse
17016 option to examine the output of things like OCTET STRINGS.
17020 * Make SSL library a little more fool-proof by not requiring any longer
17021 that `SSL_set_{accept,connect}_state` be called before
17022 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17023 in many applications because usually everything *appeared* to work as
17024 intended anyway -- now it really works as intended).
17028 * Move openssl.cnf out of lib/.
17032 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17033 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17034 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17036 *Ralf S. Engelschall*
17038 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17039 handle PKCS#7 enveloped data properly.
17041 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17043 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17044 copying pointers. The cert_st handling is changed by this in
17045 various ways (and thus what used to be known as ctx->default_cert
17046 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17047 any longer when s->cert does not give us what we need).
17048 ssl_cert_instantiate becomes obsolete by this change.
17049 As soon as we've got the new code right (possibly it already is?),
17050 we have solved a couple of bugs of the earlier code where s->cert
17051 was used as if it could not have been shared with other SSL structures.
17053 Note that using the SSL API in certain dirty ways now will result
17054 in different behaviour than observed with earlier library versions:
17055 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17056 does not influence s as it used to.
17058 In order to clean up things more thoroughly, inside SSL_SESSION
17059 we don't use CERT any longer, but a new structure SESS_CERT
17060 that holds per-session data (if available); currently, this is
17061 the peer's certificate chain and, for clients, the server's certificate
17062 and temporary key. CERT holds only those values that can have
17063 meaningful defaults in an SSL_CTX.
17067 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17068 from the internal representation. Various PKCS#7 fixes: remove some
17069 evil casts and set the enc_dig_alg field properly based on the signing
17074 * Allow PKCS#12 password to be set from the command line or the
17075 environment. Let 'ca' get its config file name from the environment
17076 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17081 * Allow certificate policies extension to use an IA5STRING for the
17082 organization field. This is contrary to the PKIX definition but
17083 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17088 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17089 without disallowing inline assembler and the like for non-pedantic builds.
17093 * Support Borland C++ builder.
17095 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17101 * SHA-1 cleanups and performance enhancements.
17103 *Andy Polyakov <appro@fy.chalmers.se>*
17105 * Sparc v8plus assembler for the bignum library.
17107 *Andy Polyakov <appro@fy.chalmers.se>*
17109 * Accept any -xxx and +xxx compiler options in Configure.
17113 * Update HPUX configuration.
17117 * Add missing `sk_<type>_unshift()` function to safestack.h
17119 *Ralf S. Engelschall*
17121 * New function SSL_CTX_use_certificate_chain_file that sets the
17122 "extra_cert"s in addition to the certificate. (This makes sense
17123 only for "PEM" format files, as chains as a whole are not
17128 * Support verify_depth from the SSL API.
17129 x509_vfy.c had what can be considered an off-by-one-error:
17130 Its depth (which was not part of the external interface)
17131 was actually counting the number of certificates in a chain;
17132 now it really counts the depth.
17136 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17137 instead of X509err, which often resulted in confusing error
17138 messages since the error codes are not globally unique
17139 (e.g. an alleged error in ssl3_accept when a certificate
17140 didn't match the private key).
17142 * New function SSL_CTX_set_session_id_context that allows to set a default
17143 value (so that you don't need SSL_set_session_id_context for each
17144 connection using the SSL_CTX).
17148 * OAEP decoding bug fix.
17152 * Support INSTALL_PREFIX for package builders, as proposed by
17157 * New Configure options "threads" and "no-threads". For systems
17158 where the proper compiler options are known (currently Solaris
17159 and Linux), "threads" is the default.
17163 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17167 * Install various scripts to $(OPENSSLDIR)/misc, not to
17168 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17169 such as /usr/local/bin.
17173 * "make linux-shared" to build shared libraries.
17175 *Niels Poppe <niels@netbox.org>*
17177 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17181 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17182 extension adding in x509 utility.
17186 * Remove NOPROTO sections and error code comments.
17190 * Partial rewrite of the DEF file generator to now parse the ANSI
17195 * New Configure options --prefix=DIR and --openssldir=DIR.
