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 * Changed all "STACK" functions to be macros instead of inline functions. Macro
27 parameters are still checked for type safety at compile time via helper
32 * Remove the RAND_DRBG API
34 The RAND_DRBG API did not fit well into the new provider concept as
35 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
36 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
37 and some of its API calls are rather low-level. This holds in particular
38 for the callback mechanism (RAND_DRBG_set_callbacks()).
40 Adding a compatibility layer to continue supporting the RAND_DRBG API as
41 a legacy API for a regular deprecation period turned out to come at the
42 price of complicating the new provider API unnecessarily. Since the
43 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
46 *Paul Dale and Matthias St. Pierre*
48 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
49 as well as actual hostnames.
53 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
54 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
55 conversely, silently ignore DTLS protocol version bounds when configuring
56 TLS-based contexts. The commands can be repeated to set bounds of both
57 types. The same applies with the corresponding "min_protocol" and
58 "max_protocol" command-line switches, in case some application uses both TLS
61 SSL_CTX instances that are created for a fixed protocol version (e.g.
62 TLSv1_server_method()) also silently ignore version bounds. Previously
63 attempts to apply bounds to these protocol versions would result in an
64 error. Now only the "version-flexible" SSL_CTX instances are subject to
65 limits in configuration files in command-line options.
69 * Deprecated the `ENGINE` API. Engines should be replaced with providers
74 * Reworked the recorded ERR codes to make better space for system errors.
75 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
76 given code is a system error (true) or an OpenSSL error (false).
80 * Reworked the test perl framework to better allow parallel testing.
82 *Nicola Tuveri and David von Oheimb*
84 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
85 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
89 * 'Configure' has been changed to figure out the configuration target if
90 none is given on the command line. Consequently, the 'config' script is
91 now only a mere wrapper. All documentation is changed to only mention
94 *Rich Salz and Richard Levitte*
96 * Added a library context that applications as well as other
97 libraries can use to form a separate context within which libcrypto
98 operations are performed.
100 There are two ways this can be used:
102 - Directly, by passing a library context to functions that take
103 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
105 - Indirectly, by creating a new library context and then assigning
106 it as the new default, with `OPENSSL_CTX_set0_default`.
108 All public OpenSSL functions that take an `OPENSSL_CTX` pointer,
109 apart from the functions directly related to `OPENSSL_CTX`, accept
110 NULL to indicate that the default library context should be used.
112 Library code that changes the default library context using
113 `OPENSSL_CTX_set0_default` should take care to restore it with a
114 second call before returning to the caller.
118 * Handshake now fails if Extended Master Secret extension is dropped
123 * Dropped interactive mode from the 'openssl' program. From now on,
124 the `openssl` command without arguments is equivalent to `openssl
129 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
130 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
131 While the old function names have been retained for backward compatibility
132 they should not be used in new developments
133 because their return values are confusing: Unlike other `_cmp()` functions
134 they do not return 0 in case their arguments are equal.
138 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
139 EC_GROUP_get_field_type().
143 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
144 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
145 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
146 Applications should rely on the library automatically assigning a suitable
147 EC_METHOD internally upon EC_GROUP construction.
151 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
152 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
153 assigned internally without application intervention.
154 Users of EC_GROUP_new() should switch to a different suitable constructor.
158 * Add CAdES-BES signature verification support, mostly derived
159 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
161 *Filipe Raimundo da Silva*
163 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
167 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
168 functions are not widely used and now OpenSSL automatically perform this
169 conversion when needed.
173 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
174 EC_KEY_precompute_mult(). These functions are not widely used and
175 applications should instead switch to named curves which OpenSSL has
176 hardcoded lookup tables for.
180 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
181 should instead use the L<EC_POINT_mul(3)> function.
185 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
186 that are not applicable to the new provider model. Applications should
187 instead use EVP_default_properties_is_fips_enabled() and
188 EVP_default_properties_enable_fips().
192 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
193 is set, an unexpected EOF is ignored, it pretends a close notify was received
194 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
198 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
199 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
200 used and applications should instead use the
201 L<EC_POINT_set_affine_coordinates(3)> and
202 L<EC_POINT_get_affine_coordinates(3)> functions.
206 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
207 arrays to be more easily constructed via a series of utility functions.
208 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
209 the various push functions and finally convert to a passable OSSL_PARAM
210 array using OSSL_PARAM_BLD_to_param().
214 * The security strength of SHA1 and MD5 based signatures in TLS has been
215 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
216 working at the default security level of 1 and instead requires security
217 level 0. The security level can be changed either using the cipher string
218 with @SECLEVEL, or calling SSL_CTX_set_security_level().
222 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
223 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
224 internal keys, if they correspond to one of those built in types.
228 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
229 contain a provider side internal key.
233 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
234 They are old functions that we don't use, and that you could disable with
235 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
239 * Project text documents not yet having a proper file name extension
240 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
241 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
245 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
246 have been converted to Markdown with the goal to produce documents
247 which not only look pretty when viewed online in the browser, but
248 remain well readable inside a plain text editor.
250 To achieve this goal, a 'minimalistic' Markdown style has been applied
251 which avoids formatting elements that interfere too much with the
252 reading flow in the text file. For example, it
254 * avoids [ATX headings][] and uses [setext headings][] instead
255 (which works for `<h1>` and `<h2>` headings only).
256 * avoids [inline links][] and uses [reference links][] instead.
257 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
259 [ATX headings]: https://github.github.com/gfm/#atx-headings
260 [setext headings]: https://github.github.com/gfm/#setext-headings
261 [inline links]: https://github.github.com/gfm/#inline-link
262 [reference links]: https://github.github.com/gfm/#reference-link
263 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
264 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
266 *Matthias St. Pierre*
268 * The test suite is changed to preserve results of each test recipe.
269 A new directory test-runs/ with subdirectories named like the
270 test recipes are created in the build tree for this purpose.
274 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
275 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
276 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
278 *David von Oheimb, Martin Peylo*
280 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
281 The legacy OCSP-focused and only partly documented API is retained.
282 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
286 * Added `util/check-format.pl`, a tool for checking adherence to the
287 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
288 The checks performed are incomplete and yield some false positives.
289 Still the tool should be useful for detecting most typical glitches.
293 * BIO_do_connect and BIO_do_handshake have been extended:
294 If domain name resolution yields multiple IP addresses all of them are tried
295 after connect() failures.
299 * All of the low level RSA functions have been deprecated including:
301 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
302 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
303 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
304 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
305 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
306 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
307 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
308 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
309 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
310 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
311 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
312 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
313 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
314 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
315 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
316 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
317 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
318 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
319 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
320 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
321 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
322 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
323 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
324 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
325 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
326 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
327 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
328 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
330 Use of these low level functions has been informally discouraged for a long
331 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
332 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
333 L<EVP_PKEY_decrypt(3)>.
337 * X509 certificates signed using SHA1 are no longer allowed at security
339 In TLS/SSL the default security level is 1. It can be set either
340 using the cipher string with @SECLEVEL, or calling
341 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
342 a call to SSL_CTX_use_certificate() will fail if the security level is not
344 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
345 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
350 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
351 modified to use PKEY APIs. These commands are now in maintenance mode
352 and no new features will be added to them.
356 * The command line utility rsautl has been deprecated.
357 Instead use the pkeyutl program.
361 * The command line utilities genrsa and rsa have been modified to use PKEY
362 APIs. They now write PKCS#8 keys by default. These commands are now in
363 maintenance mode and no new features will be added to them.
367 * All of the low level DH functions have been deprecated including:
369 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
370 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
371 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
372 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
373 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
374 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
375 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
376 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
377 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
378 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
379 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
380 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
381 DH_meth_get_generate_params and DH_meth_set_generate_params.
383 Use of these low level functions has been informally discouraged for a long
384 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
385 and L<EVP_PKEY_derive(3)>.
389 * All of the low level DSA functions have been deprecated including:
391 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
392 DSA_get_default_method, DSA_set_method, DSA_get_method,
393 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
394 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
395 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
396 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
397 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
398 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
399 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
400 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
401 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
402 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
403 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
406 Use of these low level functions has been informally discouraged for a long
407 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
408 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
412 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
413 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
414 This means that applications don't have to look at the curve NID and
415 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
416 However, they still can, that EVP_PKEY_set_alias_type() call acts as
417 a no-op when the EVP_PKEY is already of the given type.
419 Parameter and key generation is also reworked to make it possible
420 to generate EVP_PKEY_SM2 parameters and keys without having to go
421 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
422 However, code that does the latter will still work as before.
426 * Deprecated low level ECDH and ECDSA functions. These include:
428 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
429 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
432 Use of these low level functions has been informally discouraged for a long
433 time. Instead applications should use the EVP_PKEY_derive(3),
434 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
438 * Deprecated the EC_KEY_METHOD functions. These include:
440 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
441 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
442 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
443 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
444 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
445 EC_KEY_METHOD_get_verify.
447 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
451 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
452 and EVP_PKEY_decrypt() instead.
453 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
454 and EVP_PKEY_encrypt() instead.
458 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
459 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
460 a new formulation to include all the things it can be used for,
461 as well as words of caution.
465 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
466 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
470 * All of the low level HMAC functions have been deprecated including:
472 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
473 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
476 Use of these low level functions has been informally discouraged for a long
477 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
478 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
479 and L<EVP_MAC_final(3)>.
483 * Over two thousand fixes were made to the documentation, including:
484 - Common options (such as -rand/-writerand, TLS version control, etc)
485 were refactored and point to newly-enhanced descriptions in openssl.pod.
486 - Added style conformance for all options (with help from Richard Levitte),
487 documented all reported missing options, added a CI build to check
488 that all options are documented and that no unimplemented options
490 - Documented some internals, such as all use of environment variables.
491 - Addressed all internal broken L<> references.
495 * All of the low level CMAC functions have been deprecated including:
497 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
498 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
500 Use of these low level functions has been informally discouraged for a long
501 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
502 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
503 and L<EVP_MAC_final(3)>.
507 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
508 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
511 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
512 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
513 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
514 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
515 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
516 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
517 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
518 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
519 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
520 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
522 Use of these low level functions has been informally discouraged
523 for a long time. Applications should use the EVP_DigestInit_ex(3),
524 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
528 * Corrected the documentation of the return values from the `EVP_DigestSign*`
529 set of functions. The documentation mentioned negative values for some
530 errors, but this was never the case, so the mention of negative values
533 Code that followed the documentation and thereby check with something
534 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
538 * All of the low level cipher functions have been deprecated including:
540 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
541 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
542 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
543 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
544 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
545 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
546 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
547 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
548 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
549 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
550 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
551 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
552 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
553 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
554 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
555 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
556 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
557 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
558 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
559 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
560 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
561 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
562 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
563 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
564 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
565 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
566 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
567 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
568 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
570 Use of these low level functions has been informally discouraged for
571 a long time. Applications should use the high level EVP APIs, e.g.
572 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
573 equivalently named decrypt functions instead.
575 *Matt Caswell and Paul Dale*
577 * Removed include/openssl/opensslconf.h.in and replaced it with
578 include/openssl/configuration.h.in, which differs in not including
579 <openssl/macros.h>. A short header include/openssl/opensslconf.h
580 was added to include both.
582 This allows internal hacks where one might need to modify the set
583 of configured macros, for example this if deprecated symbols are
584 still supposed to be available internally:
586 #include <openssl/configuration.h>
588 #undef OPENSSL_NO_DEPRECATED
589 #define OPENSSL_SUPPRESS_DEPRECATED
591 #include <openssl/macros.h>
593 This should not be used by applications that use the exported
594 symbols, as that will lead to linking errors.
598 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
599 used in exponentiation with 512-bit moduli. No EC algorithms are
600 affected. Analysis suggests that attacks against 2-prime RSA1024,
601 3-prime RSA1536, and DSA1024 as a result of this defect would be very
602 difficult to perform and are not believed likely. Attacks against DH512
603 are considered just feasible. However, for an attack the target would
604 have to re-use the DH512 private key, which is not recommended anyway.
605 Also applications directly using the low level API BN_mod_exp may be
606 affected if they use BN_FLG_CONSTTIME.
611 * Most memory-debug features have been deprecated, and the functionality
612 replaced with no-ops.
616 * Added documentation for the STACK API.
620 * Introduced a new method type and API, OSSL_ENCODER, to
621 represent generic encoders. An implementation is expected to
622 be able to encode an object associated with a given name (such
623 as an algorithm name for an asymmetric key) into forms given by
624 implementation properties.
626 Encoders are primarily used from inside libcrypto, through
627 calls to functions like EVP_PKEY_print_private(),
628 PEM_write_bio_PrivateKey() and similar.
630 Encoders are specified in such a way that they can be made to
631 directly handle the provider side portion of an object, if this
632 provider side part comes from the same provider as the encoder
633 itself, but can also be made to handle objects in parametrized
634 form (as an OSSL_PARAM array of data). This allows a provider to
635 offer generic encoders as a service for any other provider.
639 * Added a .pragma directive to the syntax of configuration files, to
640 allow varying behavior in a supported and predictable manner.
641 Currently added pragma:
645 This allows dollar signs to be a keyword character unless it's
646 followed by a opening brace or parenthesis. This is useful for
647 platforms where dollar signs are commonly used in names, such as
648 volume names and system directory names on VMS.
652 * Added functionality to create an EVP_PKEY from user data. This
653 is effectively the same as creating a RSA, DH or DSA object and
654 then assigning them to an EVP_PKEY, but directly using algorithm
655 agnostic EVP functions. A benefit is that this should be future
656 proof for public key algorithms to come.
660 * Change the interpretation of the '--api' configuration option to
661 mean that this is a desired API compatibility level with no
662 further meaning. The previous interpretation, that this would
663 also mean to remove all deprecated symbols up to and including
664 the given version, no requires that 'no-deprecated' is also used
665 in the configuration.
667 When building applications, the desired API compatibility level
668 can be set with the OPENSSL_API_COMPAT macro like before. For
669 API compatibility version below 3.0, the old style numerical
670 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
671 For version 3.0 and on, the value is expected to be the decimal
672 value calculated from the major and minor version like this:
674 MAJOR * 10000 + MINOR * 100
678 -DOPENSSL_API_COMPAT=30000 For 3.0
679 -DOPENSSL_API_COMPAT=30200 For 3.2
681 To hide declarations that are deprecated up to and including the
682 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
683 given when building the application as well.
687 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
688 access to certificate and CRL stores via URIs and OSSL_STORE
691 This adds the following functions:
693 - X509_LOOKUP_store()
694 - X509_STORE_load_file()
695 - X509_STORE_load_path()
696 - X509_STORE_load_store()
697 - SSL_add_store_cert_subjects_to_stack()
698 - SSL_CTX_set_default_verify_store()
699 - SSL_CTX_load_verify_file()
700 - SSL_CTX_load_verify_dir()
701 - SSL_CTX_load_verify_store()
705 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
706 The presence of this system service is determined at run-time.
710 * Added functionality to create an EVP_PKEY context based on data
711 for methods from providers. This takes an algorithm name and a
712 property query string and simply stores them, with the intent
713 that any operation that uses this context will use those strings
714 to fetch the needed methods implicitly, thereby making the port
715 of application written for pre-3.0 OpenSSL easier.
719 * The undocumented function NCONF_WIN32() has been deprecated; for
720 conversion details see the HISTORY section of doc/man5/config.pod
724 * Introduced the new functions EVP_DigestSignInit_ex() and
725 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
726 EVP_DigestVerifyUpdate() have been converted to functions. See the man
727 pages for further details.
731 * Over two thousand fixes were made to the documentation, including:
732 adding missing command flags, better style conformance, documentation
735 *Rich Salz, Richard Levitte*
737 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
738 X25519, X448, Ed25519 and Ed448.
742 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
747 * Deprecated the public definition of ERR_STATE as well as the function
748 ERR_get_state(). This is done in preparation of making ERR_STATE an
753 * Added ERR functionality to give callers access to the stored function
754 names that have replaced the older function code based functions.
756 New functions are ERR_get_error_func(), ERR_peek_error_func(),
757 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
758 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
759 and ERR_peek_last_error_all().
761 These functions have become deprecated: ERR_get_error_line_data(),
762 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
763 ERR_func_error_string().
767 * Extended testing to be verbose for failing tests only. The make variables
768 VERBOSE_FAILURE or VF can be used to enable this:
770 $ make VF=1 test # Unix
771 $ mms /macro=(VF=1) test ! OpenVMS
772 $ nmake VF=1 test # Windows
776 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
777 used even when parsing explicit parameters, when loading a encoded key
778 or calling `EC_GROUP_new_from_ecpkparameters()`/
779 `EC_GROUP_new_from_ecparameters()`.
780 This prevents bypass of security hardening and performance gains,
781 especially for curves with specialized EC_METHODs.
782 By default, if a key encoded with explicit parameters is loaded and later
783 encoded, the output is still encoded with explicit parameters, even if
784 internally a "named" EC_GROUP is used for computation.
788 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
789 this change, EC_GROUP_set_generator would accept order and/or cofactor as
790 NULL. After this change, only the cofactor parameter can be NULL. It also
791 does some minimal sanity checks on the passed order.
796 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
797 An attack is simple, if the first CMS_recipientInfo is valid but the
798 second CMS_recipientInfo is chosen ciphertext. If the second
799 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
800 encryption key will be replaced by garbage, and the message cannot be
801 decoded, but if the RSA decryption fails, the correct encryption key is
802 used and the recipient will not notice the attack.
803 As a work around for this potential attack the length of the decrypted
804 key must be equal to the cipher default key length, in case the
805 certifiate is not given and all recipientInfo are tried out.
806 The old behaviour can be re-enabled in the CMS code by setting the
807 CMS_DEBUG_DECRYPT flag.
811 * Early start up entropy quality from the DEVRANDOM seed source has been
812 improved for older Linux systems. The RAND subsystem will wait for
813 /dev/random to be producing output before seeding from /dev/urandom.
814 The seeded state is stored for future library initialisations using
815 a system global shared memory segment. The shared memory identifier
816 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
817 the desired value. The default identifier is 114.
821 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
822 when primes for RSA keys are computed.
823 Since we previously always generated primes == 2 (mod 3) for RSA keys,
824 the 2-prime and 3-prime RSA modules were easy to distinguish, since
825 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
826 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
827 This avoids possible fingerprinting of newly generated RSA modules.
831 * Correct the extended master secret constant on EBCDIC systems. Without this
832 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
833 negotiate EMS will fail. Unfortunately this also means that TLS connections
834 between EBCDIC systems with this fix, and EBCDIC systems without this
835 fix will fail if they negotiate EMS.
839 * Changed the library initialisation so that the config file is now loaded
840 by default. This was already the case for libssl. It now occurs for both
841 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
842 OPENSSL_init_crypto() to suppress automatic loading of a config file.
846 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
847 where the former acts as a replacement for ERR_put_error(), and the
848 latter replaces the combination ERR_put_error()+ERR_add_error_data().
849 ERR_raise_data() adds more flexibility by taking a format string and
850 an arbitrary number of arguments following it, to be processed with
855 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
856 to check if a named provider is loaded and available. When called, it
857 will also activate all fallback providers if such are still present.
861 * Enforce a minimum DH modulus size of 512 bits.
865 * Changed DH parameters to generate the order q subgroup instead of 2q.
866 Previously generated DH parameters are still accepted by DH_check
867 but DH_generate_key works around that by clearing bit 0 of the
868 private key for those. This avoids leaking bit 0 of the private key.
872 * Significantly reduce secure memory usage by the randomness pools.
876 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
881 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
882 algorithms. An implementation of a key exchange algorithm can be obtained
883 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
884 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
885 the older EVP_PKEY_derive_init() function. See the man pages for the new
886 functions for further details.
890 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
894 * Removed the function names from error messages and deprecated the
897 * Removed NextStep support and the macro OPENSSL_UNISTD
901 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
902 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
903 Also removed "export var as function" capability; we do not export
904 variables, only functions.
908 * RC5_32_set_key has been changed to return an int type, with 0 indicating
909 an error and 1 indicating success. In previous versions of OpenSSL this
910 was a void type. If a key was set longer than the maximum possible this
915 * Support SM2 signing and verification schemes with X509 certificate.
919 * Use SHA256 as the default digest for TS query in the ts app.
923 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
924 This checks that the salt length is at least 128 bits, the derived key
925 length is at least 112 bits, and that the iteration count is at least 1000.
926 For backwards compatibility these checks are disabled by default in the
927 default provider, but are enabled by default in the fips provider.
928 To enable or disable these checks use the control
929 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
933 * Default cipher lists/suites are now available via a function, the
934 #defines are deprecated.
938 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
939 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
940 for Windows Store apps easier. Also, the "no-uplink" option has been added.
944 * Join the directories crypto/x509 and crypto/x509v3
948 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
949 This changes the size when using the genpkey app when no size is given. It
950 fixes an omission in earlier changes that changed all RSA, DSA and DH
951 generation apps to use 2048 bits by default.
955 * Added command 'openssl kdf' that uses the EVP_KDF API.
959 * Added command 'openssl mac' that uses the EVP_MAC API.
963 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
964 as default directories. Also added the command 'openssl info'
965 for scripting purposes.
969 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
970 deprecated. These undocumented functions were never integrated into the EVP
971 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
972 Bi-directional IGE mode. These modes were never formally standardised and
973 usage of these functions is believed to be very small. In particular
974 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
975 is ever used. The security implications are believed to be minimal, but
976 this issue was never fixed for backwards compatibility reasons. New code
977 should not use these modes.
981 * Add prediction resistance to the DRBG reseeding process.
985 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
986 mandated by IEEE Std 1619-2018.
990 * Added newline escaping functionality to a filename when using openssl dgst.
991 This output format is to replicate the output format found in the `*sum`
992 checksum programs. This aims to preserve backward compatibility.
994 *Matt Eaton, Richard Levitte, and Paul Dale*
996 * Removed the heartbeat message in DTLS feature, as it has very
997 little usage and doesn't seem to fulfill a valuable purpose.
998 The configuration option is now deprecated.
1002 * Changed the output of 'openssl {digestname} < file' to display the
1003 digest name in its output.
1007 * Added a new generic trace API which provides support for enabling
1008 instrumentation through trace output. This feature is mainly intended
1009 as an aid for developers and is disabled by default. To utilize it,
1010 OpenSSL needs to be configured with the `enable-trace` option.
1012 If the tracing API is enabled, the application can activate trace output
1013 by registering BIOs as trace channels for a number of tracing and debugging
1016 The 'openssl' application has been expanded to enable any of the types
1017 available via environment variables defined by the user, and serves as
1018 one possible example on how to use this functionality.
1020 *Richard Levitte & Matthias St. Pierre*
1022 * Added build tests for C++. These are generated files that only do one
1023 thing, to include one public OpenSSL head file each. This tests that
1024 the public header files can be usefully included in a C++ application.
1026 This test isn't enabled by default. It can be enabled with the option
1027 'enable-buildtest-c++'.
1031 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1035 * Add KMAC to EVP_MAC.
1039 * Added property based algorithm implementation selection framework to
1044 * Added SCA hardening for modular field inversion in EC_GROUP through
1045 a new dedicated field_inv() pointer in EC_METHOD.
1046 This also addresses a leakage affecting conversions from projective
1047 to affine coordinates.
1049 *Billy Bob Brumley, Nicola Tuveri*
1051 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1052 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1053 those algorithms that were already supported through the EVP_PKEY API
1054 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1055 and scrypt are now wrappers that call EVP_KDF.
1059 * Build devcrypto engine as a dynamic engine.
1061 *Eneas U de Queiroz*
1063 * Add keyed BLAKE2 to EVP_MAC.
1067 * Fix a bug in the computation of the endpoint-pair shared secret used
1068 by DTLS over SCTP. This breaks interoperability with older versions
1069 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1070 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1071 interoperability with such broken implementations. However, enabling
1072 this switch breaks interoperability with correct implementations.
1074 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1075 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1079 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1083 * Change the license to the Apache License v2.0.
1087 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1089 - Major releases (indicated by incrementing the MAJOR release number)
1090 may introduce incompatible API/ABI changes.
1091 - Minor releases (indicated by incrementing the MINOR release number)
1092 may introduce new features but retain API/ABI compatibility.
1093 - Patch releases (indicated by incrementing the PATCH number)
1094 are intended for bug fixes and other improvements of existing
1095 features only (like improving performance or adding documentation)
1096 and retain API/ABI compatibility.
1100 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1104 * Remove the 'dist' target and add a tarball building script. The
1105 'dist' target has fallen out of use, and it shouldn't be
1106 necessary to configure just to create a source distribution.
1110 * Recreate the OS390-Unix config target. It no longer relies on a
1111 special script like it did for OpenSSL pre-1.1.0.
1115 * Instead of having the source directories listed in Configure, add
1116 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1121 * Add GMAC to EVP_MAC.
1125 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1129 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1130 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1131 to facilitate the continued use of MACs through raw private keys in
1132 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1136 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1137 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1141 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1142 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1143 are retained for backwards compatibility.
1147 * AES-XTS mode now enforces that its two keys are different to mitigate
1148 the attacked described in "Efficient Instantiations of Tweakable
1149 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1150 Details of this attack can be obtained from:
1151 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1155 * Rename the object files, i.e. give them other names than in previous
1156 versions. Their names now include the name of the final product, as
1157 well as its type mnemonic (bin, lib, shlib).
1161 * Added new option for 'openssl list', '-objects', which will display the
1162 list of built in objects, i.e. OIDs with names.
1166 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1167 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1172 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1173 improves application performance by removing data copies and providing
1174 applications with zero-copy system calls such as sendfile and splice.
1178 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1179 option is set, openssl cleanses (zeroize) plaintext bytes from
1180 internal buffers after delivering them to the application. Note,
1181 the application is still responsible for cleansing other copies
1182 (e.g.: data received by SSL_read(3)).
1186 * `PKCS12_parse` now maintains the order of the parsed certificates
1187 when outputting them via `*ca` (rather than reversing it).
1194 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1196 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1198 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1199 while reading in libssl then we would report an error back to the
1200 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1201 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1202 therefore give a hint as to what went wrong.
1206 * Check that ed25519 and ed448 are allowed by the security level. Previously
1207 signature algorithms not using an MD were not being checked that they were
1208 allowed by the security level.
1212 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1213 was not quite right. The behaviour was not consistent between resumption
1214 and normal handshakes, and also not quite consistent with historical
1215 behaviour. The behaviour in various scenarios has been clarified and
1216 it has been updated to make it match historical behaviour as closely as
1221 * *[VMS only]* The header files that the VMS compilers include automatically,
1222 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1223 that the C++ compiler doesn't understand. This is a shortcoming in the
1224 compiler, but can be worked around with `__cplusplus` guards.
1226 C++ applications that use OpenSSL libraries must be compiled using the
1227 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1228 functions. Otherwise, only functions with symbols of less than 31
1229 characters can be used, as the linker will not be able to successfully
1230 resolve symbols with longer names.
1234 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1235 The presence of this system service is determined at run-time.
1239 * Added newline escaping functionality to a filename when using openssl dgst.
1240 This output format is to replicate the output format found in the `*sum`
1241 checksum programs. This aims to preserve backward compatibility.
1243 *Matt Eaton, Richard Levitte, and Paul Dale*
1245 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1250 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1252 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1253 number generator (RNG). This was intended to include protection in the
1254 event of a fork() system call in order to ensure that the parent and child
1255 processes did not share the same RNG state. However this protection was not
1256 being used in the default case.
1258 A partial mitigation for this issue is that the output from a high
1259 precision timer is mixed into the RNG state so the likelihood of a parent
1260 and child process sharing state is significantly reduced.
1262 If an application already calls OPENSSL_init_crypto() explicitly using
1263 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1266 *Matthias St. Pierre*
1268 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1269 used even when parsing explicit parameters, when loading a encoded key
1270 or calling `EC_GROUP_new_from_ecpkparameters()`/
1271 `EC_GROUP_new_from_ecparameters()`.
1272 This prevents bypass of security hardening and performance gains,
1273 especially for curves with specialized EC_METHODs.
1274 By default, if a key encoded with explicit parameters is loaded and later
1275 encoded, the output is still encoded with explicit parameters, even if
1276 internally a "named" EC_GROUP is used for computation.
1280 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1281 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1282 NULL. After this change, only the cofactor parameter can be NULL. It also
1283 does some minimal sanity checks on the passed order.
1288 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1289 An attack is simple, if the first CMS_recipientInfo is valid but the
1290 second CMS_recipientInfo is chosen ciphertext. If the second
1291 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1292 encryption key will be replaced by garbage, and the message cannot be
1293 decoded, but if the RSA decryption fails, the correct encryption key is
1294 used and the recipient will not notice the attack.
1295 As a work around for this potential attack the length of the decrypted
1296 key must be equal to the cipher default key length, in case the
1297 certifiate is not given and all recipientInfo are tried out.
1298 The old behaviour can be re-enabled in the CMS code by setting the
1299 CMS_DEBUG_DECRYPT flag.
1304 * Early start up entropy quality from the DEVRANDOM seed source has been
1305 improved for older Linux systems. The RAND subsystem will wait for
1306 /dev/random to be producing output before seeding from /dev/urandom.
1307 The seeded state is stored for future library initialisations using
1308 a system global shared memory segment. The shared memory identifier
1309 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1310 the desired value. The default identifier is 114.
1314 * Correct the extended master secret constant on EBCDIC systems. Without this
1315 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1316 negotiate EMS will fail. Unfortunately this also means that TLS connections
1317 between EBCDIC systems with this fix, and EBCDIC systems without this
1318 fix will fail if they negotiate EMS.
1322 * Use Windows installation paths in the mingw builds
1324 Mingw isn't a POSIX environment per se, which means that Windows
1325 paths should be used for installation.
1330 * Changed DH_check to accept parameters with order q and 2q subgroups.
1331 With order 2q subgroups the bit 0 of the private key is not secret
1332 but DH_generate_key works around that by clearing bit 0 of the
1333 private key for those. This avoids leaking bit 0 of the private key.
1337 * Significantly reduce secure memory usage by the randomness pools.
1341 * Revert the DEVRANDOM_WAIT feature for Linux systems
1343 The DEVRANDOM_WAIT feature added a select() call to wait for the
1344 /dev/random device to become readable before reading from the
1345 /dev/urandom device.
1347 It turned out that this change had negative side effects on
1348 performance which were not acceptable. After some discussion it
1349 was decided to revert this feature and leave it up to the OS
1350 resp. the platform maintainer to ensure a proper initialization
1351 during early boot time.
1353 *Matthias St. Pierre*
1355 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1357 * Add build tests for C++. These are generated files that only do one
1358 thing, to include one public OpenSSL head file each. This tests that
1359 the public header files can be usefully included in a C++ application.
1361 This test isn't enabled by default. It can be enabled with the option
1362 'enable-buildtest-c++'.
1366 * Enable SHA3 pre-hashing for ECDSA and DSA.
1370 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1371 This changes the size when using the genpkey app when no size is given. It
1372 fixes an omission in earlier changes that changed all RSA, DSA and DH
1373 generation apps to use 2048 bits by default.
1377 * Reorganize the manual pages to consistently have RETURN VALUES,
1378 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1379 util/fix-doc-nits accordingly.
1381 *Paul Yang, Joshua Lock*
1383 * Add the missing accessor EVP_PKEY_get0_engine()
1387 * Have apps like 's_client' and 's_server' output the signature scheme
1388 along with other cipher suite parameters when debugging.
1392 * Make OPENSSL_config() error agnostic again.
1396 * Do the error handling in RSA decryption constant time.
1400 * Prevent over long nonces in ChaCha20-Poly1305.
1402 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1403 for every encryption operation. RFC 7539 specifies that the nonce value
1404 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1405 and front pads the nonce with 0 bytes if it is less than 12
1406 bytes. However it also incorrectly allows a nonce to be set of up to 16
1407 bytes. In this case only the last 12 bytes are significant and any
1408 additional leading bytes are ignored.
1410 It is a requirement of using this cipher that nonce values are
1411 unique. Messages encrypted using a reused nonce value are susceptible to
1412 serious confidentiality and integrity attacks. If an application changes
1413 the default nonce length to be longer than 12 bytes and then makes a
1414 change to the leading bytes of the nonce expecting the new value to be a
1415 new unique nonce then such an application could inadvertently encrypt
1416 messages with a reused nonce.
1418 Additionally the ignored bytes in a long nonce are not covered by the
1419 integrity guarantee of this cipher. Any application that relies on the
1420 integrity of these ignored leading bytes of a long nonce may be further
1421 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1422 is safe because no such use sets such a long nonce value. However user
1423 applications that use this cipher directly and set a non-default nonce
1424 length to be longer than 12 bytes may be vulnerable.
1426 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1432 * Add DEVRANDOM_WAIT feature for Linux systems
1434 On older Linux systems where the getrandom() system call is not available,
1435 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1436 Contrary to getrandom(), the /dev/urandom device will not block during
1437 early boot when the kernel CSPRNG has not been seeded yet.
1439 To mitigate this known weakness, use select() to wait for /dev/random to
1440 become readable before reading from /dev/urandom.
1442 * Ensure that SM2 only uses SM3 as digest algorithm
1446 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1448 * Change the info callback signals for the start and end of a post-handshake
1449 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1450 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1451 confused by this and assume that a TLSv1.2 renegotiation has started. This
1452 can break KeyUpdate handling. Instead we no longer signal the start and end
1453 of a post handshake message exchange (although the messages themselves are
1454 still signalled). This could break some applications that were expecting
1455 the old signals. However without this KeyUpdate is not usable for many
1460 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1462 * Timing vulnerability in DSA signature generation
1464 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1465 timing side channel attack. An attacker could use variations in the signing
1466 algorithm to recover the private key.
1468 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1473 * Timing vulnerability in ECDSA signature generation
1475 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1476 timing side channel attack. An attacker could use variations in the signing
1477 algorithm to recover the private key.
1479 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1484 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1485 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1486 of two gigabytes and the error handling improved.
1488 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1489 categorized as a normal bug, not a security issue, because the DRBG reseeds
1490 automatically and is fully functional even without additional randomness
1491 provided by the application.
1493 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1495 * Add a new ClientHello callback. Provides a callback interface that gives
1496 the application the ability to adjust the nascent SSL object at the
1497 earliest stage of ClientHello processing, immediately after extensions have
1498 been collected but before they have been processed. In particular, this
1499 callback can adjust the supported TLS versions in response to the contents
1504 * Add SM2 base algorithm support.
1508 * s390x assembly pack: add (improved) hardware-support for the following
1509 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1510 aes-cfb/cfb8, aes-ecb.
1514 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1515 parameter is no longer accepted, as it leads to a corrupt table. NULL
1516 pem_str is reserved for alias entries only.
1520 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1521 step for prime curves. The new implementation is based on formulae from
1522 differential addition-and-doubling in homogeneous projective coordinates
1523 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1524 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1525 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1526 to work in projective coordinates.
1528 *Billy Bob Brumley, Nicola Tuveri*
1530 * Change generating and checking of primes so that the error rate of not
1531 being prime depends on the intended use based on the size of the input.
1532 For larger primes this will result in more rounds of Miller-Rabin.
1533 The maximal error rate for primes with more than 1080 bits is lowered
1536 *Kurt Roeckx, Annie Yousar*
1538 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1542 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1543 moving between systems, and to avoid confusion when a Windows build is
1544 done with mingw vs with MSVC. For POSIX installs, there's still a
1545 symlink or copy named 'tsget' to avoid that confusion as well.
1549 * Revert blinding in ECDSA sign and instead make problematic addition
1550 length-invariant. Switch even to fixed-length Montgomery multiplication.
1554 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1555 step for binary curves. The new implementation is based on formulae from
1556 differential addition-and-doubling in mixed Lopez-Dahab projective
1557 coordinates, modified to independently blind the operands.
1559 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1561 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1562 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1563 EC_METHODs to implement their own specialized "ladder step", to take
1564 advantage of more favorable coordinate systems or more efficient
1565 differential addition-and-doubling algorithms.
1567 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1569 * Modified the random device based seed sources to keep the relevant
1570 file descriptors open rather than reopening them on each access.
1571 This allows such sources to operate in a chroot() jail without
1572 the associated device nodes being available. This behaviour can be
1573 controlled using RAND_keep_random_devices_open().
1577 * Numerous side-channel attack mitigations have been applied. This may have
1578 performance impacts for some algorithms for the benefit of improved
1579 security. Specific changes are noted in this change log by their respective
1584 * AIX shared library support overhaul. Switch to AIX "natural" way of
1585 handling shared libraries, which means collecting shared objects of
1586 different versions and bitnesses in one common archive. This allows to
1587 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1588 doesn't affect the way 3rd party applications are linked, only how
1589 multi-version installation is managed.
1593 * Make ec_group_do_inverse_ord() more robust and available to other
1594 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1595 mitigations are applied to the fallback BN_mod_inverse().
1596 When using this function rather than BN_mod_inverse() directly, new
1597 EC cryptosystem implementations are then safer-by-default.
1601 * Add coordinate blinding for EC_POINT and implement projective
1602 coordinate blinding for generic prime curves as a countermeasure to
1603 chosen point SCA attacks.
1605 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1607 * Add blinding to ECDSA and DSA signatures to protect against side channel
1608 attacks discovered by Keegan Ryan (NCC Group).
1612 * Enforce checking in the pkeyutl command line app to ensure that the input
1613 length does not exceed the maximum supported digest length when performing
1614 a sign, verify or verifyrecover operation.
1618 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1619 I/O in combination with something like select() or poll() will hang. This
1620 can be turned off again using SSL_CTX_clear_mode().
1621 Many applications do not properly handle non-application data records, and
1622 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1623 around the problems in those applications, but can also break some.
1624 It's recommended to read the manpages about SSL_read(), SSL_write(),
1625 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1626 SSL_CTX_set_read_ahead() again.
1630 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1631 now allow empty (zero character) pass phrases.
1635 * Apply blinding to binary field modular inversion and remove patent
1636 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1640 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1641 binary and prime elliptic curves.
1645 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1646 constant time fixed point multiplication.
1650 * Revise elliptic curve scalar multiplication with timing attack
1651 defenses: ec_wNAF_mul redirects to a constant time implementation
1652 when computing fixed point and variable point multiplication (which
1653 in OpenSSL are mostly used with secret scalars in keygen, sign,
1654 ECDH derive operations).
1655 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1658 * Updated CONTRIBUTING
1662 * Updated DRBG / RAND to request nonce and additional low entropy
1663 randomness from the system.
1665 *Matthias St. Pierre*
1667 * Updated 'openssl rehash' to use OpenSSL consistent default.
1671 * Moved the load of the ssl_conf module to libcrypto, which helps
1672 loading engines that libssl uses before libssl is initialised.
1676 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1680 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1682 *Ingo Schwarze, Rich Salz*
1684 * Added output of accepting IP address and port for 'openssl s_server'
1688 * Added a new API for TLSv1.3 ciphersuites:
1689 SSL_CTX_set_ciphersuites()
1690 SSL_set_ciphersuites()
1694 * Memory allocation failures consistently add an error to the error
1699 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1700 in libcrypto when run as setuid/setgid.
1704 * Load any config file by default when libssl is used.
1708 * Added new public header file <openssl/rand_drbg.h> and documentation
1709 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1711 *Matthias St. Pierre*
1713 * QNX support removed (cannot find contributors to get their approval
1714 for the license change).
1718 * TLSv1.3 replay protection for early data has been implemented. See the
1719 SSL_read_early_data() man page for further details.
1723 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1724 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1725 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1726 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1727 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1728 configuration has been separated out. See the ciphers man page or the
1729 SSL_CTX_set_ciphersuites() man page for more information.
1733 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1734 in responder mode now supports the new "-multi" option, which
1735 spawns the specified number of child processes to handle OCSP
1736 requests. The "-timeout" option now also limits the OCSP
1737 responder's patience to wait to receive the full client request
1738 on a newly accepted connection. Child processes are respawned
1739 as needed, and the CA index file is automatically reloaded
1740 when changed. This makes it possible to run the "ocsp" responder
1741 as a long-running service, making the OpenSSL CA somewhat more
1742 feature-complete. In this mode, most diagnostic messages logged
1743 after entering the event loop are logged via syslog(3) rather than
1748 * Added support for X448 and Ed448. Heavily based on original work by
1753 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1754 objects loaded. This adds the functions OSSL_STORE_expect() and
1755 OSSL_STORE_find() as well as needed tools to construct searches and
1756 get the search data out of them.
1760 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1761 version of OpenSSL should review their configuration settings to ensure
1762 that they are still appropriate for TLSv1.3. For further information see:
1763 <https://wiki.openssl.org/index.php/TLS1.3>
1767 * Grand redesign of the OpenSSL random generator
1769 The default RAND method now utilizes an AES-CTR DRBG according to
1770 NIST standard SP 800-90Ar1. The new random generator is essentially
1771 a port of the default random generator from the OpenSSL FIPS 2.0
1772 object module. It is a hybrid deterministic random bit generator
1773 using an AES-CTR bit stream and which seeds and reseeds itself
1774 automatically using trusted system entropy sources.
1776 Some of its new features are:
1777 - Support for multiple DRBG instances with seed chaining.
1778 - The default RAND method makes use of a DRBG.
1779 - There is a public and private DRBG instance.
1780 - The DRBG instances are fork-safe.
1781 - Keep all global DRBG instances on the secure heap if it is enabled.
1782 - The public and private DRBG instance are per thread for lock free
1785 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1787 * Changed Configure so it only says what it does and doesn't dump
1788 so much data. Instead, ./configdata.pm should be used as a script
1789 to display all sorts of configuration data.
1793 * Added processing of "make variables" to Configure.
1797 * Added SHA512/224 and SHA512/256 algorithm support.
1801 * The last traces of Netware support, first removed in 1.1.0, have
1806 * Get rid of Makefile.shared, and in the process, make the processing
1807 of certain files (rc.obj, or the .def/.map/.opt files produced from
1808 the ordinal files) more visible and hopefully easier to trace and
1809 debug (or make silent).
1813 * Make it possible to have environment variable assignments as
1814 arguments to config / Configure.
1818 * Add multi-prime RSA (RFC 8017) support.
1822 * Add SM3 implemented according to GB/T 32905-2016
1823 *Jack Lloyd <jack.lloyd@ribose.com>,*
1824 *Ronald Tse <ronald.tse@ribose.com>,*
1825 *Erick Borsboom <erick.borsboom@ribose.com>*
1827 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1828 as documented in RFC6066.
1829 Based on a patch from Tomasz Moń
1831 *Filipe Raimundo da Silva*
1833 * Add SM4 implemented according to GB/T 32907-2016.
1834 *Jack Lloyd <jack.lloyd@ribose.com>,*
1835 *Ronald Tse <ronald.tse@ribose.com>,*
1836 *Erick Borsboom <erick.borsboom@ribose.com>*
1838 * Reimplement -newreq-nodes and ERR_error_string_n; the
1839 original author does not agree with the license change.
1843 * Add ARIA AEAD TLS support.
1847 * Some macro definitions to support VS6 have been removed. Visual
1848 Studio 6 has not worked since 1.1.0
1852 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1853 without clearing the errors.
1857 * Add "atfork" functions. If building on a system that without
1858 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1859 requirements. The RAND facility now uses/requires this.
1867 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1868 not possible to disable entirely. However, it's still possible to
1869 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1872 To disable, configure with 'no-ui-console'. 'no-ui' is still
1873 possible to use as an alias. Check at compile time with the
1874 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1875 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1879 * Add a STORE module, which implements a uniform and URI based reader of
1880 stores that can contain keys, certificates, CRLs and numerous other
1881 objects. The main API is loosely based on a few stdio functions,
1882 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1883 OSSL_STORE_error and OSSL_STORE_close.
1884 The implementation uses backends called "loaders" to implement arbitrary
1885 URI schemes. There is one built in "loader" for the 'file' scheme.
1889 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1890 then adjusted to work on FreeBSD 8.4 as well.
1891 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1892 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1896 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1897 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1898 error code calls like this:
1900 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1902 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1903 that can be encoded in C. For the foreseeable future, this will only
1906 *Richard Levitte and Tim Hudson*
1908 * Removed BSD cryptodev engine.
1912 * Add a build target 'build_all_generated', to build all generated files
1913 and only that. This can be used to prepare everything that requires
1914 things like perl for a system that lacks perl and then move everything
1915 to that system and do the rest of the build there.
1919 * In the UI interface, make it possible to duplicate the user data. This
1920 can be used by engines that need to retain the data for a longer time
1921 than just the call where this user data is passed.