17199 * Complete rewrite of the error code script(s). It is all now handled
17200 by one script at the top level which handles error code gathering,
17201 header rewriting and C source file generation. It should be much better
17202 than the old method: it now uses a modified version of Ulf's parser to
17203 read the ANSI prototypes in all header files (thus the old K&R definitions
17204 aren't needed for error creation any more) and do a better job of
17205 translating function codes into names. The old 'ASN1 error code embedded
17206 in a comment' is no longer necessary and it doesn't use .err files which
17207 have now been deleted. Also the error code call doesn't have to appear all
17208 on one line (which resulted in some large lines...).
17212 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17216 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17217 0 (which usually indicates a closed connection), but continue reading.
17221 * Fix some race conditions.
17225 * Add support for CRL distribution points extension. Add Certificate
17226 Policies and CRL distribution points documentation.
17230 * Move the autogenerated header file parts to crypto/opensslconf.h.
17234 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17235 8 of keying material. Merlin has also confirmed interop with this fix
17236 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17238 *Merlin Hughes <merlin@baltimore.ie>*
17240 * Fix lots of warnings.
17242 *Richard Levitte <levitte@stacken.kth.se>*
17244 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17245 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17247 *Richard Levitte <levitte@stacken.kth.se>*
17249 * Fix problems with sizeof(long) == 8.
17251 *Andy Polyakov <appro@fy.chalmers.se>*
17253 * Change functions to ANSI C.
17257 * Fix typos in error codes.
17259 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17261 * Remove defunct assembler files from Configure.
17265 * SPARC v8 assembler BIGNUM implementation.
17267 *Andy Polyakov <appro@fy.chalmers.se>*
17269 * Support for Certificate Policies extension: both print and set.
17270 Various additions to support the r2i method this uses.
17274 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17275 return a const string when you are expecting an allocated buffer.
17279 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17280 types DirectoryString and DisplayText.
17284 * Add code to allow r2i extensions to access the configuration database,
17285 add an LHASH database driver and add several ctx helper functions.
17289 * Fix an evil bug in bn_expand2() which caused various BN functions to
17290 fail when they extended the size of a BIGNUM.
17294 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17295 support typesafe stack.
17299 * Fix typo in SSL_[gs]et_options().
17301 *Nils Frostberg <nils@medcom.se>*
17303 * Delete various functions and files that belonged to the (now obsolete)
17304 old X509V3 handling code.
17308 * New Configure option "rsaref".
17312 * Don't auto-generate pem.h.
17316 * Introduce type-safe ASN.1 SETs.
17320 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17322 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17324 * Introduce type-safe STACKs. This will almost certainly break lots of code
17325 that links with OpenSSL (well at least cause lots of warnings), but fear
17326 not: the conversion is trivial, and it eliminates loads of evil casts. A
17327 few STACKed things have been converted already. Feel free to convert more.
17328 In the fullness of time, I'll do away with the STACK type altogether.
17332 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17333 specified in `<certfile>` by updating the entry in the index.txt file.
17334 This way one no longer has to edit the index.txt file manually for
17335 revoking a certificate. The -revoke option does the gory details now.
17337 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17339 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17340 `-text` option at all and this way the `-noout -text` combination was
17341 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17343 *Ralf S. Engelschall*
17345 * Make sure a corresponding plain text error message exists for the
17346 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17347 verify callback function determined that a certificate was revoked.
17349 *Ralf S. Engelschall*
17351 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17352 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17353 all available ciphers including rc5, which was forgotten until now.
17354 In order to let the testing shell script know which algorithms
17355 are available, a new (up to now undocumented) command
17356 `openssl list-cipher-commands` is used.
17360 * Bugfix: s_client occasionally would sleep in select() when
17361 it should have checked SSL_pending() first.
17365 * New functions DSA_do_sign and DSA_do_verify to provide access to
17366 the raw DSA values prior to ASN.1 encoding.
17370 * Tweaks to Configure
17372 *Niels Poppe <niels@netbox.org>*
17374 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17379 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17383 * New config option to avoid instructions that are illegal on the 80386.
17384 The default code is faster, but requires at least a 486.
17388 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17389 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17390 same as SSL2_VERSION anyway.
17394 * New "-showcerts" option for s_client.