1925 * Ignore the '-named_curve auto' value for compatibility of applications
1928 *Tomas Mraz <tmraz@fedoraproject.org>*
1930 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1931 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1932 alerts across multiple records (some of which could be empty). In practice
1933 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1934 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1935 support this at all. Supporting it adds significant complexity to the
1936 record layer, and its removal is unlikely to cause interoperability
1941 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1942 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1943 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1948 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1949 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1951 *Richard Levitte, Andy Polyakov*
1953 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1958 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1959 platform rather than 'mingw'.
1963 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1964 success if they are asked to add an object which already exists
1965 in the store. This change cascades to other functions which load
1966 certificates and CRLs.
1970 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1971 facilitate stack unwinding even from assembly subroutines.
1975 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1976 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1980 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1981 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1982 which is the minimum version we support.
1986 * Certificate time validation (X509_cmp_time) enforces stricter
1987 compliance with RFC 5280. Fractional seconds and timezone offsets
1988 are no longer allowed.
1992 * Add support for ARIA
1996 * s_client will now send the Server Name Indication (SNI) extension by
1997 default unless the new "-noservername" option is used. The server name is
1998 based on the host provided to the "-connect" option unless overridden by
1999 using "-servername".
2003 * Add support for SipHash
2007 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2008 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2009 prevent issues where no progress is being made and the peer continually
2010 sends unrecognised record types, using up resources processing them.
2014 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2015 using the algorithm defined in
2016 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2020 * Heartbeat support has been removed; the ABI is changed for now.
2022 *Richard Levitte, Rich Salz*
2024 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2028 * The RSA "null" method, which was partially supported to avoid patent
2029 issues, has been replaced to always returns NULL.
2036 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2038 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2039 used even when parsing explicit parameters, when loading a encoded key
2040 or calling `EC_GROUP_new_from_ecpkparameters()`/
2041 `EC_GROUP_new_from_ecparameters()`.
2042 This prevents bypass of security hardening and performance gains,
2043 especially for curves with specialized EC_METHODs.
2044 By default, if a key encoded with explicit parameters is loaded and later
2045 encoded, the output is still encoded with explicit parameters, even if
2046 internally a "named" EC_GROUP is used for computation.
2050 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2051 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2052 NULL. After this change, only the cofactor parameter can be NULL. It also
2053 does some minimal sanity checks on the passed order.
2058 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2059 An attack is simple, if the first CMS_recipientInfo is valid but the
2060 second CMS_recipientInfo is chosen ciphertext. If the second
2061 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2062 encryption key will be replaced by garbage, and the message cannot be
2063 decoded, but if the RSA decryption fails, the correct encryption key is
2064 used and the recipient will not notice the attack.
2065 As a work around for this potential attack the length of the decrypted
2066 key must be equal to the cipher default key length, in case the
2067 certifiate is not given and all recipientInfo are tried out.
2068 The old behaviour can be re-enabled in the CMS code by setting the
2069 CMS_DEBUG_DECRYPT flag.
2074 * Use Windows installation paths in the mingw builds
2076 Mingw isn't a POSIX environment per se, which means that Windows
2077 paths should be used for installation.
2082 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2084 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2085 This changes the size when using the genpkey app when no size is given. It
2086 fixes an omission in earlier changes that changed all RSA, DSA and DH
2087 generation apps to use 2048 bits by default.
2091 * Prevent over long nonces in ChaCha20-Poly1305.
2093 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2094 for every encryption operation. RFC 7539 specifies that the nonce value
2095 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2096 and front pads the nonce with 0 bytes if it is less than 12
2097 bytes. However it also incorrectly allows a nonce to be set of up to 16
2098 bytes. In this case only the last 12 bytes are significant and any
2099 additional leading bytes are ignored.
2101 It is a requirement of using this cipher that nonce values are
2102 unique. Messages encrypted using a reused nonce value are susceptible to
2103 serious confidentiality and integrity attacks. If an application changes
2104 the default nonce length to be longer than 12 bytes and then makes a
2105 change to the leading bytes of the nonce expecting the new value to be a
2106 new unique nonce then such an application could inadvertently encrypt
2107 messages with a reused nonce.
2109 Additionally the ignored bytes in a long nonce are not covered by the
2110 integrity guarantee of this cipher. Any application that relies on the
2111 integrity of these ignored leading bytes of a long nonce may be further
2112 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2113 is safe because no such use sets such a long nonce value. However user
2114 applications that use this cipher directly and set a non-default nonce
2115 length to be longer than 12 bytes may be vulnerable.
2117 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2123 * Added SCA hardening for modular field inversion in EC_GROUP through
2124 a new dedicated field_inv() pointer in EC_METHOD.
2125 This also addresses a leakage affecting conversions from projective
2126 to affine coordinates.
2128 *Billy Bob Brumley, Nicola Tuveri*
2130 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2131 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2135 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2139 * Remove the 'dist' target and add a tarball building script. The
2140 'dist' target has fallen out of use, and it shouldn't be
2141 necessary to configure just to create a source distribution.
2145 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2147 * Timing vulnerability in DSA signature generation
2149 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2150 timing side channel attack. An attacker could use variations in the signing
2151 algorithm to recover the private key.
2153 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2158 * Timing vulnerability in ECDSA signature generation
2160 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2161 timing side channel attack. An attacker could use variations in the signing
2162 algorithm to recover the private key.
2164 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2169 * Add coordinate blinding for EC_POINT and implement projective
2170 coordinate blinding for generic prime curves as a countermeasure to
2171 chosen point SCA attacks.
2173 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2175 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2177 * Client DoS due to large DH parameter
2179 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2180 malicious server can send a very large prime value to the client. This will
2181 cause the client to spend an unreasonably long period of time generating a
2182 key for this prime resulting in a hang until the client has finished. This
2183 could be exploited in a Denial Of Service attack.
2185 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2190 * Cache timing vulnerability in RSA Key Generation
2192 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2193 a cache timing side channel attack. An attacker with sufficient access to
2194 mount cache timing attacks during the RSA key generation process could
2195 recover the private key.
2197 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2198 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2203 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2204 parameter is no longer accepted, as it leads to a corrupt table. NULL
2205 pem_str is reserved for alias entries only.
2209 * Revert blinding in ECDSA sign and instead make problematic addition
2210 length-invariant. Switch even to fixed-length Montgomery multiplication.
2214 * Change generating and checking of primes so that the error rate of not
2215 being prime depends on the intended use based on the size of the input.
2216 For larger primes this will result in more rounds of Miller-Rabin.
2217 The maximal error rate for primes with more than 1080 bits is lowered
2220 *Kurt Roeckx, Annie Yousar*
2222 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2226 * Add blinding to ECDSA and DSA signatures to protect against side channel
2227 attacks discovered by Keegan Ryan (NCC Group).
2231 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2232 now allow empty (zero character) pass phrases.
2236 * Certificate time validation (X509_cmp_time) enforces stricter
2237 compliance with RFC 5280. Fractional seconds and timezone offsets
2238 are no longer allowed.
2242 * Fixed a text canonicalisation bug in CMS
2244 Where a CMS detached signature is used with text content the text goes
2245 through a canonicalisation process first prior to signing or verifying a
2246 signature. This process strips trailing space at the end of lines, converts
2247 line terminators to CRLF and removes additional trailing line terminators
2248 at the end of a file. A bug in the canonicalisation process meant that
2249 some characters, such as form-feed, were incorrectly treated as whitespace
2250 and removed. This is contrary to the specification (RFC5485). This fix
2251 could mean that detached text data signed with an earlier version of
2252 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2253 signed with a fixed OpenSSL may fail to verify with an earlier version of
2254 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2255 and use the "-binary" flag (for the "cms" command line application) or set
2256 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2260 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2262 * Constructed ASN.1 types with a recursive definition could exceed the stack
2264 Constructed ASN.1 types with a recursive definition (such as can be found
2265 in PKCS7) could eventually exceed the stack given malicious input with
2266 excessive recursion. This could result in a Denial Of Service attack. There
2267 are no such structures used within SSL/TLS that come from untrusted sources
2268 so this is considered safe.
2270 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2276 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2278 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2279 effectively reduced to only comparing the least significant bit of each
2280 byte. This allows an attacker to forge messages that would be considered as
2281 authenticated in an amount of tries lower than that guaranteed by the
2282 security claims of the scheme. The module can only be compiled by the
2283 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2285 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2291 * Add a build target 'build_all_generated', to build all generated files
2292 and only that. This can be used to prepare everything that requires
2293 things like perl for a system that lacks perl and then move everything
2294 to that system and do the rest of the build there.
2298 * Backport SSL_OP_NO_RENGOTIATION
2300 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2301 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2302 changes this is no longer possible in 1.1.0. Therefore the new
2303 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2304 1.1.0 to provide equivalent functionality.
2306 Note that if an application built against 1.1.0h headers (or above) is run
2307 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2308 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2312 * Removed the OS390-Unix config target. It relied on a script that doesn't
2317 * rsaz_1024_mul_avx2 overflow bug on x86_64
2319 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2320 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2321 Analysis suggests that attacks against RSA and DSA as a result of this
2322 defect would be very difficult to perform and are not believed likely.
2323 Attacks against DH1024 are considered just feasible, because most of the
2324 work necessary to deduce information about a private key may be performed
2325 offline. The amount of resources required for such an attack would be
2326 significant. However, for an attack on TLS to be meaningful, the server
2327 would have to share the DH1024 private key among multiple clients, which is
2328 no longer an option since CVE-2016-0701.
2330 This only affects processors that support the AVX2 but not ADX extensions
2331 like Intel Haswell (4th generation).
2333 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2334 was originally found via the OSS-Fuzz project.
2339 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2341 * bn_sqrx8x_internal carry bug on x86_64
2343 There is a carry propagating bug in the x86_64 Montgomery squaring
2344 procedure. No EC algorithms are affected. Analysis suggests that attacks
2345 against RSA and DSA as a result of this defect would be very difficult to
2346 perform and are not believed likely. Attacks against DH are considered just
2347 feasible (although very difficult) because most of the work necessary to
2348 deduce information about a private key may be performed offline. The amount
2349 of resources required for such an attack would be very significant and
2350 likely only accessible to a limited number of attackers. An attacker would
2351 additionally need online access to an unpatched system using the target
2352 private key in a scenario with persistent DH parameters and a private
2353 key that is shared between multiple clients.
2355 This only affects processors that support the BMI1, BMI2 and ADX extensions
2356 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2358 This issue was reported to OpenSSL by the OSS-Fuzz project.
2363 * Malformed X.509 IPAddressFamily could cause OOB read
2365 If an X.509 certificate has a malformed IPAddressFamily extension,
2366 OpenSSL could do a one-byte buffer overread. The most likely result
2367 would be an erroneous display of the certificate in text format.
2369 This issue was reported to OpenSSL by the OSS-Fuzz project.
2374 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2376 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2377 platform rather than 'mingw'.
2381 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2382 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2383 which is the minimum version we support.
2387 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2389 * Encrypt-Then-Mac renegotiation crash
2391 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2392 negotiated where it was not in the original handshake (or vice-versa) then
2393 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2394 and servers are affected.
2396 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2401 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2403 * Truncated packet could crash via OOB read
2405 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2406 cipher is being used, then a truncated packet can cause that host to
2407 perform an out-of-bounds read, usually resulting in a crash.
2409 This issue was reported to OpenSSL by Robert Święcki of Google.
2414 * Bad (EC)DHE parameters cause a client crash
2416 If a malicious server supplies bad parameters for a DHE or ECDHE key
2417 exchange then this can result in the client attempting to dereference a
2418 NULL pointer leading to a client crash. This could be exploited in a Denial
2421 This issue was reported to OpenSSL by Guido Vranken.
2426 * BN_mod_exp may produce incorrect results on x86_64
2428 There is a carry propagating bug in the x86_64 Montgomery squaring
2429 procedure. No EC algorithms are affected. Analysis suggests that attacks
2430 against RSA and DSA as a result of this defect would be very difficult to
2431 perform and are not believed likely. Attacks against DH are considered just
2432 feasible (although very difficult) because most of the work necessary to
2433 deduce information about a private key may be performed offline. The amount
2434 of resources required for such an attack would be very significant and
2435 likely only accessible to a limited number of attackers. An attacker would
2436 additionally need online access to an unpatched system using the target
2437 private key in a scenario with persistent DH parameters and a private
2438 key that is shared between multiple clients. For example this can occur by
2439 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2440 similar to CVE-2015-3193 but must be treated as a separate problem.
2442 This issue was reported to OpenSSL by the OSS-Fuzz project.
2447 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2449 * ChaCha20/Poly1305 heap-buffer-overflow
2451 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2452 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2453 crash. This issue is not considered to be exploitable beyond a DoS.
2455 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2460 * CMS Null dereference
2462 Applications parsing invalid CMS structures can crash with a NULL pointer
2463 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2464 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2465 structure callback if an attempt is made to free certain invalid encodings.
2466 Only CHOICE structures using a callback which do not handle NULL value are
2469 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2474 * Montgomery multiplication may produce incorrect results
2476 There is a carry propagating bug in the Broadwell-specific Montgomery
2477 multiplication procedure that handles input lengths divisible by, but
2478 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2479 and DH private keys are impossible. This is because the subroutine in
2480 question is not used in operations with the private key itself and an input
2481 of the attacker's direct choice. Otherwise the bug can manifest itself as
2482 transient authentication and key negotiation failures or reproducible
2483 erroneous outcome of public-key operations with specially crafted input.
2484 Among EC algorithms only Brainpool P-512 curves are affected and one
2485 presumably can attack ECDH key negotiation. Impact was not analyzed in
2486 detail, because pre-requisites for attack are considered unlikely. Namely
2487 multiple clients have to choose the curve in question and the server has to
2488 share the private key among them, neither of which is default behaviour.
2489 Even then only clients that chose the curve will be affected.
2491 This issue was publicly reported as transient failures and was not
2492 initially recognized as a security issue. Thanks to Richard Morgan for
2493 providing reproducible case.
2498 * Removed automatic addition of RPATH in shared libraries and executables,
2499 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2503 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2505 * Fix Use After Free for large message sizes
2507 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2508 message larger than approx 16k is received then the underlying buffer to
2509 store the incoming message is reallocated and moved. Unfortunately a
2510 dangling pointer to the old location is left which results in an attempt to
2511 write to the previously freed location. This is likely to result in a
2512 crash, however it could potentially lead to execution of arbitrary code.
2514 This issue only affects OpenSSL 1.1.0a.
2516 This issue was reported to OpenSSL by Robert Święcki.
2521 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2523 * OCSP Status Request extension unbounded memory growth
2525 A malicious client can send an excessively large OCSP Status Request
2526 extension. If that client continually requests renegotiation, sending a
2527 large OCSP Status Request extension each time, then there will be unbounded
2528 memory growth on the server. This will eventually lead to a Denial Of
2529 Service attack through memory exhaustion. Servers with a default
2530 configuration are vulnerable even if they do not support OCSP. Builds using
2531 the "no-ocsp" build time option are not affected.
2533 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2538 * SSL_peek() hang on empty record
2540 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2541 sends an empty record. This could be exploited by a malicious peer in a
2542 Denial Of Service attack.
2544 This issue was reported to OpenSSL by Alex Gaynor.
2549 * Excessive allocation of memory in tls_get_message_header() and
2550 dtls1_preprocess_fragment()
2552 A (D)TLS message includes 3 bytes for its length in the header for the
2553 message. This would allow for messages up to 16Mb in length. Messages of
2554 this length are excessive and OpenSSL includes a check to ensure that a
2555 peer is sending reasonably sized messages in order to avoid too much memory
2556 being consumed to service a connection. A flaw in the logic of version
2557 1.1.0 means that memory for the message is allocated too early, prior to
2558 the excessive message length check. Due to way memory is allocated in
2559 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2560 to service a connection. This could lead to a Denial of Service through
2561 memory exhaustion. However, the excessive message length check still takes
2562 place, and this would cause the connection to immediately fail. Assuming
2563 that the application calls SSL_free() on the failed connection in a timely
2564 manner then the 21Mb of allocated memory will then be immediately freed
2565 again. Therefore the excessive memory allocation will be transitory in
2566 nature. This then means that there is only a security impact if:
2568 1) The application does not call SSL_free() in a timely manner in the event
2569 that the connection fails
2571 2) The application is working in a constrained environment where there is
2572 very little free memory
2574 3) The attacker initiates multiple connection attempts such that there are
2575 multiple connections in a state where memory has been allocated for the
2576 connection; SSL_free() has not yet been called; and there is insufficient
2577 memory to service the multiple requests.
2579 Except in the instance of (1) above any Denial Of Service is likely to be
2580 transitory because as soon as the connection fails the memory is
2581 subsequently freed again in the SSL_free() call. However there is an
2582 increased risk during this period of application crashes due to the lack of
2583 memory - which would then mean a more serious Denial of Service.
2585 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2586 (CVE-2016-6307 and CVE-2016-6308)
2590 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2591 had to be removed. Primary reason is that vendor assembler can't
2592 assemble our modules with -KPIC flag. As result it, assembly
2593 support, was not even available as option. But its lack means
2594 lack of side-channel resistant code, which is incompatible with
2595 security by todays standards. Fortunately gcc is readily available
2596 prepackaged option, which we firmly point at...
2600 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2602 * Windows command-line tool supports UTF-8 opt-in option for arguments
2603 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2604 (to any value) allows Windows user to access PKCS#12 file generated
2605 with Windows CryptoAPI and protected with non-ASCII password, as well
2606 as files generated under UTF-8 locale on Linux also protected with
2611 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2612 have been disabled by default and removed from DEFAULT, just like RC4.
2613 See the RC4 item below to re-enable both.
2617 * The method for finding the storage location for the Windows RAND seed file
2618 has changed. First we check %RANDFILE%. If that is not set then we check
2619 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2620 all else fails we fall back to C:\.
2624 * The EVP_EncryptUpdate() function has had its return type changed from void
2625 to int. A return of 0 indicates and error while a return of 1 indicates
2630 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2631 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2632 off the constant time implementation for RSA, DSA and DH have been made
2633 no-ops and deprecated.
2637 * Windows RAND implementation was simplified to only get entropy by
2638 calling CryptGenRandom(). Various other RAND-related tickets
2641 *Joseph Wylie Yandle, Rich Salz*
2643 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2644 and `OPENSSL_LH_`, respectively. The old names are available
2645 with API compatibility. They new names are now completely documented.
2649 * Unify TYPE_up_ref(obj) methods signature.
2650 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2651 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2652 int (instead of void) like all others TYPE_up_ref() methods.
2653 So now these methods also check the return value of CRYPTO_atomic_add(),
2654 and the validity of object reference counter.
2656 *fdasilvayy@gmail.com*
2658 * With Windows Visual Studio builds, the .pdb files are installed
2659 alongside the installed libraries and executables. For a static
2660 library installation, ossl_static.pdb is the associate compiler
2661 generated .pdb file to be used when linking programs.
2665 * Remove openssl.spec. Packaging files belong with the packagers.
2669 * Automatic Darwin/OSX configuration has had a refresh, it will now
2670 recognise x86_64 architectures automatically. You can still decide
2671 to build for a different bitness with the environment variable
2672 KERNEL_BITS (can be 32 or 64), for example:
2674 KERNEL_BITS=32 ./config
2678 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2679 256 bit AES and HMAC with SHA256.
2683 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2687 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2691 * To enable users to have their own config files and build file templates,
2692 Configure looks in the directory indicated by the environment variable
2693 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2694 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2695 name and is used as is.
2699 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2700 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2701 X509_CERT_FILE_CTX was removed.
2705 * "shared" builds are now the default. To create only static libraries use
2706 the "no-shared" Configure option.
2710 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2711 All of these option have not worked for some while and are fundamental
2716 * Make various cleanup routines no-ops and mark them as deprecated. Most
2717 global cleanup functions are no longer required because they are handled
2718 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2719 Explicitly de-initing can cause problems (e.g. where a library that uses
2720 OpenSSL de-inits, but an application is still using it). The affected
2721 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2722 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2723 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2724 COMP_zlib_cleanup().
2728 * --strict-warnings no longer enables runtime debugging options
2729 such as REF_DEBUG. Instead, debug options are automatically
2730 enabled with '--debug' builds.
2732 *Andy Polyakov, Emilia Käsper*
2734 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2735 have been moved out of the public header files. New functions for managing
2736 these have been added.
2740 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2741 objects have been moved out of the public header files. New
2742 functions for managing these have been added.
2746 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2747 have been moved out of the public header files. New functions for managing
2748 these have been added.
2752 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2753 moved out of the public header files. New functions for managing these
2758 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2762 * Removed the mk1mf build scripts.
2766 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2767 it is always safe to #include a header now.
2771 * Removed the aged BC-32 config and all its supporting scripts
2775 * Removed support for Ultrix, Netware, and OS/2.
2779 * Add support for HKDF.
2781 *Alessandro Ghedini*
2783 * Add support for blake2b and blake2s
2787 * Added support for "pipelining". Ciphers that have the
2788 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2789 encryptions/decryptions simultaneously. There are currently no built-in
2790 ciphers with this property but the expectation is that engines will be able
2791 to offer it to significantly improve throughput. Support has been extended
2792 into libssl so that multiple records for a single connection can be
2793 processed in one go (for >=TLS 1.1).
2797 * Added the AFALG engine. This is an async capable engine which is able to
2798 offload work to the Linux kernel. In this initial version it only supports
2799 AES128-CBC. The kernel must be version 4.1.0 or greater.
2803 * OpenSSL now uses a new threading API. It is no longer necessary to
2804 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2805 are two supported threading models: pthreads and windows threads. It is
2806 also possible to configure OpenSSL at compile time for "no-threads". The
2807 old threading API should no longer be used. The functions have been
2808 replaced with "no-op" compatibility macros.
2810 *Alessandro Ghedini, Matt Caswell*
2812 * Modify behavior of ALPN to invoke callback after SNI/servername
2813 callback, such that updates to the SSL_CTX affect ALPN.
2817 * Add SSL_CIPHER queries for authentication and key-exchange.
2821 * Changes to the DEFAULT cipherlist:
2822 - Prefer (EC)DHE handshakes over plain RSA.
2823 - Prefer AEAD ciphers over legacy ciphers.
2824 - Prefer ECDSA over RSA when both certificates are available.
2825 - Prefer TLSv1.2 ciphers/PRF.
2826 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2831 * Change the ECC default curve list to be this, in order: x25519,
2832 secp256r1, secp521r1, secp384r1.
2836 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2837 disabled by default. They can be re-enabled using the
2838 enable-weak-ssl-ciphers option to Configure.
2842 * If the server has ALPN configured, but supports no protocols that the
2843 client advertises, send a fatal "no_application_protocol" alert.
2844 This behaviour is SHALL in RFC 7301, though it isn't universally
2845 implemented by other servers.
2849 * Add X25519 support.
2850 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2851 for public and private key encoding using the format documented in
2852 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2853 key generation and key derivation.
2855 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2860 * Deprecate SRP_VBASE_get_by_user.
2861 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2862 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2863 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2864 seed, even if the seed is configured.
2866 Users should use SRP_VBASE_get1_by_user instead. Note that in
2867 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2868 also that even though configuring the SRP seed attempts to hide
2869 invalid usernames by continuing the handshake with fake
2870 credentials, this behaviour is not constant time and no strong
2871 guarantees are made that the handshake is indistinguishable from
2872 that of a valid user.
2876 * Configuration change; it's now possible to build dynamic engines
2877 without having to build shared libraries and vice versa. This
2878 only applies to the engines in engines/, those in crypto/engine/
2879 will always be built into libcrypto (i.e. "static").
2881 Building dynamic engines is enabled by default; to disable, use
2882 the configuration option "disable-dynamic-engine".
2884 The only requirements for building dynamic engines are the
2885 presence of the DSO module and building with position independent
2886 code, so they will also automatically be disabled if configuring
2887 with "disable-dso" or "disable-pic".
2889 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2890 are also taken away from openssl/opensslconf.h, as they are
2895 * Configuration change; if there is a known flag to compile
2896 position independent code, it will always be applied on the
2897 libcrypto and libssl object files, and never on the application
2898 object files. This means other libraries that use routines from
2899 libcrypto / libssl can be made into shared libraries regardless
2900 of how OpenSSL was configured.
2902 If this isn't desirable, the configuration options "disable-pic"
2903 or "no-pic" can be used to disable the use of PIC. This will
2904 also disable building shared libraries and dynamic engines.
2908 * Removed JPAKE code. It was experimental and has no wide use.
2912 * The INSTALL_PREFIX Makefile variable has been renamed to
2913 DESTDIR. That makes for less confusion on what this variable
2914 is for. Also, the configuration option --install_prefix is
2919 * Heartbeat for TLS has been removed and is disabled by default
2920 for DTLS; configure with enable-heartbeats. Code that uses the
2921 old #define's might need to be updated.
2923 *Emilia Käsper, Rich Salz*
2925 * Rename REF_CHECK to REF_DEBUG.
2929 * New "unified" build system
2931 The "unified" build system is aimed to be a common system for all
2932 platforms we support. With it comes new support for VMS.
2934 This system builds supports building in a different directory tree
2935 than the source tree. It produces one Makefile (for unix family
2936 or lookalikes), or one descrip.mms (for VMS).
2938 The source of information to make the Makefile / descrip.mms is
2939 small files called 'build.info', holding the necessary
2940 information for each directory with source to compile, and a
2941 template in Configurations, like unix-Makefile.tmpl or
2944 With this change, the library names were also renamed on Windows
2945 and on VMS. They now have names that are closer to the standard
2946 on Unix, and include the major version number, and in certain
2947 cases, the architecture they are built for. See "Notes on shared
2948 libraries" in INSTALL.
2950 We rely heavily on the perl module Text::Template.
2954 * Added support for auto-initialisation and de-initialisation of the library.
2955 OpenSSL no longer requires explicit init or deinit routines to be called,
2956 except in certain circumstances. See the OPENSSL_init_crypto() and
2957 OPENSSL_init_ssl() man pages for further information.
2961 * The arguments to the DTLSv1_listen function have changed. Specifically the
2962 "peer" argument is now expected to be a BIO_ADDR object.
2964 * Rewrite of BIO networking library. The BIO library lacked consistent
2965 support of IPv6, and adding it required some more extensive
2966 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2967 which hold all types of addresses and chains of address information.
2968 It also introduces a new API, with functions like BIO_socket,
2969 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2970 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2971 have been adapted accordingly.
2975 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2980 * CRIME protection: disable compression by default, even if OpenSSL is
2981 compiled with zlib enabled. Applications can still enable compression
2982 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2983 using the SSL_CONF library to configure compression.
2987 * The signature of the session callback configured with
2988 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2989 was explicitly marked as `const unsigned char*` instead of
2994 * Always DPURIFY. Remove the use of uninitialized memory in the
2995 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2999 * Removed many obsolete configuration items, including
3000 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3001 MD2_CHAR, MD2_INT, MD2_LONG
3003 IDEA_SHORT, IDEA_LONG
3004 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3006 *Rich Salz, with advice from Andy Polyakov*
3008 * Many BN internals have been moved to an internal header file.
3010 *Rich Salz with help from Andy Polyakov*
3012 * Configuration and writing out the results from it has changed.
3013 Files such as Makefile include/openssl/opensslconf.h and are now
3014 produced through general templates, such as Makefile.in and
3015 crypto/opensslconf.h.in and some help from the perl module
3018 Also, the center of configuration information is no longer
3019 Makefile. Instead, Configure produces a perl module in
3020 configdata.pm which holds most of the config data (in the hash
3021 table %config), the target data that comes from the target
3022 configuration in one of the `Configurations/*.conf` files (in
3027 * To clarify their intended purposes, the Configure options
3028 --prefix and --openssldir change their semantics, and become more
3029 straightforward and less interdependent.
3031 --prefix shall be used exclusively to give the location INSTALLTOP
3032 where programs, scripts, libraries, include files and manuals are
3033 going to be installed. The default is now /usr/local.
3035 --openssldir shall be used exclusively to give the default
3036 location OPENSSLDIR where certificates, private keys, CRLs are
3037 managed. This is also where the default openssl.cnf gets
3039 If the directory given with this option is a relative path, the
3040 values of both the --prefix value and the --openssldir value will
3041 be combined to become OPENSSLDIR.
3042 The default for --openssldir is INSTALLTOP/ssl.
3044 Anyone who uses --openssldir to specify where OpenSSL is to be
3045 installed MUST change to use --prefix instead.
3049 * The GOST engine was out of date and therefore it has been removed. An up
3050 to date GOST engine is now being maintained in an external repository.
3051 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3052 support for GOST ciphersuites (these are only activated if a GOST engine
3057 * EGD is no longer supported by default; use enable-egd when
3060 *Ben Kaduk and Rich Salz*
3062 * The distribution now has Makefile.in files, which are used to
3063 create Makefile's when Configure is run. *Configure must be run
3064 before trying to build now.*
3068 * The return value for SSL_CIPHER_description() for error conditions
3073 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3075 Obtaining and performing DNSSEC validation of TLSA records is
3076 the application's responsibility. The application provides
3077 the TLSA records of its choice to OpenSSL, and these are then
3078 used to authenticate the peer.
3080 The TLSA records need not even come from DNS. They can, for
3081 example, be used to implement local end-entity certificate or
3082 trust-anchor "pinning", where the "pin" data takes the form
3083 of TLSA records, which can augment or replace verification
3084 based on the usual WebPKI public certification authorities.
3088 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3089 continues to support deprecated interfaces in default builds.
3090 However, applications are strongly advised to compile their
3091 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3092 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3093 or the 1.1.0 releases.
3095 In environments in which all applications have been ported to
3096 not use any deprecated interfaces OpenSSL's Configure script
3097 should be used with the --api=1.1.0 option to entirely remove
3098 support for the deprecated features from the library and
3099 unconditionally disable them in the installed headers.
3100 Essentially the same effect can be achieved with the "no-deprecated"
3101 argument to Configure, except that this will always restrict
3102 the build to just the latest API, rather than a fixed API
3105 As applications are ported to future revisions of the API,
3106 they should update their compile-time OPENSSL_API_COMPAT define
3107 accordingly, but in most cases should be able to continue to
3108 compile with later releases.
3110 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3111 0x10000000L and 0x00908000L, respectively. However those
3112 versions did not support the OPENSSL_API_COMPAT feature, and
3113 so applications are not typically tested for explicit support
3114 of just the undeprecated features of either release.
3118 * Add support for setting the minimum and maximum supported protocol.
3119 It can bet set via the SSL_set_min_proto_version() and
3120 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3121 MaxProtocol. It's recommended to use the new APIs to disable
3122 protocols instead of disabling individual protocols using
3123 SSL_set_options() or SSL_CONF's Protocol. This change also
3124 removes support for disabling TLS 1.2 in the OpenSSL TLS
3125 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3129 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3133 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3134 and integrates ECDSA and ECDH functionality into EC. Implementations can
3135 now redirect key generation and no longer need to convert to or from
3138 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3139 include the ec.h header file instead.
3143 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3144 ciphers who are no longer supported and drops support the ephemeral RSA key
3145 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3149 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3150 opaque. For HMAC_CTX, the following constructors and destructors
3153 HMAC_CTX *HMAC_CTX_new(void);
3154 void HMAC_CTX_free(HMAC_CTX *ctx);
3156 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3157 destroy such methods has been added. See EVP_MD_meth_new(3) and
3158 EVP_CIPHER_meth_new(3) for documentation.
3161 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3162 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3163 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3164 an already created structure.
3165 2) For consistency with the majority of our object creators and
3166 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3167 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3168 for deprecated builds.
3172 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3173 cryptographic operations to be performed asynchronously as long as an
3174 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3175 further details. Libssl has also had this capability integrated with the
3176 introduction of the new mode SSL_MODE_ASYNC and associated error
3177 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3178 pages. This work was developed in partnership with Intel Corp.
3182 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3183 always enabled now. If you want to disable the support you should
3184 exclude it using the list of supported ciphers. This also means that the
3185 "-no_ecdhe" option has been removed from s_server.
3189 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3190 SSL_{CTX_}set1_curves() which can set a list.
3194 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3195 curve you want to support using SSL_{CTX_}set1_curves().
3199 * State machine rewrite. The state machine code has been significantly
3200 refactored in order to remove much duplication of code and solve issues
3201 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3202 further details). This change does have some associated API changes.
3203 Notably the SSL_state() function has been removed and replaced by
3204 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3205 SSL_set_state() has been removed altogether. The previous handshake states
3206 defined in ssl.h and ssl3.h have also been removed.
3210 * All instances of the string "ssleay" in the public API were replaced
3211 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3212 Some error codes related to internal RSA_eay API's were renamed.
3216 * The demo files in crypto/threads were moved to demo/threads.
3220 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3223 *Matt Caswell, Rich Salz*
3225 * New ASN.1 embed macro.
3227 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3228 structure is not allocated: it is part of the parent. That is instead of
3236 This reduces memory fragmentation and make it impossible to accidentally
3237 set a mandatory field to NULL.
3239 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3240 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3241 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3246 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3250 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3251 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3252 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3253 DES and RC4 ciphersuites.
3257 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3258 This changes the decoding behaviour for some invalid messages,
3259 though the change is mostly in the more lenient direction, and
3260 legacy behaviour is preserved as much as possible.
3264 * Fix no-stdio build.
3265 *David Woodhouse <David.Woodhouse@intel.com> and also*
3266 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3268 * New testing framework
3269 The testing framework has been largely rewritten and is now using
3270 perl and the perl modules Test::Harness and an extended variant of
3271 Test::More called OpenSSL::Test to do its work. All test scripts in
3272 test/ have been rewritten into test recipes, and all direct calls to
3273 executables in test/Makefile have become individual recipes using the
3274 simplified testing OpenSSL::Test::Simple.
3276 For documentation on our testing modules, do:
3278 perldoc test/testlib/OpenSSL/Test/Simple.pm
3279 perldoc test/testlib/OpenSSL/Test.pm
3283 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3284 are used; the latter aborts on memory leaks (usually checked on exit).
3285 Some undocumented "set malloc, etc., hooks" functions were removed
3286 and others were changed. All are now documented.
3290 * In DSA_generate_parameters_ex, if the provided seed is too short,
3293 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3295 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3296 from RFC4279, RFC4785, RFC5487, RFC5489.
3298 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3299 original RSA_PSK patch.
3303 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3304 era flag was never set throughout the codebase (only read). Also removed
3305 SSL3_FLAGS_POP_BUFFER which was only used if
3306 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3310 * Changed the default name options in the "ca", "crl", "req" and "x509"
3311 to be "oneline" instead of "compat".
3315 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3316 not aware of clients that still exhibit this bug, and the workaround
3317 hasn't been working properly for a while.
3321 * The return type of BIO_number_read() and BIO_number_written() as well as
3322 the corresponding num_read and num_write members in the BIO structure has
3323 changed from unsigned long to uint64_t. On platforms where an unsigned
3324 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3329 * Given the pervasive nature of TLS extensions it is inadvisable to run
3330 OpenSSL without support for them. It also means that maintaining
3331 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3332 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3336 * Removed support for the two export grade static DH ciphersuites
3337 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3338 were newly added (along with a number of other static DH ciphersuites) to
3339 1.0.2. However the two export ones have *never* worked since they were
3340 introduced. It seems strange in any case to be adding new export
3341 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3345 * Version negotiation has been rewritten. In particular SSLv23_method(),
3346 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3347 and turned into macros which simply call the new preferred function names
3348 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3349 should use the new names instead. Also as part of this change the ssl23.h
3350 header file has been removed.
3354 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3355 code and the associated standard is no longer considered fit-for-purpose.
3359 * RT2547 was closed. When generating a private key, try to make the
3360 output file readable only by the owner. This behavior change might
3361 be noticeable when interacting with other software.
3363 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3368 * Added HTTP GET support to the ocsp command.
3372 * Changed default digest for the dgst and enc commands from MD5 to
3377 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3381 * Added support for TLS extended master secret from
3382 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3383 initial patch which was a great help during development.
3387 * All libssl internal structures have been removed from the public header
3388 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3389 now redundant). Users should not attempt to access internal structures
3390 directly. Instead they should use the provided API functions.
3394 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3395 Access to deprecated functions can be re-enabled by running config with
3396 "enable-deprecated". In addition applications wishing to use deprecated
3397 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3398 will, by default, disable some transitive includes that previously existed
3399 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3403 * Added support for OCB mode. OpenSSL has been granted a patent license
3404 compatible with the OpenSSL license for use of OCB. Details are available
3405 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3406 for OCB can be removed by calling config with no-ocb.
3410 * SSLv2 support has been removed. It still supports receiving a SSLv2
3411 compatible client hello.
3415 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3416 done while fixing the error code for the key-too-small case.
3418 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3420 * CA.sh has been removed; use CA.pl instead.
3424 * Removed old DES API.
3428 * Remove various unsupported platforms:
3434 Sinix/ReliantUNIX RM400
3439 16-bit platforms such as WIN16
3443 * Clean up OPENSSL_NO_xxx #define's
3444 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3445 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3446 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3447 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3448 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3449 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3450 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3451 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3452 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3453 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3457 * Cleaned up dead code
3458 Remove all but one '#ifdef undef' which is to be looked at.
3462 * Clean up calling of xxx_free routines.
3463 Just like free(), fix most of the xxx_free routines to accept
3464 NULL. Remove the non-null checks from callers. Save much code.
3468 * Add secure heap for storage of private keys (when possible).
3469 Add BIO_s_secmem(), CBIGNUM, etc.
3470 Contributed by Akamai Technologies under our Corporate CLA.
3474 * Experimental support for a new, fast, unbiased prime candidate generator,
3475 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3477 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3479 * New output format NSS in the sess_id command line tool. This allows
3480 exporting the session id and the master key in NSS keylog format.
3482 *Martin Kaiser <martin@kaiser.cx>*
3484 * Harmonize version and its documentation. -f flag is used to display
3487 *mancha <mancha1@zoho.com>*
3489 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3490 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3492 *mancha <mancha1@zoho.com>*
3494 * Fix some double frees. These are not thought to be exploitable.
3496 *mancha <mancha1@zoho.com>*
3498 * A missing bounds check in the handling of the TLS heartbeat extension
3499 can be used to reveal up to 64k of memory to a connected client or
3502 Thanks for Neel Mehta of Google Security for discovering this bug and to
3503 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3504 preparing the fix [CVE-2014-0160][]
3506 *Adam Langley, Bodo Moeller*
3508 * Fix for the attack described in the paper "Recovering OpenSSL
3509 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3510 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3511 <http://eprint.iacr.org/2014/140>
3513 Thanks to Yuval Yarom and Naomi Benger for discovering this
3514 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3516 *Yuval Yarom and Naomi Benger*
3518 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3519 this fixes a limitation in previous versions of OpenSSL.
3523 * Experimental encrypt-then-mac support.
3525 Experimental support for encrypt then mac from
3526 draft-gutmann-tls-encrypt-then-mac-02.txt
3528 To enable it set the appropriate extension number (0x42 for the test
3529 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3531 For non-compliant peers (i.e. just about everything) this should have no
3534 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3538 * Add EVP support for key wrapping algorithms, to avoid problems with
3539 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3540 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3541 algorithms and include tests cases.
3545 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3550 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3551 MGF1 digest and OAEP label.
3555 * Make openssl verify return errors.
3557 *Chris Palmer <palmer@google.com> and Ben Laurie*
3559 * New function ASN1_TIME_diff to calculate the difference between two
3560 ASN1_TIME structures or one structure and the current time.
3564 * Update fips_test_suite to support multiple command line options. New
3565 test to induce all self test errors in sequence and check expected
3570 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3571 sign or verify all in one operation.
3575 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3576 test programs and fips_test_suite. Includes functionality to parse
3577 the minimal script output of fipsalgest.pl directly.
3581 * Add authorisation parameter to FIPS_module_mode_set().
3585 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3589 * Use separate DRBG fields for internal and external flags. New function
3590 FIPS_drbg_health_check() to perform on demand health checking. Add
3591 generation tests to fips_test_suite with reduced health check interval to
3592 demonstrate periodic health checking. Add "nodh" option to
3593 fips_test_suite to skip very slow DH test.
3597 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3602 * More extensive health check for DRBG checking many more failure modes.
3603 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3604 combination: call this in fips_test_suite.
3608 * Add support for canonical generation of DSA parameter 'g'. See
3611 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3612 POST to handle HMAC cases.
3616 * Add functions FIPS_module_version() and FIPS_module_version_text()
3617 to return numerical and string versions of the FIPS module number.
3621 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3622 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3623 outside the validated module in the FIPS capable OpenSSL.
3627 * Minor change to DRBG entropy callback semantics. In some cases
3628 there is no multiple of the block length between min_len and
3629 max_len. Allow the callback to return more than max_len bytes
3630 of entropy but discard any extra: it is the callback's responsibility
3631 to ensure that the extra data discarded does not impact the
3632 requested amount of entropy.
3636 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3637 information in FIPS186-3, SP800-57 and SP800-131A.
3641 * CCM support via EVP. Interface is very similar to GCM case except we
3642 must supply all data in one chunk (i.e. no update, final) and the
3643 message length must be supplied if AAD is used. Add algorithm test
3648 * Initial version of POST overhaul. Add POST callback to allow the status
3649 of POST to be monitored and/or failures induced. Modify fips_test_suite
3650 to use callback. Always run all selftests even if one fails.
3654 * XTS support including algorithm test driver in the fips_gcmtest program.
3655 Note: this does increase the maximum key length from 32 to 64 bytes but
3656 there should be no binary compatibility issues as existing applications
3657 will never use XTS mode.
3661 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3662 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3663 performs algorithm blocking for unapproved PRNG types. Also do not
3664 set PRNG type in FIPS_mode_set(): leave this to the application.
3665 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3666 the standard OpenSSL PRNG: set additional data to a date time vector.
3670 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3671 This shouldn't present any incompatibility problems because applications
3672 shouldn't be using these directly and any that are will need to rethink
3673 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3677 * Extensive self tests and health checking required by SP800-90 DRBG.
3678 Remove strength parameter from FIPS_drbg_instantiate and always
3679 instantiate at maximum supported strength.
3683 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3687 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3691 * New function DH_compute_key_padded() to compute a DH key and pad with
3692 leading zeroes if needed: this complies with SP800-56A et al.
3696 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3697 anything, incomplete, subject to change and largely untested at present.
3701 * Modify fipscanisteronly build option to only build the necessary object
3702 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3706 * Add experimental option FIPSSYMS to give all symbols in
3707 fipscanister.o and FIPS or fips prefix. This will avoid
3708 conflicts with future versions of OpenSSL. Add perl script
3709 util/fipsas.pl to preprocess assembly language source files
3710 and rename any affected symbols.
3714 * Add selftest checks and algorithm block of non-fips algorithms in
3715 FIPS mode. Remove DES2 from selftests.
3719 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3720 return internal method without any ENGINE dependencies. Add new
3721 tiny fips sign and verify functions.
3725 * New build option no-ec2m to disable characteristic 2 code.
3729 * New build option "fipscanisteronly". This only builds fipscanister.o
3730 and (currently) associated fips utilities. Uses the file Makefile.fips
3731 instead of Makefile.org as the prototype.
3735 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3736 Update fips_gcmtest to use IV generator.
3740 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3741 setting output buffer to NULL. The `*Final` function must be
3742 called although it will not retrieve any additional data. The tag
3743 can be set or retrieved with a ctrl. The IV length is by default 12
3744 bytes (96 bits) but can be set to an alternative value. If the IV
3745 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3750 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3751 underlying do_cipher function handles all cipher semantics itself
3752 including padding and finalisation. This is useful if (for example)
3753 an ENGINE cipher handles block padding itself. The behaviour of
3754 do_cipher is subtly changed if this flag is set: the return value
3755 is the number of characters written to the output buffer (zero is
3756 no longer an error code) or a negative error code. Also if the
3757 input buffer is NULL and length 0 finalisation should be performed.
3761 * If a candidate issuer certificate is already part of the constructed
3762 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3766 * Improve forward-security support: add functions
3768 void SSL_CTX_set_not_resumable_session_callback(
3769 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3770 void SSL_set_not_resumable_session_callback(
3771 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3773 for use by SSL/TLS servers; the callback function will be called whenever a
3774 new session is created, and gets to decide whether the session may be
3775 cached to make it resumable (return 0) or not (return 1). (As by the
3776 SSL/TLS protocol specifications, the session_id sent by the server will be
3777 empty to indicate that the session is not resumable; also, the server will
3778 not generate RFC 4507 (RFC 5077) session tickets.)