17398 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17399 application. Various cleanups and fixes.
17403 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17404 modify error routines to work internally. Add error codes and PBE init
17405 to library startup routines.
17409 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17410 packing functions to asn1 and evp. Changed function names and error
17411 codes along the way.
17415 * PKCS12 integration: and so it begins... First of several patches to
17416 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17417 objects to objects.h
17421 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17422 and display support for Thawte strong extranet extension.
17426 * Add LinuxPPC support.
17428 *Jeff Dubrule <igor@pobox.org>*
17430 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17431 bn_div_words in alpha.s.
17433 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17435 * Make sure the RSA OAEP test is skipped under -DRSAref because
17436 OAEP isn't supported when OpenSSL is built with RSAref.
17438 *Ulf Moeller <ulf@fitug.de>*
17440 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17441 so they no longer are missing under -DNOPROTO.
17443 *Soren S. Jorvang <soren@t.dk>*
17445 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17447 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17448 doesn't work when the session is reused. Coming soon!
17452 * Fix a security hole, that allows sessions to be reused in the wrong
17453 context thus bypassing client cert protection! All software that uses
17454 client certs and session caches in multiple contexts NEEDS PATCHING to
17455 allow session reuse! A fuller solution is in the works.
17457 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17459 * Some more source tree cleanups (removed obsolete files
17460 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17461 permission on "config" script to be executable) and a fix for the INSTALL
17464 *Ulf Moeller <ulf@fitug.de>*
17466 * Remove some legacy and erroneous uses of malloc, free instead of
17469 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17471 * Make rsa_oaep_test return non-zero on error.
17473 *Ulf Moeller <ulf@fitug.de>*
17475 * Add support for native Solaris shared libraries. Configure
17476 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17477 if someone would make that last step automatic.
17479 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17481 * ctx_size was not built with the right compiler during "make links". Fixed.
17485 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17486 except NULL ciphers". This means the default cipher list will no longer
17487 enable NULL ciphers. They need to be specifically enabled e.g. with
17488 the string "DEFAULT:eNULL".
17492 * Fix to RSA private encryption routines: if p < q then it would
17493 occasionally produce an invalid result. This will only happen with
17494 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17498 * Be less restrictive and allow also `perl util/perlpath.pl
17499 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17500 because this way one can also use an interpreter named `perl5` (which is
17501 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17502 installed as `perl`).
17504 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17506 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17508 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17510 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17511 advapi32.lib to Win32 build and change the pem test comparison
17512 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17513 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17514 and crypto/des/ede_cbcm_enc.c.
17518 * DES quad checksum was broken on big-endian architectures. Fixed.
17522 * Comment out two functions in bio.h that aren't implemented. Fix up the
17523 Win32 test batch file so it (might) work again. The Win32 test batch file
17524 is horrible: I feel ill....
17528 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17529 in e_os.h. Audit of header files to check ANSI and non ANSI
17530 sections: 10 functions were absent from non ANSI section and not exported
17531 from Windows DLLs. Fixed up libeay.num for new functions.
17535 * Make `openssl version` output lines consistent.
17537 *Ralf S. Engelschall*
17539 * Fix Win32 symbol export lists for BIO functions: Added
17540 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17541 to ms/libeay{16,32}.def.
17543 *Ralf S. Engelschall*
17545 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17546 fine under Unix and passes some trivial tests I've now added. But the
17547 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17548 added to make sure no one expects that this stuff really works in the
17549 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17550 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17553 *Ralf S. Engelschall*
17555 * Fix the generation of two part addresses in perl.
17557 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17559 * Add config entry for Linux on MIPS.
17561 *John Tobey <jtobey@channel1.com>*
17563 * Make links whenever Configure is run, unless we are on Windoze.
17567 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17568 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17573 * Add a useful kludge to allow package maintainers to specify compiler and
17574 other platforms details on the command line without having to patch the
17575 Configure script every time: One now can use
17576 `perl Configure <id>:<details>`,
17577 i.e. platform ids are allowed to have details appended
17578 to them (separated by colons). This is treated as there would be a static
17579 pre-configured entry in Configure's %table under key `<id>` with value
17580 `<details>` and `perl Configure <id>` is called. So, when you want to
17581 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17582 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17583 now, which overrides the FreeBSD-elf entry on-the-fly.