3780 A simple reasonable callback implementation is to return is_forward_secure.
3781 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3782 by the SSL/TLS server library, indicating whether it can provide forward
3785 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3787 * New -verify_name option in command line utilities to set verification
3792 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3793 Add CMAC pkey methods.
3797 * Experimental renegotiation in s_server -www mode. If the client
3798 browses /reneg connection is renegotiated. If /renegcert it is
3799 renegotiated requesting a certificate.
3803 * Add an "external" session cache for debugging purposes to s_server. This
3804 should help trace issues which normally are only apparent in deployed
3805 multi-process servers.
3809 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3810 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3811 BIO_set_cipher() and some obscure PEM functions were changed so they
3812 can now return an error. The RAND changes required a change to the
3813 RAND_METHOD structure.
3817 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3818 a gcc attribute to warn if the result of a function is ignored. This
3819 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3820 whose return value is often ignored.
3824 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3825 These allow SCTs (signed certificate timestamps) to be requested and
3826 validated when establishing a connection.
3828 *Rob Percival <robpercival@google.com>*
3833 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3835 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3836 used even when parsing explicit parameters, when loading a encoded key
3837 or calling `EC_GROUP_new_from_ecpkparameters()`/
3838 `EC_GROUP_new_from_ecparameters()`.
3839 This prevents bypass of security hardening and performance gains,
3840 especially for curves with specialized EC_METHODs.
3841 By default, if a key encoded with explicit parameters is loaded and later
3842 encoded, the output is still encoded with explicit parameters, even if
3843 internally a "named" EC_GROUP is used for computation.
3847 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3848 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3849 NULL. After this change, only the cofactor parameter can be NULL. It also
3850 does some minimal sanity checks on the passed order.
3855 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3856 An attack is simple, if the first CMS_recipientInfo is valid but the
3857 second CMS_recipientInfo is chosen ciphertext. If the second
3858 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3859 encryption key will be replaced by garbage, and the message cannot be
3860 decoded, but if the RSA decryption fails, the correct encryption key is
3861 used and the recipient will not notice the attack.
3862 As a work around for this potential attack the length of the decrypted
3863 key must be equal to the cipher default key length, in case the
3864 certifiate is not given and all recipientInfo are tried out.
3865 The old behaviour can be re-enabled in the CMS code by setting the
3866 CMS_DEBUG_DECRYPT flag.
3871 * Document issue with installation paths in diverse Windows builds
3873 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3874 binaries and run-time config file.
3879 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3881 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3882 This changes the size when using the genpkey app when no size is given. It
3883 fixes an omission in earlier changes that changed all RSA, DSA and DH
3884 generation apps to use 2048 bits by default.
3888 * Add FIPS support for Android Arm 64-bit
3890 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3891 Module in Version 2.0.10. For some reason, the corresponding target
3892 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3893 built with FIPS support on Android Arm 64-bit. This omission has been
3896 *Matthias St. Pierre*
3898 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3900 * 0-byte record padding oracle
3902 If an application encounters a fatal protocol error and then calls
3903 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3904 then OpenSSL can respond differently to the calling application if a 0 byte
3905 record is received with invalid padding compared to if a 0 byte record is
3906 received with an invalid MAC. If the application then behaves differently
3907 based on that in a way that is detectable to the remote peer, then this
3908 amounts to a padding oracle that could be used to decrypt data.
3910 In order for this to be exploitable "non-stitched" ciphersuites must be in
3911 use. Stitched ciphersuites are optimised implementations of certain
3912 commonly used ciphersuites. Also the application must call SSL_shutdown()
3913 twice even if a protocol error has occurred (applications should not do
3914 this but some do anyway).
3916 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3917 Aviram, with additional investigation by Steven Collison and Andrew
3918 Hourselt. It was reported to OpenSSL on 10th December 2018.
3923 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3927 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3929 * Microarchitecture timing vulnerability in ECC scalar multiplication
3931 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3932 shown to be vulnerable to a microarchitecture timing side channel attack.
3933 An attacker with sufficient access to mount local timing attacks during
3934 ECDSA signature generation could recover the private key.
3936 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3937 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3943 * Timing vulnerability in DSA signature generation
3945 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3946 timing side channel attack. An attacker could use variations in the signing
3947 algorithm to recover the private key.
3949 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3954 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3955 Module, accidentally introduced while backporting security fixes from the
3956 development branch and hindering the use of ECC in FIPS mode.
3960 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3962 * Client DoS due to large DH parameter
3964 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3965 malicious server can send a very large prime value to the client. This will
3966 cause the client to spend an unreasonably long period of time generating a
3967 key for this prime resulting in a hang until the client has finished. This
3968 could be exploited in a Denial Of Service attack.
3970 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3975 * Cache timing vulnerability in RSA Key Generation
3977 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3978 a cache timing side channel attack. An attacker with sufficient access to
3979 mount cache timing attacks during the RSA key generation process could
3980 recover the private key.
3982 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3983 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3988 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3989 parameter is no longer accepted, as it leads to a corrupt table. NULL
3990 pem_str is reserved for alias entries only.
3994 * Revert blinding in ECDSA sign and instead make problematic addition
3995 length-invariant. Switch even to fixed-length Montgomery multiplication.
3999 * Change generating and checking of primes so that the error rate of not
4000 being prime depends on the intended use based on the size of the input.
4001 For larger primes this will result in more rounds of Miller-Rabin.
4002 The maximal error rate for primes with more than 1080 bits is lowered
4005 *Kurt Roeckx, Annie Yousar*
4007 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4011 * Add blinding to ECDSA and DSA signatures to protect against side channel
4012 attacks discovered by Keegan Ryan (NCC Group).
4016 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4017 now allow empty (zero character) pass phrases.
4021 * Certificate time validation (X509_cmp_time) enforces stricter
4022 compliance with RFC 5280. Fractional seconds and timezone offsets
4023 are no longer allowed.
4027 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4029 * Constructed ASN.1 types with a recursive definition could exceed the stack
4031 Constructed ASN.1 types with a recursive definition (such as can be found
4032 in PKCS7) could eventually exceed the stack given malicious input with
4033 excessive recursion. This could result in a Denial Of Service attack. There
4034 are no such structures used within SSL/TLS that come from untrusted sources
4035 so this is considered safe.
4037 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4043 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4045 * Read/write after SSL object in error state
4047 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4048 mechanism. The intent was that if a fatal error occurred during a handshake
4049 then OpenSSL would move into the error state and would immediately fail if
4050 you attempted to continue the handshake. This works as designed for the
4051 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4052 SSL_connect()), however due to a bug it does not work correctly if
4053 SSL_read() or SSL_write() is called directly. In that scenario, if the
4054 handshake fails then a fatal error will be returned in the initial function
4055 call. If SSL_read()/SSL_write() is subsequently called by the application
4056 for the same SSL object then it will succeed and the data is passed without
4057 being decrypted/encrypted directly from the SSL/TLS record layer.
4059 In order to exploit this issue an application bug would have to be present
4060 that resulted in a call to SSL_read()/SSL_write() being issued after having
4061 already received a fatal error.
4063 This issue was reported to OpenSSL by David Benjamin (Google).
4068 * rsaz_1024_mul_avx2 overflow bug on x86_64
4070 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4071 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4072 Analysis suggests that attacks against RSA and DSA as a result of this
4073 defect would be very difficult to perform and are not believed likely.
4074 Attacks against DH1024 are considered just feasible, because most of the
4075 work necessary to deduce information about a private key may be performed
4076 offline. The amount of resources required for such an attack would be
4077 significant. However, for an attack on TLS to be meaningful, the server
4078 would have to share the DH1024 private key among multiple clients, which is
4079 no longer an option since CVE-2016-0701.
4081 This only affects processors that support the AVX2 but not ADX extensions
4082 like Intel Haswell (4th generation).
4084 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4085 was originally found via the OSS-Fuzz project.
4090 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4092 * bn_sqrx8x_internal carry bug on x86_64
4094 There is a carry propagating bug in the x86_64 Montgomery squaring
4095 procedure. No EC algorithms are affected. Analysis suggests that attacks
4096 against RSA and DSA as a result of this defect would be very difficult to
4097 perform and are not believed likely. Attacks against DH are considered just
4098 feasible (although very difficult) because most of the work necessary to
4099 deduce information about a private key may be performed offline. The amount
4100 of resources required for such an attack would be very significant and
4101 likely only accessible to a limited number of attackers. An attacker would
4102 additionally need online access to an unpatched system using the target
4103 private key in a scenario with persistent DH parameters and a private
4104 key that is shared between multiple clients.
4106 This only affects processors that support the BMI1, BMI2 and ADX extensions
4107 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4109 This issue was reported to OpenSSL by the OSS-Fuzz project.
4114 * Malformed X.509 IPAddressFamily could cause OOB read
4116 If an X.509 certificate has a malformed IPAddressFamily extension,
4117 OpenSSL could do a one-byte buffer overread. The most likely result
4118 would be an erroneous display of the certificate in text format.
4120 This issue was reported to OpenSSL by the OSS-Fuzz project.
4125 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4127 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4128 platform rather than 'mingw'.
4132 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4134 * Truncated packet could crash via OOB read
4136 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4137 cipher is being used, then a truncated packet can cause that host to
4138 perform an out-of-bounds read, usually resulting in a crash.
4140 This issue was reported to OpenSSL by Robert Święcki of Google.
4145 * BN_mod_exp may produce incorrect results on x86_64
4147 There is a carry propagating bug in the x86_64 Montgomery squaring
4148 procedure. No EC algorithms are affected. Analysis suggests that attacks
4149 against RSA and DSA as a result of this defect would be very difficult to
4150 perform and are not believed likely. Attacks against DH are considered just
4151 feasible (although very difficult) because most of the work necessary to
4152 deduce information about a private key may be performed offline. The amount
4153 of resources required for such an attack would be very significant and
4154 likely only accessible to a limited number of attackers. An attacker would
4155 additionally need online access to an unpatched system using the target
4156 private key in a scenario with persistent DH parameters and a private
4157 key that is shared between multiple clients. For example this can occur by
4158 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4159 similar to CVE-2015-3193 but must be treated as a separate problem.
4161 This issue was reported to OpenSSL by the OSS-Fuzz project.
4166 * Montgomery multiplication may produce incorrect results
4168 There is a carry propagating bug in the Broadwell-specific Montgomery
4169 multiplication procedure that handles input lengths divisible by, but
4170 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4171 and DH private keys are impossible. This is because the subroutine in
4172 question is not used in operations with the private key itself and an input
4173 of the attacker's direct choice. Otherwise the bug can manifest itself as
4174 transient authentication and key negotiation failures or reproducible
4175 erroneous outcome of public-key operations with specially crafted input.
4176 Among EC algorithms only Brainpool P-512 curves are affected and one
4177 presumably can attack ECDH key negotiation. Impact was not analyzed in
4178 detail, because pre-requisites for attack are considered unlikely. Namely
4179 multiple clients have to choose the curve in question and the server has to
4180 share the private key among them, neither of which is default behaviour.
4181 Even then only clients that chose the curve will be affected.
4183 This issue was publicly reported as transient failures and was not
4184 initially recognized as a security issue. Thanks to Richard Morgan for
4185 providing reproducible case.
4190 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4191 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4192 prevent issues where no progress is being made and the peer continually
4193 sends unrecognised record types, using up resources processing them.
4197 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4199 * Missing CRL sanity check
4201 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4202 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4203 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4205 This issue only affects the OpenSSL 1.0.2i
4210 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4212 * OCSP Status Request extension unbounded memory growth
4214 A malicious client can send an excessively large OCSP Status Request
4215 extension. If that client continually requests renegotiation, sending a
4216 large OCSP Status Request extension each time, then there will be unbounded
4217 memory growth on the server. This will eventually lead to a Denial Of
4218 Service attack through memory exhaustion. Servers with a default
4219 configuration are vulnerable even if they do not support OCSP. Builds using
4220 the "no-ocsp" build time option are not affected.
4222 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4227 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4230 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4236 * OOB write in MDC2_Update()
4238 An overflow can occur in MDC2_Update() either if called directly or
4239 through the EVP_DigestUpdate() function using MDC2. If an attacker
4240 is able to supply very large amounts of input data after a previous
4241 call to EVP_EncryptUpdate() with a partial block then a length check
4242 can overflow resulting in a heap corruption.
4244 The amount of data needed is comparable to SIZE_MAX which is impractical
4247 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4252 * Malformed SHA512 ticket DoS
4254 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4255 DoS attack where a malformed ticket will result in an OOB read which will
4258 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4259 a custom server callback and ticket lookup mechanism.
4261 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4266 * OOB write in BN_bn2dec()
4268 The function BN_bn2dec() does not check the return value of BN_div_word().
4269 This can cause an OOB write if an application uses this function with an
4270 overly large BIGNUM. This could be a problem if an overly large certificate
4271 or CRL is printed out from an untrusted source. TLS is not affected because
4272 record limits will reject an oversized certificate before it is parsed.
4274 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4279 * OOB read in TS_OBJ_print_bio()
4281 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4282 the total length the OID text representation would use and not the amount
4283 of data written. This will result in OOB reads when large OIDs are
4286 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4291 * Pointer arithmetic undefined behaviour
4293 Avoid some undefined pointer arithmetic
4295 A common idiom in the codebase is to check limits in the following manner:
4298 Where "p" points to some malloc'd data of SIZE bytes and
4301 "len" here could be from some externally supplied data (e.g. from a TLS
4304 The rules of C pointer arithmetic are such that "p + len" is only well
4305 defined where len <= SIZE. Therefore the above idiom is actually
4306 undefined behaviour.
4308 For example this could cause problems if some malloc implementation
4309 provides an address for "p" such that "p + len" actually overflows for
4310 values of len that are too big and therefore p + len < limit.
4312 This issue was reported to OpenSSL by Guido Vranken
4317 * Constant time flag not preserved in DSA signing
4319 Operations in the DSA signing algorithm should run in constant time in
4320 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4321 implementation means that a non-constant time codepath is followed for
4322 certain operations. This has been demonstrated through a cache-timing
4323 attack to be sufficient for an attacker to recover the private DSA key.
4325 This issue was reported by César Pereida (Aalto University), Billy Brumley
4326 (Tampere University of Technology), and Yuval Yarom (The University of
4327 Adelaide and NICTA).
4332 * DTLS buffered message DoS
4334 In a DTLS connection where handshake messages are delivered out-of-order
4335 those messages that OpenSSL is not yet ready to process will be buffered
4336 for later use. Under certain circumstances, a flaw in the logic means that
4337 those messages do not get removed from the buffer even though the handshake
4338 has been completed. An attacker could force up to approx. 15 messages to
4339 remain in the buffer when they are no longer required. These messages will
4340 be cleared when the DTLS connection is closed. The default maximum size for
4341 a message is 100k. Therefore the attacker could force an additional 1500k
4342 to be consumed per connection. By opening many simulataneous connections an
4343 attacker could cause a DoS attack through memory exhaustion.
4345 This issue was reported to OpenSSL by Quan Luo.
4350 * DTLS replay protection DoS
4352 A flaw in the DTLS replay attack protection mechanism means that records
4353 that arrive for future epochs update the replay protection "window" before
4354 the MAC for the record has been validated. This could be exploited by an
4355 attacker by sending a record for the next epoch (which does not have to
4356 decrypt or have a valid MAC), with a very large sequence number. This means
4357 that all subsequent legitimate packets are dropped causing a denial of
4358 service for a specific DTLS connection.
4360 This issue was reported to OpenSSL by the OCAP audit team.
4365 * Certificate message OOB reads
4367 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4368 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4369 theoretical DoS risk but this has not been observed in practice on common
4372 The messages affected are client certificate, client certificate request
4373 and server certificate. As a result the attack can only be performed
4374 against a client or a server which enables client authentication.
4376 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4381 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4383 * Prevent padding oracle in AES-NI CBC MAC check
4385 A MITM attacker can use a padding oracle attack to decrypt traffic
4386 when the connection uses an AES CBC cipher and the server support
4389 This issue was introduced as part of the fix for Lucky 13 padding
4390 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4391 constant time by making sure that always the same bytes are read and
4392 compared against either the MAC or padding bytes. But it no longer
4393 checked that there was enough data to have both the MAC and padding
4396 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4401 * Fix EVP_EncodeUpdate overflow
4403 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4404 Base64 encoding of binary data. If an attacker is able to supply very large
4405 amounts of input data then a length check can overflow resulting in a heap
4408 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4409 the `PEM_write_bio*` family of functions. These are mainly used within the
4410 OpenSSL command line applications, so any application which processes data
4411 from an untrusted source and outputs it as a PEM file should be considered
4412 vulnerable to this issue. User applications that call these APIs directly
4413 with large amounts of untrusted data may also be vulnerable.
4415 This issue was reported by Guido Vranken.
4420 * Fix EVP_EncryptUpdate overflow
4422 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4423 is able to supply very large amounts of input data after a previous call to
4424 EVP_EncryptUpdate() with a partial block then a length check can overflow
4425 resulting in a heap corruption. Following an analysis of all OpenSSL
4426 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4427 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4428 the first called function after an EVP_EncryptInit(), and therefore that
4429 specific call must be safe. The second form is where the length passed to
4430 EVP_EncryptUpdate() can be seen from the code to be some small value and
4431 therefore there is no possibility of an overflow. Since all instances are
4432 one of these two forms, it is believed that there can be no overflows in
4433 internal code due to this problem. It should be noted that
4434 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4435 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4436 of these calls have also been analysed too and it is believed there are no
4437 instances in internal usage where an overflow could occur.
4439 This issue was reported by Guido Vranken.
4444 * Prevent ASN.1 BIO excessive memory allocation
4446 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4447 a short invalid encoding can cause allocation of large amounts of memory
4448 potentially consuming excessive resources or exhausting memory.
4450 Any application parsing untrusted data through d2i BIO functions is
4451 affected. The memory based functions such as d2i_X509() are *not* affected.
4452 Since the memory based functions are used by the TLS library, TLS
4453 applications are not affected.
4455 This issue was reported by Brian Carpenter.
4462 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4463 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4464 in arbitrary stack data being returned in the buffer.
4466 This issue was reported by Guido Vranken.
4471 * Modify behavior of ALPN to invoke callback after SNI/servername
4472 callback, such that updates to the SSL_CTX affect ALPN.
4476 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4481 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4482 methods are enabled and ssl2 is disabled the methods return NULL.
4486 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4488 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4489 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4490 provide any "EXPORT" or "LOW" strength ciphers.
4494 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4495 is by default disabled at build-time. Builds that are not configured with
4496 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4497 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4498 will need to explicitly call either of:
4500 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4502 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4504 as appropriate. Even if either of those is used, or the application
4505 explicitly uses the version-specific SSLv2_method() or its client and
4506 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4507 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4508 ciphers, and SSLv2 56-bit DES are no longer available.
4513 * Fix a double-free in DSA code
4515 A double free bug was discovered when OpenSSL parses malformed DSA private
4516 keys and could lead to a DoS attack or memory corruption for applications
4517 that receive DSA private keys from untrusted sources. This scenario is
4520 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4526 * Disable SRP fake user seed to address a server memory leak.
4528 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4530 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4531 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4532 was changed to ignore the "fake user" SRP seed, even if the seed
4535 Users should use SRP_VBASE_get1_by_user instead. Note that in
4536 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4537 also that even though configuring the SRP seed attempts to hide
4538 invalid usernames by continuing the handshake with fake
4539 credentials, this behaviour is not constant time and no strong
4540 guarantees are made that the handshake is indistinguishable from
4541 that of a valid user.
4546 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4548 In the BN_hex2bn function the number of hex digits is calculated using an
4549 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4550 large values of `i` this can result in `bn_expand` not allocating any
4551 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4552 field as NULL leading to a subsequent NULL ptr deref. For very large values
4553 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4554 In this case memory is allocated to the internal BIGNUM data field, but it
4555 is insufficiently sized leading to heap corruption. A similar issue exists
4556 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4557 is ever called by user applications with very large untrusted hex/dec data.
4558 This is anticipated to be a rare occurrence.
4560 All OpenSSL internal usage of these functions use data that is not expected
4561 to be untrusted, e.g. config file data or application command line
4562 arguments. If user developed applications generate config file data based
4563 on untrusted data then it is possible that this could also lead to security
4564 consequences. This is also anticipated to be rare.
4566 This issue was reported to OpenSSL by Guido Vranken.
4571 * Fix memory issues in `BIO_*printf` functions
4573 The internal `fmtstr` function used in processing a "%s" format string in
4574 the `BIO_*printf` functions could overflow while calculating the length of a
4575 string and cause an OOB read when printing very long strings.
4577 Additionally the internal `doapr_outch` function can attempt to write to an
4578 OOB memory location (at an offset from the NULL pointer) in the event of a
4579 memory allocation failure. In 1.0.2 and below this could be caused where
4580 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4581 could be in processing a very long "%s" format string. Memory leaks can
4584 The first issue may mask the second issue dependent on compiler behaviour.
4585 These problems could enable attacks where large amounts of untrusted data
4586 is passed to the `BIO_*printf` functions. If applications use these functions
4587 in this way then they could be vulnerable. OpenSSL itself uses these
4588 functions when printing out human-readable dumps of ASN.1 data. Therefore
4589 applications that print this data could be vulnerable if the data is from
4590 untrusted sources. OpenSSL command line applications could also be
4591 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4592 as command line arguments.
4594 Libssl is not considered directly vulnerable. Additionally certificates etc
4595 received via remote connections via libssl are also unlikely to be able to
4596 trigger these issues because of message size limits enforced within libssl.
4598 This issue was reported to OpenSSL Guido Vranken.
4603 * Side channel attack on modular exponentiation
4605 A side-channel attack was found which makes use of cache-bank conflicts on
4606 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4607 of RSA keys. The ability to exploit this issue is limited as it relies on
4608 an attacker who has control of code in a thread running on the same
4609 hyper-threaded core as the victim thread which is performing decryptions.
4611 This issue was reported to OpenSSL by Yuval Yarom, The University of
4612 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4613 Nadia Heninger, University of Pennsylvania with more information at
4614 <http://cachebleed.info>.
4619 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4620 if no keysize is specified with default_bits. This fixes an
4621 omission in an earlier change that changed all RSA/DSA key generation
4622 apps to use 2048 bits by default.
4626 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4628 * DH small subgroups
4630 Historically OpenSSL only ever generated DH parameters based on "safe"
4631 primes. More recently (in version 1.0.2) support was provided for
4632 generating X9.42 style parameter files such as those required for RFC 5114
4633 support. The primes used in such files may not be "safe". Where an
4634 application is using DH configured with parameters based on primes that are
4635 not "safe" then an attacker could use this fact to find a peer's private
4636 DH exponent. This attack requires that the attacker complete multiple
4637 handshakes in which the peer uses the same private DH exponent. For example
4638 this could be used to discover a TLS server's private DH exponent if it's
4639 reusing the private DH exponent or it's using a static DH ciphersuite.
4641 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4642 TLS. It is not on by default. If the option is not set then the server
4643 reuses the same private DH exponent for the life of the server process and
4644 would be vulnerable to this attack. It is believed that many popular
4645 applications do set this option and would therefore not be at risk.
4647 The fix for this issue adds an additional check where a "q" parameter is
4648 available (as is the case in X9.42 based parameters). This detects the
4649 only known attack, and is the only possible defense for static DH
4650 ciphersuites. This could have some performance impact.
4652 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4653 default and cannot be disabled. This could have some performance impact.
4655 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4660 * SSLv2 doesn't block disabled ciphers
4662 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4663 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4664 been disabled, provided that the SSLv2 protocol was not also disabled via
4667 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4668 and Sebastian Schinzel.
4673 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4675 * BN_mod_exp may produce incorrect results on x86_64
4677 There is a carry propagating bug in the x86_64 Montgomery squaring
4678 procedure. No EC algorithms are affected. Analysis suggests that attacks
4679 against RSA and DSA as a result of this defect would be very difficult to
4680 perform and are not believed likely. Attacks against DH are considered just
4681 feasible (although very difficult) because most of the work necessary to
4682 deduce information about a private key may be performed offline. The amount
4683 of resources required for such an attack would be very significant and
4684 likely only accessible to a limited number of attackers. An attacker would
4685 additionally need online access to an unpatched system using the target
4686 private key in a scenario with persistent DH parameters and a private
4687 key that is shared between multiple clients. For example this can occur by
4688 default in OpenSSL DHE based SSL/TLS ciphersuites.
4690 This issue was reported to OpenSSL by Hanno Böck.
4695 * Certificate verify crash with missing PSS parameter
4697 The signature verification routines will crash with a NULL pointer
4698 dereference if presented with an ASN.1 signature using the RSA PSS
4699 algorithm and absent mask generation function parameter. Since these
4700 routines are used to verify certificate signature algorithms this can be
4701 used to crash any certificate verification operation and exploited in a
4702 DoS attack. Any application which performs certificate verification is
4703 vulnerable including OpenSSL clients and servers which enable client
4706 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4711 * X509_ATTRIBUTE memory leak
4713 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4714 memory. This structure is used by the PKCS#7 and CMS routines so any
4715 application which reads PKCS#7 or CMS data from untrusted sources is
4716 affected. SSL/TLS is not affected.
4718 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4724 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4725 This changes the decoding behaviour for some invalid messages,
4726 though the change is mostly in the more lenient direction, and
4727 legacy behaviour is preserved as much as possible.
4731 * In DSA_generate_parameters_ex, if the provided seed is too short,
4734 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4736 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4738 * Alternate chains certificate forgery
4740 During certificate verification, OpenSSL will attempt to find an
4741 alternative certificate chain if the first attempt to build such a chain
4742 fails. An error in the implementation of this logic can mean that an
4743 attacker could cause certain checks on untrusted certificates to be
4744 bypassed, such as the CA flag, enabling them to use a valid leaf
4745 certificate to act as a CA and "issue" an invalid certificate.
4747 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4752 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4754 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4755 incompatibility in the handling of HMAC. The previous ABI has now been
4760 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4762 * Malformed ECParameters causes infinite loop
4764 When processing an ECParameters structure OpenSSL enters an infinite loop
4765 if the curve specified is over a specially malformed binary polynomial
4768 This can be used to perform denial of service against any
4769 system which processes public keys, certificate requests or
4770 certificates. This includes TLS clients and TLS servers with
4771 client authentication enabled.
4773 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4778 * Exploitable out-of-bounds read in X509_cmp_time
4780 X509_cmp_time does not properly check the length of the ASN1_TIME
4781 string and can read a few bytes out of bounds. In addition,
4782 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4785 An attacker can use this to craft malformed certificates and CRLs of
4786 various sizes and potentially cause a segmentation fault, resulting in
4787 a DoS on applications that verify certificates or CRLs. TLS clients
4788 that verify CRLs are affected. TLS clients and servers with client
4789 authentication enabled may be affected if they use custom verification
4792 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4793 independently by Hanno Böck.
4798 * PKCS7 crash with missing EnvelopedContent
4800 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4801 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4802 with missing content and trigger a NULL pointer dereference on parsing.
4804 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4805 structures from untrusted sources are affected. OpenSSL clients and
4806 servers are not affected.
4808 This issue was reported to OpenSSL by Michal Zalewski (Google).
4813 * CMS verify infinite loop with unknown hash function
4815 When verifying a signedData message the CMS code can enter an infinite loop
4816 if presented with an unknown hash function OID. This can be used to perform
4817 denial of service against any system which verifies signedData messages using
4819 This issue was reported to OpenSSL by Johannes Bauer.
4824 * Race condition handling NewSessionTicket
4826 If a NewSessionTicket is received by a multi-threaded client when attempting to
4827 reuse a previous ticket then a race condition can occur potentially leading to
4828 a double free of the ticket data.
4833 * Only support 256-bit or stronger elliptic curves with the
4834 'ecdh_auto' setting (server) or by default (client). Of supported
4835 curves, prefer P-256 (both).
4839 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4841 * ClientHello sigalgs DoS fix
4843 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4844 invalid signature algorithms extension a NULL pointer dereference will
4845 occur. This can be exploited in a DoS attack against the server.
4847 This issue was was reported to OpenSSL by David Ramos of Stanford
4851 *Stephen Henson and Matt Caswell*
4853 * Multiblock corrupted pointer fix
4855 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4856 feature only applies on 64 bit x86 architecture platforms that support AES
4857 NI instructions. A defect in the implementation of "multiblock" can cause
4858 OpenSSL's internal write buffer to become incorrectly set to NULL when
4859 using non-blocking IO. Typically, when the user application is using a
4860 socket BIO for writing, this will only result in a failed connection.
4861 However if some other BIO is used then it is likely that a segmentation
4862 fault will be triggered, thus enabling a potential DoS attack.
4864 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4869 * Segmentation fault in DTLSv1_listen fix
4871 The DTLSv1_listen function is intended to be stateless and processes the
4872 initial ClientHello from many peers. It is common for user code to loop
4873 over the call to DTLSv1_listen until a valid ClientHello is received with
4874 an associated cookie. A defect in the implementation of DTLSv1_listen means
4875 that state is preserved in the SSL object from one invocation to the next
4876 that can lead to a segmentation fault. Errors processing the initial
4877 ClientHello can trigger this scenario. An example of such an error could be
4878 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4881 This issue was reported to OpenSSL by Per Allansson.
4886 * Segmentation fault in ASN1_TYPE_cmp fix
4888 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4889 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4890 certificate signature algorithm consistency this can be used to crash any
4891 certificate verification operation and exploited in a DoS attack. Any
4892 application which performs certificate verification is vulnerable including
4893 OpenSSL clients and servers which enable client authentication.
4898 * Segmentation fault for invalid PSS parameters fix
4900 The signature verification routines will crash with a NULL pointer
4901 dereference if presented with an ASN.1 signature using the RSA PSS
4902 algorithm and invalid parameters. Since these routines are used to verify
4903 certificate signature algorithms this can be used to crash any
4904 certificate verification operation and exploited in a DoS attack. Any
4905 application which performs certificate verification is vulnerable including
4906 OpenSSL clients and servers which enable client authentication.
4908 This issue was was reported to OpenSSL by Brian Carpenter.
4913 * ASN.1 structure reuse memory corruption fix
4915 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4916 memory corruption via an invalid write. Such reuse is and has been
4917 strongly discouraged and is believed to be rare.
4919 Applications that parse structures containing CHOICE or ANY DEFINED BY
4920 components may be affected. Certificate parsing (d2i_X509 and related
4921 functions) are however not affected. OpenSSL clients and servers are
4927 * PKCS7 NULL pointer dereferences fix
4929 The PKCS#7 parsing code does not handle missing outer ContentInfo
4930 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4931 missing content and trigger a NULL pointer dereference on parsing.
4933 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4934 otherwise parse PKCS#7 structures from untrusted sources are
4935 affected. OpenSSL clients and servers are not affected.
4937 This issue was reported to OpenSSL by Michal Zalewski (Google).
4942 * DoS via reachable assert in SSLv2 servers fix
4944 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4945 servers that both support SSLv2 and enable export cipher suites by sending
4946 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4948 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4949 (OpenSSL development team).
4954 * Empty CKE with client auth and DHE fix
4956 If client auth is used then a server can seg fault in the event of a DHE
4957 ciphersuite being selected and a zero length ClientKeyExchange message
4958 being sent by the client. This could be exploited in a DoS attack.
4963 * Handshake with unseeded PRNG fix
4965 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4966 with an unseeded PRNG. The conditions are:
4967 - The client is on a platform where the PRNG has not been seeded
4968 automatically, and the user has not seeded manually
4969 - A protocol specific client method version has been used (i.e. not
4970 SSL_client_methodv23)
4971 - A ciphersuite is used that does not require additional random data from
4972 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4974 If the handshake succeeds then the client random that has been used will
4975 have been generated from a PRNG with insufficient entropy and therefore the
4976 output may be predictable.
4978 For example using the following command with an unseeded openssl will
4979 succeed on an unpatched platform:
4981 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4986 * Use After Free following d2i_ECPrivatekey error fix
4988 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4989 could cause a use after free condition. This, in turn, could cause a double
4990 free in several private key parsing functions (such as d2i_PrivateKey
4991 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4992 for applications that receive EC private keys from untrusted
4993 sources. This scenario is considered rare.
4995 This issue was discovered by the BoringSSL project and fixed in their
5001 * X509_to_X509_REQ NULL pointer deref fix
5003 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5004 the certificate key is invalid. This function is rarely used in practice.
5006 This issue was discovered by Brian Carpenter.
5011 * Removed the export ciphers from the DEFAULT ciphers
5015 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5017 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5018 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5019 So far those who have to target multiple platforms would compromise
5020 and argue that binary targeting say ARMv5 would still execute on
5021 ARMv8. "Universal" build resolves this compromise by providing
5022 near-optimal performance even on newer platforms.
5026 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5027 (other platforms pending).
5029 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5031 * Add support for the SignedCertificateTimestampList certificate and
5032 OCSP response extensions from RFC6962.
5036 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5037 for corner cases. (Certain input points at infinity could lead to
5038 bogus results, with non-infinity inputs mapped to infinity too.)
5042 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5043 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5044 common cases are optimized and there still is room for further
5045 improvements. Vector Permutation AES for Altivec is also added.
5049 * Add support for little-endian ppc64 Linux target.
5051 *Marcelo Cerri (IBM)*
5053 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5054 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5055 are optimized and there still is room for further improvements.
5056 Both 32- and 64-bit modes are supported.
5058 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5060 * Improved ARMv7 NEON support.
5064 * Support for SPARC Architecture 2011 crypto extensions, first
5065 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5066 SHA256/512, MD5, GHASH and modular exponentiation.
5068 *Andy Polyakov, David Miller*
5070 * Accelerated modular exponentiation for Intel processors, a.k.a.
5073 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5075 * Support for new and upcoming Intel processors, including AVX2,
5076 BMI and SHA ISA extensions. This includes additional "stitched"
5077 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5080 This work was sponsored by Intel Corp.
5084 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5085 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5086 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5090 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5091 this fixes a limitation in previous versions of OpenSSL.
5095 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5096 MGF1 digest and OAEP label.
5100 * Add EVP support for key wrapping algorithms, to avoid problems with
5101 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5102 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5103 algorithms and include tests cases.
5107 * Add functions to allocate and set the fields of an ECDSA_METHOD
5110 *Douglas E. Engert, Steve Henson*
5112 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5113 difference in days and seconds between two tm or ASN1_TIME structures.
5117 * Add -rev test option to s_server to just reverse order of characters
5118 received by client and send back to server. Also prints an abbreviated
5119 summary of the connection parameters.
5123 * New option -brief for s_client and s_server to print out a brief summary
5124 of connection parameters.
5128 * Add callbacks for arbitrary TLS extensions.
5130 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5132 * New option -crl_download in several openssl utilities to download CRLs
5133 from CRLDP extension in certificates.
5137 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5141 * New function X509_CRL_diff to generate a delta CRL from the difference
5142 of two full CRLs. Add support to "crl" utility.
5146 * New functions to set lookup_crls function and to retrieve
5147 X509_STORE from X509_STORE_CTX.
5151 * Print out deprecated issuer and subject unique ID fields in
5156 * Extend OCSP I/O functions so they can be used for simple general purpose
5157 HTTP as well as OCSP. New wrapper function which can be used to download
5158 CRLs using the OCSP API.
5162 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5166 * `SSL_CONF*` functions. These provide a common framework for application
5167 configuration using configuration files or command lines.
5171 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5172 message callback and prints the results. Needs compile time option
5173 "enable-ssl-trace". New options to s_client and s_server to enable
5178 * New ctrl and macro to retrieve supported points extensions.
5179 Print out extension in s_server and s_client.
5183 * New functions to retrieve certificate signature and signature
5188 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5193 * New Suite B modes for TLS code. These use and enforce the requirements
5194 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5195 only use Suite B curves. The Suite B modes can be set by using the
5196 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5200 * New chain verification flags for Suite B levels of security. Check
5201 algorithms are acceptable when flags are set in X509_verify_cert.
5205 * Make tls1_check_chain return a set of flags indicating checks passed
5206 by a certificate chain. Add additional tests to handle client
5207 certificates: checks for matching certificate type and issuer name
5212 * If an attempt is made to use a signature algorithm not in the peer
5213 preference list abort the handshake. If client has no suitable
5214 signature algorithms in response to a certificate request do not
5215 use the certificate.
5219 * If server EC tmp key is not in client preference list abort handshake.
5223 * Add support for certificate stores in CERT structure. This makes it
5224 possible to have different stores per SSL structure or one store in
5225 the parent SSL_CTX. Include distinct stores for certificate chain
5226 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5227 to build and store a certificate chain in CERT structure: returning
5228 an error if the chain cannot be built: this will allow applications
5229 to test if a chain is correctly configured.
5231 Note: if the CERT based stores are not set then the parent SSL_CTX
5232 store is used to retain compatibility with existing behaviour.
5236 * New function ssl_set_client_disabled to set a ciphersuite disabled
5237 mask based on the current session, check mask when sending client
5238 hello and checking the requested ciphersuite.
5242 * New ctrls to retrieve and set certificate types in a certificate
5243 request message. Print out received values in s_client. If certificate
5244 types is not set with custom values set sensible values based on
5245 supported signature algorithms.
5249 * Support for distinct client and server supported signature algorithms.
5253 * Add certificate callback. If set this is called whenever a certificate
5254 is required by client or server. An application can decide which
5255 certificate chain to present based on arbitrary criteria: for example
5256 supported signature algorithms. Add very simple example to s_server.
5257 This fixes many of the problems and restrictions of the existing client
5258 certificate callback: for example you can now clear an existing
5259 certificate and specify the whole chain.
5263 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5264 the certificate can be used for (if anything). Set valid_flags field
5265 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5266 to have similar checks in it.
5268 Add new "cert_flags" field to CERT structure and include a "strict mode".
5269 This enforces some TLS certificate requirements (such as only permitting
5270 certificate signature algorithms contained in the supported algorithms
5271 extension) which some implementations ignore: this option should be used
5272 with caution as it could cause interoperability issues.
5276 * Update and tidy signature algorithm extension processing. Work out
5277 shared signature algorithms based on preferences and peer algorithms
5278 and print them out in s_client and s_server. Abort handshake if no
5279 shared signature algorithms.
5283 * Add new functions to allow customised supported signature algorithms
5284 for SSL and SSL_CTX structures. Add options to s_client and s_server
5289 * New function SSL_certs_clear() to delete all references to certificates
5290 from an SSL structure. Before this once a certificate had been added
5291 it couldn't be removed.
5295 * Integrate hostname, email address and IP address checking with certificate
5296 verification. New verify options supporting checking in openssl utility.
5300 * Fixes and wildcard matching support to hostname and email checking
5301 functions. Add manual page.
5303 *Florian Weimer (Red Hat Product Security Team)*
5305 * New functions to check a hostname email or IP address against a
5306 certificate. Add options x509 utility to print results of checks against
5311 * Fix OCSP checking.
5313 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5315 * Initial experimental support for explicitly trusted non-root CAs.
5316 OpenSSL still tries to build a complete chain to a root but if an
5317 intermediate CA has a trust setting included that is used. The first
5318 setting is used: whether to trust (e.g., -addtrust option to the x509
5323 * Add -trusted_first option which attempts to find certificates in the
5324 trusted store even if an untrusted chain is also supplied.
5328 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5329 platform support for Linux and Android.
5333 * Support for linux-x32, ILP32 environment in x86_64 framework.
5337 * Experimental multi-implementation support for FIPS capable OpenSSL.
5338 When in FIPS mode the approved implementations are used as normal,
5339 when not in FIPS mode the internal unapproved versions are used instead.
5340 This means that the FIPS capable OpenSSL isn't forced to use the
5341 (often lower performance) FIPS implementations outside FIPS mode.
5345 * Transparently support X9.42 DH parameters when calling
5346 PEM_read_bio_DHparameters. This means existing applications can handle
5347 the new parameter format automatically.
5351 * Initial experimental support for X9.42 DH parameter format: mainly
5352 to support use of 'q' parameter for RFC5114 parameters.
5356 * Add DH parameters from RFC5114 including test data to dhtest.
5360 * Support for automatic EC temporary key parameter selection. If enabled
5361 the most preferred EC parameters are automatically used instead of
5362 hardcoded fixed parameters. Now a server just has to call:
5363 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5364 support ECDH and use the most appropriate parameters.
5368 * Enhance and tidy EC curve and point format TLS extension code. Use
5369 static structures instead of allocation if default values are used.
5370 New ctrls to set curves we wish to support and to retrieve shared curves.
5371 Print out shared curves in s_server. New options to s_server and s_client
5372 to set list of supported curves.
5376 * New ctrls to retrieve supported signature algorithms and
5377 supported curve values as an array of NIDs. Extend openssl utility
5378 to print out received values.
5382 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5383 between NIDs and the more common NIST names such as "P-256". Enhance
5384 ecparam utility and ECC method to recognise the NIST names for curves.
5388 * Enhance SSL/TLS certificate chain handling to support different
5389 chains for each certificate instead of one chain in the parent SSL_CTX.
5393 * Support for fixed DH ciphersuite client authentication: where both
5394 server and client use DH certificates with common parameters.
5398 * Support for fixed DH ciphersuites: those requiring DH server
5403 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5405 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5406 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5407 X509_CINF_get_signature were reverted post internal team review.
5412 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5414 * OCSP Status Request extension unbounded memory growth
5416 A malicious client can send an excessively large OCSP Status Request
5417 extension. If that client continually requests renegotiation, sending a
5418 large OCSP Status Request extension each time, then there will be unbounded
5419 memory growth on the server. This will eventually lead to a Denial Of
5420 Service attack through memory exhaustion. Servers with a default
5421 configuration are vulnerable even if they do not support OCSP. Builds using
5422 the "no-ocsp" build time option are not affected.
5424 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5429 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5432 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5438 * OOB write in MDC2_Update()
5440 An overflow can occur in MDC2_Update() either if called directly or
5441 through the EVP_DigestUpdate() function using MDC2. If an attacker
5442 is able to supply very large amounts of input data after a previous
5443 call to EVP_EncryptUpdate() with a partial block then a length check
5444 can overflow resulting in a heap corruption.
5446 The amount of data needed is comparable to SIZE_MAX which is impractical
5449 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5454 * Malformed SHA512 ticket DoS
5456 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5457 DoS attack where a malformed ticket will result in an OOB read which will
5460 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5461 a custom server callback and ticket lookup mechanism.
5463 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5468 * OOB write in BN_bn2dec()
5470 The function BN_bn2dec() does not check the return value of BN_div_word().
5471 This can cause an OOB write if an application uses this function with an
5472 overly large BIGNUM. This could be a problem if an overly large certificate
5473 or CRL is printed out from an untrusted source. TLS is not affected because
5474 record limits will reject an oversized certificate before it is parsed.
5476 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5481 * OOB read in TS_OBJ_print_bio()
5483 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5484 the total length the OID text representation would use and not the amount
5485 of data written. This will result in OOB reads when large OIDs are
5488 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5493 * Pointer arithmetic undefined behaviour
5495 Avoid some undefined pointer arithmetic
5497 A common idiom in the codebase is to check limits in the following manner:
5500 Where "p" points to some malloc'd data of SIZE bytes and
5503 "len" here could be from some externally supplied data (e.g. from a TLS
5506 The rules of C pointer arithmetic are such that "p + len" is only well
5507 defined where len <= SIZE. Therefore the above idiom is actually
5508 undefined behaviour.
5510 For example this could cause problems if some malloc implementation
5511 provides an address for "p" such that "p + len" actually overflows for
5512 values of len that are too big and therefore p + len < limit.
5514 This issue was reported to OpenSSL by Guido Vranken
5519 * Constant time flag not preserved in DSA signing
5521 Operations in the DSA signing algorithm should run in constant time in
5522 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5523 implementation means that a non-constant time codepath is followed for
5524 certain operations. This has been demonstrated through a cache-timing
5525 attack to be sufficient for an attacker to recover the private DSA key.