17585 *Ralf S. Engelschall*
17587 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17591 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17592 on the `perl Configure ...` command line. This way one can compile
17593 OpenSSL libraries with Position Independent Code (PIC) which is needed
17594 for linking it into DSOs.
17596 *Ralf S. Engelschall*
17598 * Remarkably, export ciphers were totally broken and no-one had noticed!
17603 * Cleaned up the LICENSE document: The official contact for any license
17604 questions now is the OpenSSL core team under openssl-core@openssl.org.
17605 And add a paragraph about the dual-license situation to make sure people
17606 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17607 to the OpenSSL toolkit.
17609 *Ralf S. Engelschall*
17611 * General source tree makefile cleanups: Made `making xxx in yyy...`
17612 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17613 Additionally cleaned up the `make links` target: Remove unnecessary
17614 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17615 to speed processing and no longer clutter the display with confusing
17616 stuff. Instead only the actually done links are displayed.
17618 *Ralf S. Engelschall*
17620 * Permit null encryption ciphersuites, used for authentication only. It used
17621 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17622 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17627 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17628 signed attributes when verifying signatures (this would break them),
17629 the detached data encoding was wrong and public keys obtained using
17630 X509_get_pubkey() weren't freed.
17634 * Add text documentation for the BUFFER functions. Also added a work around
17635 to a Win95 console bug. This was triggered by the password read stuff: the
17636 last character typed gets carried over to the next fread(). If you were
17637 generating a new cert request using 'req' for example then the last
17638 character of the passphrase would be CR which would then enter the first
17643 * Added the new 'Includes OpenSSL Cryptography Software' button as
17644 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17645 button and can be used by applications based on OpenSSL to show the
17646 relationship to the OpenSSL project.
17648 *Ralf S. Engelschall*
17650 * Remove confusing variables in function signatures in files
17651 ssl/ssl_lib.c and ssl/ssl.h.
17653 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17655 * Don't install bss_file.c under PREFIX/include/
17657 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17659 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17660 functions that return function pointers and has support for NT specific
17661 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17662 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17663 unsigned to signed types: this was killing the Win32 compile.
17667 * Add new certificate file to stack functions,
17668 SSL_add_dir_cert_subjects_to_stack() and
17669 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17670 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17671 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17672 This means that Apache-SSL and similar packages don't have to mess around
17673 to add as many CAs as they want to the preferred list.
17677 * Experiment with doxygen documentation. Currently only partially applied to
17679 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17680 openssl.doxy as the configuration file.
17684 * Get rid of remaining C++-style comments which strict C compilers hate.
17686 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17688 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17689 compiled in by default: it has problems with large keys.
17693 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17694 DH private keys and/or callback functions which directly correspond to
17695 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17696 is needed for applications which have to configure certificates on a
17697 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17699 For the RSA certificate situation is makes no difference, but
17700 for the DSA certificate situation this fixes the "no shared cipher"
17701 problem where the OpenSSL cipher selection procedure failed because the
17702 temporary keys were not overtaken from the context and the API provided
17703 no way to reconfigure them.
17704 The new functions now let applications reconfigure the stuff and they
17705 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17706 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17707 non-public-API function ssl_cert_instantiate() is used as a helper
17708 function and also to reduce code redundancy inside ssl_rsa.c.
17710 *Ralf S. Engelschall*
17712 * Move s_server -dcert and -dkey options out of the undocumented feature
17713 area because they are useful for the DSA situation and should be
17714 recognized by the users.
17716 *Ralf S. Engelschall*
17718 * Fix the cipher decision scheme for export ciphers: the export bits are
17719 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17720 SSL_EXP_MASK. So, the original variable has to be used instead of the
17721 already masked variable.
17723 *Richard Levitte <levitte@stacken.kth.se>*
17725 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17727 *Richard Levitte <levitte@stacken.kth.se>*
17729 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17730 from `int` to `unsigned int` because it is a length and initialized by
17731 EVP_DigestFinal() which expects an `unsigned int *`.
17733 *Richard Levitte <levitte@stacken.kth.se>*
17735 * Don't hard-code path to Perl interpreter on shebang line of Configure
17736 script. Instead use the usual Shell->Perl transition trick.
17738 *Ralf S. Engelschall*
17740 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17741 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17742 -noout -modulus` as it's already the case for `openssl rsa -noout
17743 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17744 currently the public key is printed (a decision which was already done by
17745 `openssl dsa -modulus` in the past) which serves a similar purpose.
17746 Additionally the NO_RSA no longer completely removes the whole -modulus
17747 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17750 *Ralf S. Engelschall*
17752 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17753 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17755 *Arne Ansper <arne@ats.cyber.ee>*
17757 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17758 to be added. Now both 'req' and 'ca' can use new objects defined in the
17763 * Add cool BIO that does syslog (or event log on NT).
17765 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17767 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17768 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17769 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17770 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17774 * Add preliminary config info for new extension code.
17778 * Make RSA_NO_PADDING really use no padding.
17780 *Ulf Moeller <ulf@fitug.de>*
17782 * Generate errors when private/public key check is done.
17786 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17787 for some CRL extensions and new objects added.
17791 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17792 key usage extension and fuller support for authority key id.
17796 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17797 padding method for RSA, which is recommended for new applications in PKCS
17798 #1 v2.0 (RFC 2437, October 1998).
17799 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17800 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17801 against Bleichbacher's attack on RSA.
17802 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17805 * Updates to the new SSL compression code
17807 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17809 * Fix so that the version number in the master secret, when passed
17810 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17811 (because the server will not accept higher), that the version number
17812 is 0x03,0x01, not 0x03,0x00
17814 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17816 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17817 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17818 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17822 * Support for RAW extensions where an arbitrary extension can be
17823 created by including its DER encoding. See apps/openssl.cnf for
17828 * Make sure latest Perl versions don't interpret some generated C array
17829 code as Perl array code in the crypto/err/err_genc.pl script.
17831 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17833 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17834 not many people have the assembler. Various Win32 compilation fixes and
17835 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17836 build instructions.
17840 * Modify configure script 'Configure' to automatically create crypto/date.h
17841 file under Win32 and also build pem.h from pem.org. New script
17842 util/mkfiles.pl to create the MINFO file on environments that can't do a
17843 'make files': perl util/mkfiles.pl >MINFO should work.
17847 * Major rework of DES function declarations, in the pursuit of correctness
17848 and purity. As a result, many evil casts evaporated, and some weirdness,
17849 too. You may find this causes warnings in your code. Zapping your evil
17850 casts will probably fix them. Mostly.
17854 * Fix for a typo in asn1.h. Bug fix to object creation script
17855 obj_dat.pl. It considered a zero in an object definition to mean
17856 "end of object": none of the objects in objects.h have any zeros
17857 so it wasn't spotted.
17859 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17861 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17862 Masking (CBCM). In the absence of test vectors, the best I have been able
17863 to do is check that the decrypt undoes the encrypt, so far. Send me test
17864 vectors if you have them.
17868 * Correct calculation of key length for export ciphers (too much space was
17869 allocated for null ciphers). This has not been tested!
17873 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17874 message is now correct (it understands "crypto" and "ssl" on its
17875 command line). There is also now an "update" option. This will update
17876 the util/ssleay.num and util/libeay.num files with any new functions.
17878 perl util/mkdef.pl crypto ssl update
17879 it will update them.
17883 * Overhauled the Perl interface:
17884 - ported BN stuff to OpenSSL's different BN library
17885 - made the perl/ source tree CVS-aware
17886 - renamed the package from SSLeay to OpenSSL (the files still contain
17887 their history because I've copied them in the repository)
17888 - removed obsolete files (the test scripts will be replaced
17889 by better Test::Harness variants in the future)
17891 *Ralf S. Engelschall*
17893 * First cut for a very conservative source tree cleanup:
17894 1. merge various obsolete readme texts into doc/ssleay.txt
17895 where we collect the old documents and readme texts.
17896 2. remove the first part of files where I'm already sure that we no
17897 longer need them because of three reasons: either they are just temporary
17898 files which were left by Eric or they are preserved original files where
17899 I've verified that the diff is also available in the CVS via "cvs diff
17900 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17901 the crypto/md/ stuff).