5527 This issue was reported by César Pereida (Aalto University), Billy Brumley
5528 (Tampere University of Technology), and Yuval Yarom (The University of
5529 Adelaide and NICTA).
5534 * DTLS buffered message DoS
5536 In a DTLS connection where handshake messages are delivered out-of-order
5537 those messages that OpenSSL is not yet ready to process will be buffered
5538 for later use. Under certain circumstances, a flaw in the logic means that
5539 those messages do not get removed from the buffer even though the handshake
5540 has been completed. An attacker could force up to approx. 15 messages to
5541 remain in the buffer when they are no longer required. These messages will
5542 be cleared when the DTLS connection is closed. The default maximum size for
5543 a message is 100k. Therefore the attacker could force an additional 1500k
5544 to be consumed per connection. By opening many simulataneous connections an
5545 attacker could cause a DoS attack through memory exhaustion.
5547 This issue was reported to OpenSSL by Quan Luo.
5552 * DTLS replay protection DoS
5554 A flaw in the DTLS replay attack protection mechanism means that records
5555 that arrive for future epochs update the replay protection "window" before
5556 the MAC for the record has been validated. This could be exploited by an
5557 attacker by sending a record for the next epoch (which does not have to
5558 decrypt or have a valid MAC), with a very large sequence number. This means
5559 that all subsequent legitimate packets are dropped causing a denial of
5560 service for a specific DTLS connection.
5562 This issue was reported to OpenSSL by the OCAP audit team.
5567 * Certificate message OOB reads
5569 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5570 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5571 theoretical DoS risk but this has not been observed in practice on common
5574 The messages affected are client certificate, client certificate request
5575 and server certificate. As a result the attack can only be performed
5576 against a client or a server which enables client authentication.
5578 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5583 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5585 * Prevent padding oracle in AES-NI CBC MAC check
5587 A MITM attacker can use a padding oracle attack to decrypt traffic
5588 when the connection uses an AES CBC cipher and the server support
5591 This issue was introduced as part of the fix for Lucky 13 padding
5592 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5593 constant time by making sure that always the same bytes are read and
5594 compared against either the MAC or padding bytes. But it no longer
5595 checked that there was enough data to have both the MAC and padding
5598 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5603 * Fix EVP_EncodeUpdate overflow
5605 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5606 Base64 encoding of binary data. If an attacker is able to supply very large
5607 amounts of input data then a length check can overflow resulting in a heap
5610 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5611 the `PEM_write_bio*` family of functions. These are mainly used within the
5612 OpenSSL command line applications, so any application which processes data
5613 from an untrusted source and outputs it as a PEM file should be considered
5614 vulnerable to this issue. User applications that call these APIs directly
5615 with large amounts of untrusted data may also be vulnerable.
5617 This issue was reported by Guido Vranken.
5622 * Fix EVP_EncryptUpdate overflow
5624 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5625 is able to supply very large amounts of input data after a previous call to
5626 EVP_EncryptUpdate() with a partial block then a length check can overflow
5627 resulting in a heap corruption. Following an analysis of all OpenSSL
5628 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5629 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5630 the first called function after an EVP_EncryptInit(), and therefore that
5631 specific call must be safe. The second form is where the length passed to
5632 EVP_EncryptUpdate() can be seen from the code to be some small value and
5633 therefore there is no possibility of an overflow. Since all instances are
5634 one of these two forms, it is believed that there can be no overflows in
5635 internal code due to this problem. It should be noted that
5636 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5637 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5638 of these calls have also been analysed too and it is believed there are no
5639 instances in internal usage where an overflow could occur.
5641 This issue was reported by Guido Vranken.
5646 * Prevent ASN.1 BIO excessive memory allocation
5648 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5649 a short invalid encoding can casuse allocation of large amounts of memory
5650 potentially consuming excessive resources or exhausting memory.
5652 Any application parsing untrusted data through d2i BIO functions is
5653 affected. The memory based functions such as d2i_X509() are *not* affected.
5654 Since the memory based functions are used by the TLS library, TLS
5655 applications are not affected.
5657 This issue was reported by Brian Carpenter.
5664 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5665 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5666 in arbitrary stack data being returned in the buffer.
5668 This issue was reported by Guido Vranken.
5673 * Modify behavior of ALPN to invoke callback after SNI/servername
5674 callback, such that updates to the SSL_CTX affect ALPN.
5678 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5683 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5684 methods are enabled and ssl2 is disabled the methods return NULL.
5688 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5690 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5691 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5692 provide any "EXPORT" or "LOW" strength ciphers.
5696 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5697 is by default disabled at build-time. Builds that are not configured with
5698 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5699 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5700 will need to explicitly call either of:
5702 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5704 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5706 as appropriate. Even if either of those is used, or the application
5707 explicitly uses the version-specific SSLv2_method() or its client and
5708 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5709 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5710 ciphers, and SSLv2 56-bit DES are no longer available.
5715 * Fix a double-free in DSA code
5717 A double free bug was discovered when OpenSSL parses malformed DSA private
5718 keys and could lead to a DoS attack or memory corruption for applications
5719 that receive DSA private keys from untrusted sources. This scenario is
5722 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5728 * Disable SRP fake user seed to address a server memory leak.
5730 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5732 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5733 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5734 was changed to ignore the "fake user" SRP seed, even if the seed
5737 Users should use SRP_VBASE_get1_by_user instead. Note that in
5738 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5739 also that even though configuring the SRP seed attempts to hide
5740 invalid usernames by continuing the handshake with fake
5741 credentials, this behaviour is not constant time and no strong
5742 guarantees are made that the handshake is indistinguishable from
5743 that of a valid user.
5748 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5750 In the BN_hex2bn function the number of hex digits is calculated using an
5751 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5752 large values of `i` this can result in `bn_expand` not allocating any
5753 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5754 field as NULL leading to a subsequent NULL ptr deref. For very large values
5755 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5756 In this case memory is allocated to the internal BIGNUM data field, but it
5757 is insufficiently sized leading to heap corruption. A similar issue exists
5758 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5759 is ever called by user applications with very large untrusted hex/dec data.
5760 This is anticipated to be a rare occurrence.
5762 All OpenSSL internal usage of these functions use data that is not expected
5763 to be untrusted, e.g. config file data or application command line
5764 arguments. If user developed applications generate config file data based
5765 on untrusted data then it is possible that this could also lead to security
5766 consequences. This is also anticipated to be rare.
5768 This issue was reported to OpenSSL by Guido Vranken.
5773 * Fix memory issues in `BIO_*printf` functions
5775 The internal `fmtstr` function used in processing a "%s" format string in
5776 the `BIO_*printf` functions could overflow while calculating the length of a
5777 string and cause an OOB read when printing very long strings.
5779 Additionally the internal `doapr_outch` function can attempt to write to an
5780 OOB memory location (at an offset from the NULL pointer) in the event of a
5781 memory allocation failure. In 1.0.2 and below this could be caused where
5782 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5783 could be in processing a very long "%s" format string. Memory leaks can
5786 The first issue may mask the second issue dependent on compiler behaviour.
5787 These problems could enable attacks where large amounts of untrusted data
5788 is passed to the `BIO_*printf` functions. If applications use these functions
5789 in this way then they could be vulnerable. OpenSSL itself uses these
5790 functions when printing out human-readable dumps of ASN.1 data. Therefore
5791 applications that print this data could be vulnerable if the data is from
5792 untrusted sources. OpenSSL command line applications could also be
5793 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5794 as command line arguments.
5796 Libssl is not considered directly vulnerable. Additionally certificates etc
5797 received via remote connections via libssl are also unlikely to be able to
5798 trigger these issues because of message size limits enforced within libssl.
5800 This issue was reported to OpenSSL Guido Vranken.
5805 * Side channel attack on modular exponentiation
5807 A side-channel attack was found which makes use of cache-bank conflicts on
5808 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5809 of RSA keys. The ability to exploit this issue is limited as it relies on
5810 an attacker who has control of code in a thread running on the same
5811 hyper-threaded core as the victim thread which is performing decryptions.
5813 This issue was reported to OpenSSL by Yuval Yarom, The University of
5814 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5815 Nadia Heninger, University of Pennsylvania with more information at
5816 <http://cachebleed.info>.
5821 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5822 if no keysize is specified with default_bits. This fixes an
5823 omission in an earlier change that changed all RSA/DSA key generation
5824 apps to use 2048 bits by default.
5828 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5830 * Protection for DH small subgroup attacks
5832 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5833 switched on by default and cannot be disabled. This could have some
5838 * SSLv2 doesn't block disabled ciphers
5840 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5841 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5842 been disabled, provided that the SSLv2 protocol was not also disabled via
5845 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5846 and Sebastian Schinzel.
5851 * Reject DH handshakes with parameters shorter than 1024 bits.
5855 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5857 * Certificate verify crash with missing PSS parameter
5859 The signature verification routines will crash with a NULL pointer
5860 dereference if presented with an ASN.1 signature using the RSA PSS
5861 algorithm and absent mask generation function parameter. Since these
5862 routines are used to verify certificate signature algorithms this can be
5863 used to crash any certificate verification operation and exploited in a
5864 DoS attack. Any application which performs certificate verification is
5865 vulnerable including OpenSSL clients and servers which enable client
5868 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5873 * X509_ATTRIBUTE memory leak
5875 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5876 memory. This structure is used by the PKCS#7 and CMS routines so any
5877 application which reads PKCS#7 or CMS data from untrusted sources is
5878 affected. SSL/TLS is not affected.
5880 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5886 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5887 This changes the decoding behaviour for some invalid messages,
5888 though the change is mostly in the more lenient direction, and
5889 legacy behaviour is preserved as much as possible.
5893 * In DSA_generate_parameters_ex, if the provided seed is too short,
5894 use a random seed, as already documented.
5896 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5898 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5900 * Alternate chains certificate forgery
5902 During certificate verfification, OpenSSL will attempt to find an
5903 alternative certificate chain if the first attempt to build such a chain
5904 fails. An error in the implementation of this logic can mean that an
5905 attacker could cause certain checks on untrusted certificates to be
5906 bypassed, such as the CA flag, enabling them to use a valid leaf
5907 certificate to act as a CA and "issue" an invalid certificate.
5909 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5915 * Race condition handling PSK identify hint
5917 If PSK identity hints are received by a multi-threaded client then
5918 the values are wrongly updated in the parent SSL_CTX structure. This can
5919 result in a race condition potentially leading to a double free of the
5925 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5927 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5928 incompatibility in the handling of HMAC. The previous ABI has now been
5931 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5933 * Malformed ECParameters causes infinite loop
5935 When processing an ECParameters structure OpenSSL enters an infinite loop
5936 if the curve specified is over a specially malformed binary polynomial
5939 This can be used to perform denial of service against any
5940 system which processes public keys, certificate requests or
5941 certificates. This includes TLS clients and TLS servers with
5942 client authentication enabled.
5944 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5949 * Exploitable out-of-bounds read in X509_cmp_time
5951 X509_cmp_time does not properly check the length of the ASN1_TIME
5952 string and can read a few bytes out of bounds. In addition,
5953 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5956 An attacker can use this to craft malformed certificates and CRLs of
5957 various sizes and potentially cause a segmentation fault, resulting in
5958 a DoS on applications that verify certificates or CRLs. TLS clients
5959 that verify CRLs are affected. TLS clients and servers with client
5960 authentication enabled may be affected if they use custom verification
5963 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5964 independently by Hanno Böck.
5969 * PKCS7 crash with missing EnvelopedContent
5971 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5972 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5973 with missing content and trigger a NULL pointer dereference on parsing.
5975 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5976 structures from untrusted sources are affected. OpenSSL clients and
5977 servers are not affected.
5979 This issue was reported to OpenSSL by Michal Zalewski (Google).
5984 * CMS verify infinite loop with unknown hash function
5986 When verifying a signedData message the CMS code can enter an infinite loop
5987 if presented with an unknown hash function OID. This can be used to perform
5988 denial of service against any system which verifies signedData messages using
5990 This issue was reported to OpenSSL by Johannes Bauer.
5995 * Race condition handling NewSessionTicket
5997 If a NewSessionTicket is received by a multi-threaded client when attempting to
5998 reuse a previous ticket then a race condition can occur potentially leading to
5999 a double free of the ticket data.
6004 * Reject DH handshakes with parameters shorter than 768 bits.
6006 *Kurt Roeckx and Emilia Kasper*
6008 * dhparam: generate 2048-bit parameters by default.
6010 *Kurt Roeckx and Emilia Kasper*
6012 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6014 * Segmentation fault in ASN1_TYPE_cmp fix
6016 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6017 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6018 certificate signature algorithm consistency this can be used to crash any
6019 certificate verification operation and exploited in a DoS attack. Any
6020 application which performs certificate verification is vulnerable including
6021 OpenSSL clients and servers which enable client authentication.
6026 * ASN.1 structure reuse memory corruption fix
6028 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6029 memory corruption via an invalid write. Such reuse is and has been
6030 strongly discouraged and is believed to be rare.
6032 Applications that parse structures containing CHOICE or ANY DEFINED BY
6033 components may be affected. Certificate parsing (d2i_X509 and related
6034 functions) are however not affected. OpenSSL clients and servers are
6040 * PKCS7 NULL pointer dereferences fix
6042 The PKCS#7 parsing code does not handle missing outer ContentInfo
6043 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6044 missing content and trigger a NULL pointer dereference on parsing.
6046 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6047 otherwise parse PKCS#7 structures from untrusted sources are
6048 affected. OpenSSL clients and servers are not affected.
6050 This issue was reported to OpenSSL by Michal Zalewski (Google).
6055 * DoS via reachable assert in SSLv2 servers fix
6057 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6058 servers that both support SSLv2 and enable export cipher suites by sending
6059 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6061 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6062 (OpenSSL development team).
6067 * Use After Free following d2i_ECPrivatekey error fix
6069 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6070 could cause a use after free condition. This, in turn, could cause a double
6071 free in several private key parsing functions (such as d2i_PrivateKey
6072 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6073 for applications that receive EC private keys from untrusted
6074 sources. This scenario is considered rare.
6076 This issue was discovered by the BoringSSL project and fixed in their
6082 * X509_to_X509_REQ NULL pointer deref fix
6084 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6085 the certificate key is invalid. This function is rarely used in practice.
6087 This issue was discovered by Brian Carpenter.
6092 * Removed the export ciphers from the DEFAULT ciphers
6096 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6098 * Build fixes for the Windows and OpenVMS platforms
6100 *Matt Caswell and Richard Levitte*
6102 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6104 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6105 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6106 dereference. This could lead to a Denial Of Service attack. Thanks to
6107 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6112 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6113 dtls1_buffer_record function under certain conditions. In particular this
6114 could occur if an attacker sent repeated DTLS records with the same
6115 sequence number but for the next epoch. The memory leak could be exploited
6116 by an attacker in a Denial of Service attack through memory exhaustion.
6117 Thanks to Chris Mueller for reporting this issue.
6122 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6123 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6124 method would be set to NULL which could later result in a NULL pointer
6125 dereference. Thanks to Frank Schmirler for reporting this issue.
6130 * Abort handshake if server key exchange message is omitted for ephemeral
6133 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6134 reporting this issue.
6139 * Remove non-export ephemeral RSA code on client and server. This code
6140 violated the TLS standard by allowing the use of temporary RSA keys in
6141 non-export ciphersuites and could be used by a server to effectively
6142 downgrade the RSA key length used to a value smaller than the server
6143 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6144 INRIA or reporting this issue.
6149 * Fixed issue where DH client certificates are accepted without verification.
6150 An OpenSSL server will accept a DH certificate for client authentication
6151 without the certificate verify message. This effectively allows a client to
6152 authenticate without the use of a private key. This only affects servers
6153 which trust a client certificate authority which issues certificates
6154 containing DH keys: these are extremely rare and hardly ever encountered.
6155 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6161 * Ensure that the session ID context of an SSL is updated when its
6162 SSL_CTX is updated via SSL_set_SSL_CTX.
6164 The session ID context is typically set from the parent SSL_CTX,
6165 and can vary with the CTX.
6169 * Fix various certificate fingerprint issues.
6171 By using non-DER or invalid encodings outside the signed portion of a
6172 certificate the fingerprint can be changed without breaking the signature.
6173 Although no details of the signed portion of the certificate can be changed
6174 this can cause problems with some applications: e.g. those using the
6175 certificate fingerprint for blacklists.
6177 1. Reject signatures with non zero unused bits.
6179 If the BIT STRING containing the signature has non zero unused bits reject
6180 the signature. All current signature algorithms require zero unused bits.
6182 2. Check certificate algorithm consistency.
6184 Check the AlgorithmIdentifier inside TBS matches the one in the
6185 certificate signature. NB: this will result in signature failure
6186 errors for some broken certificates.
6188 Thanks to Konrad Kraszewski from Google for reporting this issue.
6190 3. Check DSA/ECDSA signatures use DER.
6192 Re-encode DSA/ECDSA signatures and compare with the original received
6193 signature. Return an error if there is a mismatch.
6195 This will reject various cases including garbage after signature
6196 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6197 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6198 (negative or with leading zeroes).
6200 Further analysis was conducted and fixes were developed by Stephen Henson
6201 of the OpenSSL core team.
6207 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6208 results on some platforms, including x86_64. This bug occurs at random
6209 with a very low probability, and is not known to be exploitable in any
6210 way, though its exact impact is difficult to determine. Thanks to Pieter
6211 Wuille (Blockstream) who reported this issue and also suggested an initial
6212 fix. Further analysis was conducted by the OpenSSL development team and
6213 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6214 the OpenSSL core team.
6219 * Do not resume sessions on the server if the negotiated protocol
6220 version does not match the session's version. Resuming with a different
6221 version, while not strictly forbidden by the RFC, is of questionable
6222 sanity and breaks all known clients.
6224 *David Benjamin, Emilia Käsper*
6226 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6227 early CCS messages during renegotiation. (Note that because
6228 renegotiation is encrypted, this early CCS was not exploitable.)
6232 * Tighten client-side session ticket handling during renegotiation:
6233 ensure that the client only accepts a session ticket if the server sends
6234 the extension anew in the ServerHello. Previously, a TLS client would
6235 reuse the old extension state and thus accept a session ticket if one was
6236 announced in the initial ServerHello.
6238 Similarly, ensure that the client requires a session ticket if one
6239 was advertised in the ServerHello. Previously, a TLS client would
6240 ignore a missing NewSessionTicket message.
6244 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6248 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6249 sends a carefully crafted handshake message, to cause OpenSSL to fail
6250 to free up to 64k of memory causing a memory leak. This could be
6251 exploited in a Denial Of Service attack. This issue affects OpenSSL
6252 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6253 whether SRTP is used or configured. Implementations of OpenSSL that
6254 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6256 The fix was developed by the OpenSSL team.
6261 * Session Ticket Memory Leak.
6263 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6264 integrity of that ticket is first verified. In the event of a session
6265 ticket integrity check failing, OpenSSL will fail to free memory
6266 causing a memory leak. By sending a large number of invalid session
6267 tickets an attacker could exploit this issue in a Denial Of Service
6273 * Build option no-ssl3 is incomplete.
6275 When OpenSSL is configured with "no-ssl3" as a build option, servers
6276 could accept and complete a SSL 3.0 handshake, and clients could be
6277 configured to send them.
6280 *Akamai and the OpenSSL team*
6282 * Add support for TLS_FALLBACK_SCSV.
6283 Client applications doing fallback retries should call
6284 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6287 *Adam Langley, Bodo Moeller*
6289 * Add additional DigestInfo checks.
6291 Re-encode DigestInto in DER and check against the original when
6292 verifying RSA signature: this will reject any improperly encoded
6293 DigestInfo structures.
6295 Note: this is a precautionary measure and no attacks are currently known.
6299 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6301 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6302 SRP code can be overrun an internal buffer. Add sanity check that
6303 g, A, B < N to SRP code.
6305 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6306 Group for discovering this issue.
6311 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6312 TLS 1.0 instead of higher protocol versions when the ClientHello message
6313 is badly fragmented. This allows a man-in-the-middle attacker to force a
6314 downgrade to TLS 1.0 even if both the server and the client support a
6315 higher protocol version, by modifying the client's TLS records.
6317 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6318 researching this issue.
6323 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6324 to a denial of service attack. A malicious server can crash the client
6325 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6326 ciphersuite and sending carefully crafted handshake messages.
6328 Thanks to Felix Gröbert (Google) for discovering and researching this
6334 * By sending carefully crafted DTLS packets an attacker could cause openssl
6335 to leak memory. This can be exploited through a Denial of Service attack.
6336 Thanks to Adam Langley for discovering and researching this issue.
6341 * An attacker can force openssl to consume large amounts of memory whilst
6342 processing DTLS handshake messages. This can be exploited through a
6343 Denial of Service attack.
6344 Thanks to Adam Langley for discovering and researching this issue.
6349 * An attacker can force an error condition which causes openssl to crash
6350 whilst processing DTLS packets due to memory being freed twice. This
6351 can be exploited through a Denial of Service attack.
6352 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6358 * If a multithreaded client connects to a malicious server using a resumed
6359 session and the server sends an ec point format extension it could write
6360 up to 255 bytes to freed memory.
6362 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6368 * A malicious server can crash an OpenSSL client with a null pointer
6369 dereference (read) by specifying an SRP ciphersuite even though it was not
6370 properly negotiated with the client. This can be exploited through a
6371 Denial of Service attack.
6373 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6374 discovering and researching this issue.
6379 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6380 X509_name_oneline, X509_name_print_ex et al. to leak some information
6381 from the stack. Applications may be affected if they echo pretty printing
6382 output to the attacker.
6384 Thanks to Ivan Fratric (Google) for discovering this issue.
6387 *Emilia Käsper, and Steve Henson*
6389 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6390 for corner cases. (Certain input points at infinity could lead to
6391 bogus results, with non-infinity inputs mapped to infinity too.)
6395 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6397 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6398 handshake can force the use of weak keying material in OpenSSL
6399 SSL/TLS clients and servers.
6401 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6402 researching this issue. [CVE-2014-0224][]
6404 *KIKUCHI Masashi, Steve Henson*
6406 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6407 OpenSSL DTLS client the code can be made to recurse eventually crashing
6410 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6413 *Imre Rad, Steve Henson*
6415 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6416 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6417 client or server. This is potentially exploitable to run arbitrary
6418 code on a vulnerable client or server.
6420 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6422 *Jüri Aedla, Steve Henson*
6424 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6425 are subject to a denial of service attack.
6427 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6428 this issue. [CVE-2014-3470][]
6430 *Felix Gröbert, Ivan Fratric, Steve Henson*
6432 * Harmonize version and its documentation. -f flag is used to display
6435 *mancha <mancha1@zoho.com>*
6437 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6438 in i2d_ECPrivateKey.