17903 *Ralf S. Engelschall*
17905 * More extension code. Incomplete support for subject and issuer alt
17906 name, issuer and authority key id. Change the i2v function parameters
17907 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17908 what that's for :-) Fix to ASN1 macro which messed up
17909 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17913 * Preliminary support for ENUMERATED type. This is largely copied from the
17918 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17920 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17922 * Make sure `make rehash` target really finds the `openssl` program.
17924 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17926 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17927 like to hear about it if this slows down other processors.
17931 * Add CygWin32 platform information to Configure script.
17933 *Alan Batie <batie@aahz.jf.intel.com>*
17935 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17937 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17939 * New program nseq to manipulate netscape certificate sequences
17943 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17948 * Fixes to BN code. Previously the default was to define BN_RECURSION
17949 but the BN code had some problems that would cause failures when
17950 doing certificate verification and some other functions.
17952 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17954 * Add ASN1 and PEM code to support netscape certificate sequences.
17958 * Add ASN1 and PEM code to support netscape certificate sequences.
17962 * Add several PKIX and private extended key usage OIDs.
17966 * Modify the 'ca' program to handle the new extension code. Modify
17967 openssl.cnf for new extension format, add comments.
17971 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17972 and add a sample to openssl.cnf so req -x509 now adds appropriate
17977 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17978 error code, add initial support to X509_print() and x509 application.
17982 * Takes a deep breath and start adding X509 V3 extension support code. Add
17983 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17984 stuff is currently isolated and isn't even compiled yet.
17988 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17989 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17990 Removed the versions check from X509 routines when loading extensions:
17991 this allows certain broken certificates that don't set the version
17992 properly to be processed.
17996 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17997 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17998 can still be regenerated with "make depend".
18002 * Spelling mistake in C version of CAST-128.
18004 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18006 * Changes to the error generation code. The perl script err-code.pl
18007 now reads in the old error codes and retains the old numbers, only
18008 adding new ones if necessary. It also only changes the .err files if new
18009 codes are added. The makefiles have been modified to only insert errors
18010 when needed (to avoid needlessly modifying header files). This is done
18011 by only inserting errors if the .err file is newer than the auto generated
18012 C file. To rebuild all the error codes from scratch (the old behaviour)
18013 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18014 or delete all the .err files.
18018 * CAST-128 was incorrectly implemented for short keys. The C version has
18019 been fixed, but is untested. The assembler versions are also fixed, but
18020 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18021 to regenerate it if needed.
18022 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18023 Hagino <itojun@kame.net>*
18025 * File was opened incorrectly in randfile.c.
18027 *Ulf Möller <ulf@fitug.de>*
18029 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18030 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18031 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18032 al: it's just almost always a UTCTime. Note this patch adds new error
18033 codes so do a "make errors" if there are problems.
18037 * Correct Linux 1 recognition in config.
18039 *Ulf Möller <ulf@fitug.de>*
18041 * Remove pointless MD5 hash when using DSA keys in ca.
18043 *Anonymous <nobody@replay.com>*
18045 * Generate an error if given an empty string as a cert directory. Also
18046 generate an error if handed NULL (previously returned 0 to indicate an
18047 error, but didn't set one).
18049 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18051 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18055 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18056 parameters. This was causing a warning which killed off the Win32 compile.
18060 * Remove C++ style comments from crypto/bn/bn_local.h.
18062 *Neil Costigan <neil.costigan@celocom.com>*
18064 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18065 based on a text string, looking up short and long names and finally
18066 "dot" format. The "dot" format stuff didn't work. Added new function
18067 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18068 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18069 OID is not part of the table.
18073 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18074 X509_LOOKUP_by_alias().
18078 * Sort openssl functions by name.
18082 * Get the gendsa program working (hopefully) and add it to app list. Remove
18083 encryption from sample DSA keys (in case anyone is interested the password
18088 * Make *all* `*_free` functions accept a NULL pointer.
18090 *Frans Heymans <fheymans@isaserver.be>*
18092 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18095 *Anonymous <nobody@replay.com>*
18097 * s_server should send the CAfile as acceptable CAs, not its own cert.
18099 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18101 * Don't blow it for numeric -newkey arguments to apps/req.
18103 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18105 * Temp key "for export" tests were wrong in s3_srvr.c.
18107 *Anonymous <nobody@replay.com>*
18109 * Add prototype for temp key callback functions
18110 SSL_CTX_set_tmp_{rsa,dh}_callback().
18114 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18115 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18119 * X509_name_add_entry() freed the wrong thing after an error.
18121 *Arne Ansper <arne@ats.cyber.ee>*
18123 * rsa_eay.c would attempt to free a NULL context.
18125 *Arne Ansper <arne@ats.cyber.ee>*
18127 * BIO_s_socket() had a broken should_retry() on Windoze.
18129 *Arne Ansper <arne@ats.cyber.ee>*
18131 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18133 *Arne Ansper <arne@ats.cyber.ee>*
18135 * Make sure the already existing X509_STORE->depth variable is initialized
18136 in X509_STORE_new(), but document the fact that this variable is still
18137 unused in the certificate verification process.
18139 *Ralf S. Engelschall*
18141 * Fix the various library and apps files to free up pkeys obtained from
18142 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18146 * Fix reference counting in X509_PUBKEY_get(). This makes
18147 demos/maurice/example2.c work, amongst others, probably.
18149 *Steve Henson and Ben Laurie*
18151 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18152 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18153 are no longer created. This way we have a single and consistent command
18154 line interface `openssl <command>`, similar to `cvs <command>`.
18156 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18158 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18159 BIT STRING wrapper always have zero unused bits.
18163 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18167 * Make the top-level INSTALL documentation easier to understand.
18171 * Makefiles updated to exit if an error occurs in a sub-directory
18172 make (including if user presses ^C) [Paul Sutton]
18174 * Make Montgomery context stuff explicit in RSA data structure.
18178 * Fix build order of pem and err to allow for generated pem.h.
18182 * Fix renumbering bug in X509_NAME_delete_entry().
18186 * Enhanced the err-ins.pl script so it makes the error library number
18187 global and can add a library name. This is needed for external ASN1 and
18188 other error libraries.
18192 * Fixed sk_insert which never worked properly.
18196 * Fix ASN1 macros so they can handle indefinite length constructed
18197 EXPLICIT tags. Some non standard certificates use these: they can now
18202 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18203 into a single doc/ssleay.txt bundle. This way the information is still
18204 preserved but no longer messes up this directory. Now it's new room for
18205 the new set of documentation files.
18207 *Ralf S. Engelschall*
18209 * SETs were incorrectly DER encoded. This was a major pain, because they
18210 shared code with SEQUENCEs, which aren't coded the same. This means that
18211 almost everything to do with SETs or SEQUENCEs has either changed name or
18212 number of arguments.
18214 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18216 * Fix test data to work with the above.
18220 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18221 was already fixed by Eric for 0.9.1 it seems.
18223 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18225 * Autodetect FreeBSD3.
18229 * Fix various bugs in Configure. This affects the following platforms:
18233 unixware-2.0-pentium
18238 * Eliminate generated files from CVS. Reorder tests to regenerate files
18239 before they are needed.
18243 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18247 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18249 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18250 changed SSLeay to OpenSSL in version strings.
18252 *Ralf S. Engelschall*
18254 * Some fixups to the top-level documents.
18258 * Fixed the nasty bug where rsaref.h was not found under compile-time
18259 because the symlink to include/ was missing.
18261 *Ralf S. Engelschall*
18263 * Incorporated the popular no-RSA/DSA-only patches
18264 which allow to compile a RSA-free SSLeay.
18266 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18268 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18269 when "ssleay" is still not found.
18271 *Ralf S. Engelschall*
18273 * Added more platforms to Configure: Cray T3E, HPUX 11,
18275 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18277 * Updated the README file.
18279 *Ralf S. Engelschall*
18281 * Added various .cvsignore files in the CVS repository subdirs
18282 to make a "cvs update" really silent.
18284 *Ralf S. Engelschall*
18286 * Recompiled the error-definition header files and added
18287 missing symbols to the Win32 linker tables.
18289 *Ralf S. Engelschall*
18291 * Cleaned up the top-level documents;
18292 o new files: CHANGES and LICENSE
18293 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18294 o merged COPYRIGHT into LICENSE
18295 o removed obsolete TODO file
18296 o renamed MICROSOFT to INSTALL.W32
18298 *Ralf S. Engelschall*
18300 * Removed dummy files from the 0.9.1b source tree:
18301 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18302 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18303 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18304 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18305 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18307 *Ralf S. Engelschall*
18309 * Added various platform portability fixes.
18313 * The Genesis of the OpenSSL rpject:
18314 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18315 Young and Tim J. Hudson created while they were working for C2Net until
18318 *The OpenSSL Project*
18320 ### Changes between 0.9.0b and 0.9.1b [not released]
18322 * Updated a few CA certificates under certs/
18326 * Changed some BIGNUM api stuff.
18330 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18331 DGUX x86, Linux Alpha, etc.
18335 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18336 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18341 * Add -strparse option to asn1pars program which parses nested
18344 *Dr Stephen Henson <shenson@bigfoot.com>*
18346 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18350 * DSA fix for "ca" program.
18354 * Added "-genkey" option to "dsaparam" program.
18358 * Added RIPE MD160 (rmd160) message digest.
18362 * Added -a (all) option to "ssleay version" command.
18366 * Added PLATFORM define which is the id given to Configure.
18370 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18374 * Extended the ASN.1 parser routines.
18378 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18382 * Added a BN_CTX to the BN library.
18386 * Fixed the weak key values in DES library
18390 * Changed API in EVP library for cipher aliases.
18394 * Added support for RC2/64bit cipher.
18398 * Converted the lhash library to the crypto/mem.c functions.
18402 * Added more recognized ASN.1 object ids.
18406 * Added more RSA padding checks for SSL/TLS.
18410 * Added BIO proxy/filter functionality.
18414 * Added extra_certs to SSL_CTX which can be used
18415 send extra CA certificates to the client in the CA cert chain sending
18416 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18420 * Now Fortezza is denied in the authentication phase because
18421 this is key exchange mechanism is not supported by SSLeay at all.
18425 * Additional PKCS1 checks.
18429 * Support the string "TLSv1" for all TLS v1 ciphers.
18433 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18434 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18438 * Fixed a few memory leaks.
18442 * Fixed various code and comment typos.
18446 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18447 bytes sent in the client random.
18449 *Edward Bishop <ebishop@spyglass.com>*
18453 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18454 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18455 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18456 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18457 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18458 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18459 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18460 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18461 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18462 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18463 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18464 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18465 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18466 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18467 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18468 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18469 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18470 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18471 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18472 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18473 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18474 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18475 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18476 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18477 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18478 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18479 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18480 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18481 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18482 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18483 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18484 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18485 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18486 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18487 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18488 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18489 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18490 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18491 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18492 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18493 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18494 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18495 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18496 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18497 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18498 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18499 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18500 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18501 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18502 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18503 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18504 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18505 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18506 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18507 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18508 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18509 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18510 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18511 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18512 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18513 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18514 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18515 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18516 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18517 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18518 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18519 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18520 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18521 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18522 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18523 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18524 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18525 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18526 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18527 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18528 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18529 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18530 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18531 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18532 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18533 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18534 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18535 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18536 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18537 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18538 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18539 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18540 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18541 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18542 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18543 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18544 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18545 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18546 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18547 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18548 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18549 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18550 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18551 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18552 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18553 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18554 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18555 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18556 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18557 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18558 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18559 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18560 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18561 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18562 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18563 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18564 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18565 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18566 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18567 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18568 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18569 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18570 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18571 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18572 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18573 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18574 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18575 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18576 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18577 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18578 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18579 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18580 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18581 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18582 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18583 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18584 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18585 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18586 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18587 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18588 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18589 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18590 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18591 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18592 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18593 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18594 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18595 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18596 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18597 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18598 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18599 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18600 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18601 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18602 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18603 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18604 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18605 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18606 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18607 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18608 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18609 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18610 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18611 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18612 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655