6440 *mancha <mancha1@zoho.com>*
6442 * Fix some double frees. These are not thought to be exploitable.
6444 *mancha <mancha1@zoho.com>*
6446 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6448 * A missing bounds check in the handling of the TLS heartbeat extension
6449 can be used to reveal up to 64k of memory to a connected client or
6452 Thanks for Neel Mehta of Google Security for discovering this bug and to
6453 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6454 preparing the fix [CVE-2014-0160][]
6456 *Adam Langley, Bodo Moeller*
6458 * Fix for the attack described in the paper "Recovering OpenSSL
6459 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6460 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6461 <http://eprint.iacr.org/2014/140>
6463 Thanks to Yuval Yarom and Naomi Benger for discovering this
6464 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6466 *Yuval Yarom and Naomi Benger*
6468 * TLS pad extension: draft-agl-tls-padding-03
6470 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6471 TLS client Hello record length value would otherwise be > 255 and
6472 less that 512 pad with a dummy extension containing zeroes so it
6473 is at least 512 bytes long.
6475 *Adam Langley, Steve Henson*
6477 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6479 * Fix for TLS record tampering bug. A carefully crafted invalid
6480 handshake could crash OpenSSL with a NULL pointer exception.
6481 Thanks to Anton Johansson for reporting this issues.
6484 * Keep original DTLS digest and encryption contexts in retransmission
6485 structures so we can use the previous session parameters if they need
6486 to be resent. [CVE-2013-6450][]
6490 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6491 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6492 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6493 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6494 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6495 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6497 *Rob Stradling, Adam Langley*
6499 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6501 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6502 supporting platforms or when small records were transferred.
6504 *Andy Polyakov, Steve Henson*
6506 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6508 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6510 This addresses the flaw in CBC record processing discovered by
6511 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6512 at: <http://www.isg.rhul.ac.uk/tls/>
6514 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6515 Security Group at Royal Holloway, University of London
6516 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6517 Emilia Käsper for the initial patch.
6520 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6522 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6523 ciphersuites which can be exploited in a denial of service attack.
6524 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6525 and detecting this bug and to Wolfgang Ettlinger
6526 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6531 * Return an error when checking OCSP signatures when key is NULL.
6532 This fixes a DoS attack. [CVE-2013-0166][]
6536 * Make openssl verify return errors.
6538 *Chris Palmer <palmer@google.com> and Ben Laurie*
6540 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6541 the right response is stapled. Also change SSL_get_certificate()
6542 so it returns the certificate actually sent.
6543 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6545 *Rob Stradling <rob.stradling@comodo.com>*
6547 * Fix possible deadlock when decoding public keys.
6551 * Don't use TLS 1.0 record version number in initial client hello
6556 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6558 * Sanity check record length before skipping explicit IV in TLS
6559 1.2, 1.1 and DTLS to fix DoS attack.
6561 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6562 fuzzing as a service testing platform.
6567 * Initialise tkeylen properly when encrypting CMS messages.
6568 Thanks to Solar Designer of Openwall for reporting this issue.
6572 * In FIPS mode don't try to use composite ciphers as they are not
6577 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6579 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6580 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6581 mean any application compiled against OpenSSL 1.0.0 headers setting
6582 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6583 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6584 0x10000000L Any application which was previously compiled against
6585 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6586 will need to be recompiled as a result. Letting be results in
6587 inability to disable specifically TLS 1.1 and in client context,
6588 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6592 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6593 disable just protocol X, but all protocols above X *if* there are
6594 protocols *below* X still enabled. In more practical terms it means
6595 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6596 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6597 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6602 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6604 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6605 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6606 in CRYPTO_realloc_clean.
6608 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6609 issue and to Adam Langley <agl@chromium.org> for fixing it.
6612 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6614 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6618 * Workarounds for some broken servers that "hang" if a client hello
6619 record length exceeds 255 bytes.
6621 1. Do not use record version number > TLS 1.0 in initial client
6622 hello: some (but not all) hanging servers will now work.
6623 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6624 the number of ciphers sent in the client hello. This should be
6625 set to an even number, such as 50, for example by passing:
6626 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6627 Most broken servers should now work.
6628 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6629 TLS 1.2 client support entirely.
6633 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6637 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6639 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6640 STRING form instead of a DigestInfo.
6644 * The format used for MDC2 RSA signatures is inconsistent between EVP
6645 and the RSA_sign/RSA_verify functions. This was made more apparent when
6646 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6647 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6648 the correct format in RSA_verify so both forms transparently work.
6652 * Some servers which support TLS 1.0 can choke if we initially indicate
6653 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6654 encrypted premaster secret. As a workaround use the maximum permitted
6655 client version in client hello, this should keep such servers happy
6656 and still work with previous versions of OpenSSL.
6660 * Add support for TLS/DTLS heartbeats.
6662 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6664 * Add support for SCTP.
6666 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6668 * Improved PRNG seeding for VOS.
6670 *Paul Green <Paul.Green@stratus.com>*
6672 * Extensive assembler packs updates, most notably:
6674 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6675 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6676 - x86_64: bit-sliced AES implementation;
6677 - ARM: NEON support, contemporary platforms optimizations;
6678 - s390x: z196 support;
6679 - `*`: GHASH and GF(2^m) multiplication implementations;
6683 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6684 (removal of unnecessary code)
6686 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6688 * Add TLS key material exporter from RFC 5705.
6692 * Add DTLS-SRTP negotiation from RFC 5764.
6696 * Add Next Protocol Negotiation,
6697 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6698 disabled with a no-npn flag to config or Configure. Code donated
6701 *Adam Langley <agl@google.com> and Ben Laurie*
6703 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6704 NIST-P256, NIST-P521, with constant-time single point multiplication on
6705 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6706 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6707 Code made available under Apache License version 2.0.
6709 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6710 line to include this in your build of OpenSSL, and run "make depend" (or
6711 "make update"). This enables the following EC_METHODs:
6713 EC_GFp_nistp224_method()
6714 EC_GFp_nistp256_method()
6715 EC_GFp_nistp521_method()
6717 EC_GROUP_new_by_curve_name() will automatically use these (while
6718 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6721 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6723 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6724 all platforms. Move ssize_t definition from e_os.h to the public
6725 header file e_os2.h as it now appears in public header file cms.h
6729 * New -sigopt option to the ca, req and x509 utilities. Additional
6730 signature parameters can be passed using this option and in
6735 * Add RSA PSS signing function. This will generate and set the
6736 appropriate AlgorithmIdentifiers for PSS based on those in the
6737 corresponding EVP_MD_CTX structure. No application support yet.
6741 * Support for companion algorithm specific ASN1 signing routines.
6742 New function ASN1_item_sign_ctx() signs a pre-initialised
6743 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6744 the appropriate parameters.
6748 * Add new algorithm specific ASN1 verification initialisation function
6749 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6750 handling will be the same no matter what EVP_PKEY_METHOD is used.
6751 Add a PSS handler to support verification of PSS signatures: checked
6752 against a number of sample certificates.
6756 * Add signature printing for PSS. Add PSS OIDs.
6758 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6760 * Add algorithm specific signature printing. An individual ASN1 method
6761 can now print out signatures instead of the standard hex dump.
6763 More complex signatures (e.g. PSS) can print out more meaningful
6764 information. Include DSA version that prints out the signature
6769 * Password based recipient info support for CMS library: implementing
6774 * Split password based encryption into PBES2 and PBKDF2 functions. This
6775 neatly separates the code into cipher and PBE sections and is required
6776 for some algorithms that split PBES2 into separate pieces (such as
6777 password based CMS).
6781 * Session-handling fixes:
6782 - Fix handling of connections that are resuming with a session ID,
6783 but also support Session Tickets.
6784 - Fix a bug that suppressed issuing of a new ticket if the client
6785 presented a ticket with an expired session.
6786 - Try to set the ticket lifetime hint to something reasonable.
6787 - Make tickets shorter by excluding irrelevant information.
6788 - On the client side, don't ignore renewed tickets.
6790 *Adam Langley, Bodo Moeller (Google)*
6792 * Fix PSK session representation.
6796 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6798 This work was sponsored by Intel.
6802 * Add GCM support to TLS library. Some custom code is needed to split
6803 the IV between the fixed (from PRF) and explicit (from TLS record)
6804 portions. This adds all GCM ciphersuites supported by RFC5288 and
6805 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6806 add a special AESGCM string for GCM only.
6810 * Expand range of ctrls for AES GCM. Permit setting invocation
6811 field on decrypt and retrieval of invocation field only on encrypt.
6815 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6816 As required by RFC5289 these ciphersuites cannot be used if for
6817 versions of TLS earlier than 1.2.
6821 * For FIPS capable OpenSSL interpret a NULL default public key method
6822 as unset and return the appropriate default but do *not* set the default.
6823 This means we can return the appropriate method in applications that
6824 switch between FIPS and non-FIPS modes.
6828 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6829 ENGINE is used then we cannot handle that in the FIPS module so we
6830 keep original code iff non-FIPS operations are allowed.
6834 * Add -attime option to openssl utilities.
6836 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6838 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6842 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6843 FIPS EC methods unconditionally for now.
6847 * New build option no-ec2m to disable characteristic 2 code.
6851 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6852 all cases can be covered as some introduce binary incompatibilities.
6856 * Redirect RSA operations to FIPS module including keygen,
6857 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6861 * Add similar low level API blocking to ciphers.
6865 * Low level digest APIs are not approved in FIPS mode: any attempt
6866 to use these will cause a fatal error. Applications that *really* want
6867 to use them can use the `private_*` version instead.
6871 * Redirect cipher operations to FIPS module for FIPS builds.
6875 * Redirect digest operations to FIPS module for FIPS builds.
6879 * Update build system to add "fips" flag which will link in fipscanister.o
6880 for static and shared library builds embedding a signature if needed.
6884 * Output TLS supported curves in preference order instead of numerical
6885 order. This is currently hardcoded for the highest order curves first.
6886 This should be configurable so applications can judge speed vs strength.
6890 * Add TLS v1.2 server support for client authentication.
6894 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6899 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6900 FIPS modules versions.
6904 * Add TLS v1.2 client side support for client authentication. Keep cache
6905 of handshake records longer as we don't know the hash algorithm to use
6906 until after the certificate request message is received.
6910 * Initial TLS v1.2 client support. Add a default signature algorithms
6911 extension including all the algorithms we support. Parse new signature
6912 format in client key exchange. Relax some ECC signing restrictions for
6913 TLS v1.2 as indicated in RFC5246.
6917 * Add server support for TLS v1.2 signature algorithms extension. Switch
6918 to new signature format when needed using client digest preference.
6919 All server ciphersuites should now work correctly in TLS v1.2. No client
6920 support yet and no support for client certificates.
6924 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6925 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6926 ciphersuites. At present only RSA key exchange ciphersuites work with
6927 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6928 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6929 and version checking.
6933 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6934 with this defined it will not be affected by any changes to ssl internal
6935 structures. Add several utility functions to allow openssl application
6936 to work with OPENSSL_NO_SSL_INTERN defined.
6940 * A long standing patch to add support for SRP from EdelWeb (Peter
6941 Sylvester and Christophe Renou) was integrated.
6942 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6943 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6946 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6950 * Permit abbreviated handshakes when renegotiating using the function
6951 SSL_renegotiate_abbreviated().
6953 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6955 * Add call to ENGINE_register_all_complete() to
6956 ENGINE_load_builtin_engines(), so some implementations get used
6957 automatically instead of needing explicit application support.
6961 * Add support for TLS key exporter as described in RFC5705.
6963 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6965 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6966 a few changes are required:
6968 Add SSL_OP_NO_TLSv1_1 flag.
6969 Add TLSv1_1 methods.
6970 Update version checking logic to handle version 1.1.
6971 Add explicit IV handling (ported from DTLS code).
6972 Add command line options to s_client/s_server.
6979 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6981 * X509_ATTRIBUTE memory leak
6983 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6984 memory. This structure is used by the PKCS#7 and CMS routines so any
6985 application which reads PKCS#7 or CMS data from untrusted sources is
6986 affected. SSL/TLS is not affected.
6988 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6994 * Race condition handling PSK identify hint
6996 If PSK identity hints are received by a multi-threaded client then
6997 the values are wrongly updated in the parent SSL_CTX structure. This can
6998 result in a race condition potentially leading to a double free of the
7004 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7006 * Malformed ECParameters causes infinite loop
7008 When processing an ECParameters structure OpenSSL enters an infinite loop
7009 if the curve specified is over a specially malformed binary polynomial
7012 This can be used to perform denial of service against any
7013 system which processes public keys, certificate requests or
7014 certificates. This includes TLS clients and TLS servers with
7015 client authentication enabled.
7017 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7022 * Exploitable out-of-bounds read in X509_cmp_time
7024 X509_cmp_time does not properly check the length of the ASN1_TIME
7025 string and can read a few bytes out of bounds. In addition,
7026 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7029 An attacker can use this to craft malformed certificates and CRLs of
7030 various sizes and potentially cause a segmentation fault, resulting in
7031 a DoS on applications that verify certificates or CRLs. TLS clients
7032 that verify CRLs are affected. TLS clients and servers with client
7033 authentication enabled may be affected if they use custom verification
7036 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7037 independently by Hanno Böck.
7042 * PKCS7 crash with missing EnvelopedContent
7044 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7045 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7046 with missing content and trigger a NULL pointer dereference on parsing.
7048 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7049 structures from untrusted sources are affected. OpenSSL clients and
7050 servers are not affected.
7052 This issue was reported to OpenSSL by Michal Zalewski (Google).
7057 * CMS verify infinite loop with unknown hash function
7059 When verifying a signedData message the CMS code can enter an infinite loop
7060 if presented with an unknown hash function OID. This can be used to perform
7061 denial of service against any system which verifies signedData messages using
7063 This issue was reported to OpenSSL by Johannes Bauer.
7068 * Race condition handling NewSessionTicket
7070 If a NewSessionTicket is received by a multi-threaded client when attempting to
7071 reuse a previous ticket then a race condition can occur potentially leading to
7072 a double free of the ticket data.
7077 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7079 * Segmentation fault in ASN1_TYPE_cmp fix
7081 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7082 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7083 certificate signature algorithm consistency this can be used to crash any
7084 certificate verification operation and exploited in a DoS attack. Any
7085 application which performs certificate verification is vulnerable including
7086 OpenSSL clients and servers which enable client authentication.
7091 * ASN.1 structure reuse memory corruption fix
7093 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7094 memory corruption via an invalid write. Such reuse is and has been
7095 strongly discouraged and is believed to be rare.
7097 Applications that parse structures containing CHOICE or ANY DEFINED BY
7098 components may be affected. Certificate parsing (d2i_X509 and related
7099 functions) are however not affected. OpenSSL clients and servers are
7105 * PKCS7 NULL pointer dereferences fix
7107 The PKCS#7 parsing code does not handle missing outer ContentInfo
7108 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7109 missing content and trigger a NULL pointer dereference on parsing.
7111 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7112 otherwise parse PKCS#7 structures from untrusted sources are
7113 affected. OpenSSL clients and servers are not affected.
7115 This issue was reported to OpenSSL by Michal Zalewski (Google).
7120 * DoS via reachable assert in SSLv2 servers fix
7122 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7123 servers that both support SSLv2 and enable export cipher suites by sending
7124 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7126 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7127 (OpenSSL development team).
7132 * Use After Free following d2i_ECPrivatekey error fix
7134 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7135 could cause a use after free condition. This, in turn, could cause a double
7136 free in several private key parsing functions (such as d2i_PrivateKey
7137 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7138 for applications that receive EC private keys from untrusted
7139 sources. This scenario is considered rare.
7141 This issue was discovered by the BoringSSL project and fixed in their
7147 * X509_to_X509_REQ NULL pointer deref fix
7149 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7150 the certificate key is invalid. This function is rarely used in practice.
7152 This issue was discovered by Brian Carpenter.
7157 * Removed the export ciphers from the DEFAULT ciphers
7161 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7163 * Build fixes for the Windows and OpenVMS platforms
7165 *Matt Caswell and Richard Levitte*
7167 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7169 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7170 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7171 dereference. This could lead to a Denial Of Service attack. Thanks to
7172 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7177 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7178 dtls1_buffer_record function under certain conditions. In particular this
7179 could occur if an attacker sent repeated DTLS records with the same
7180 sequence number but for the next epoch. The memory leak could be exploited
7181 by an attacker in a Denial of Service attack through memory exhaustion.
7182 Thanks to Chris Mueller for reporting this issue.
7187 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7188 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7189 method would be set to NULL which could later result in a NULL pointer
7190 dereference. Thanks to Frank Schmirler for reporting this issue.
7195 * Abort handshake if server key exchange message is omitted for ephemeral
7198 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7199 reporting this issue.
7204 * Remove non-export ephemeral RSA code on client and server. This code
7205 violated the TLS standard by allowing the use of temporary RSA keys in
7206 non-export ciphersuites and could be used by a server to effectively
7207 downgrade the RSA key length used to a value smaller than the server
7208 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7209 INRIA or reporting this issue.
7214 * Fixed issue where DH client certificates are accepted without verification.
7215 An OpenSSL server will accept a DH certificate for client authentication
7216 without the certificate verify message. This effectively allows a client to
7217 authenticate without the use of a private key. This only affects servers
7218 which trust a client certificate authority which issues certificates
7219 containing DH keys: these are extremely rare and hardly ever encountered.
7220 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7226 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7227 results on some platforms, including x86_64. This bug occurs at random
7228 with a very low probability, and is not known to be exploitable in any
7229 way, though its exact impact is difficult to determine. Thanks to Pieter
7230 Wuille (Blockstream) who reported this issue and also suggested an initial
7231 fix. Further analysis was conducted by the OpenSSL development team and
7232 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7233 the OpenSSL core team.
7238 * Fix various certificate fingerprint issues.
7240 By using non-DER or invalid encodings outside the signed portion of a
7241 certificate the fingerprint can be changed without breaking the signature.
7242 Although no details of the signed portion of the certificate can be changed
7243 this can cause problems with some applications: e.g. those using the
7244 certificate fingerprint for blacklists.
7246 1. Reject signatures with non zero unused bits.
7248 If the BIT STRING containing the signature has non zero unused bits reject
7249 the signature. All current signature algorithms require zero unused bits.
7251 2. Check certificate algorithm consistency.
7253 Check the AlgorithmIdentifier inside TBS matches the one in the
7254 certificate signature. NB: this will result in signature failure
7255 errors for some broken certificates.
7257 Thanks to Konrad Kraszewski from Google for reporting this issue.
7259 3. Check DSA/ECDSA signatures use DER.
7261 Reencode DSA/ECDSA signatures and compare with the original received
7262 signature. Return an error if there is a mismatch.
7264 This will reject various cases including garbage after signature
7265 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7266 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7267 (negative or with leading zeroes).
7269 Further analysis was conducted and fixes were developed by Stephen Henson
7270 of the OpenSSL core team.
7276 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7278 * Session Ticket Memory Leak.
7280 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7281 integrity of that ticket is first verified. In the event of a session
7282 ticket integrity check failing, OpenSSL will fail to free memory
7283 causing a memory leak. By sending a large number of invalid session
7284 tickets an attacker could exploit this issue in a Denial Of Service
7290 * Build option no-ssl3 is incomplete.
7292 When OpenSSL is configured with "no-ssl3" as a build option, servers
7293 could accept and complete a SSL 3.0 handshake, and clients could be
7294 configured to send them.
7297 *Akamai and the OpenSSL team*
7299 * Add support for TLS_FALLBACK_SCSV.
7300 Client applications doing fallback retries should call
7301 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7304 *Adam Langley, Bodo Moeller*
7306 * Add additional DigestInfo checks.
7308 Reencode DigestInto in DER and check against the original when
7309 verifying RSA signature: this will reject any improperly encoded
7310 DigestInfo structures.
7312 Note: this is a precautionary measure and no attacks are currently known.
7316 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7318 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7319 to a denial of service attack. A malicious server can crash the client
7320 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7321 ciphersuite and sending carefully crafted handshake messages.
7323 Thanks to Felix Gröbert (Google) for discovering and researching this
7329 * By sending carefully crafted DTLS packets an attacker could cause openssl
7330 to leak memory. This can be exploited through a Denial of Service attack.
7331 Thanks to Adam Langley for discovering and researching this issue.
7336 * An attacker can force openssl to consume large amounts of memory whilst
7337 processing DTLS handshake messages. This can be exploited through a
7338 Denial of Service attack.
7339 Thanks to Adam Langley for discovering and researching this issue.
7344 * An attacker can force an error condition which causes openssl to crash
7345 whilst processing DTLS packets due to memory being freed twice. This
7346 can be exploited through a Denial of Service attack.
7347 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7353 * If a multithreaded client connects to a malicious server using a resumed
7354 session and the server sends an ec point format extension it could write
7355 up to 255 bytes to freed memory.
7357 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7363 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7364 X509_name_oneline, X509_name_print_ex et al. to leak some information
7365 from the stack. Applications may be affected if they echo pretty printing
7366 output to the attacker.
7368 Thanks to Ivan Fratric (Google) for discovering this issue.
7371 *Emilia Käsper, and Steve Henson*
7373 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7374 for corner cases. (Certain input points at infinity could lead to
7375 bogus results, with non-infinity inputs mapped to infinity too.)
7379 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7381 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7382 handshake can force the use of weak keying material in OpenSSL
7383 SSL/TLS clients and servers.
7385 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7386 researching this issue. [CVE-2014-0224][]
7388 *KIKUCHI Masashi, Steve Henson*
7390 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7391 OpenSSL DTLS client the code can be made to recurse eventually crashing
7394 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7397 *Imre Rad, Steve Henson*
7399 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7400 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7401 client or server. This is potentially exploitable to run arbitrary
7402 code on a vulnerable client or server.
7404 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7406 *Jüri Aedla, Steve Henson*
7408 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7409 are subject to a denial of service attack.
7411 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7412 this issue. [CVE-2014-3470][]
7414 *Felix Gröbert, Ivan Fratric, Steve Henson*
7416 * Harmonize version and its documentation. -f flag is used to display
7419 *mancha <mancha1@zoho.com>*
7421 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7422 in i2d_ECPrivateKey.
7424 *mancha <mancha1@zoho.com>*
7426 * Fix some double frees. These are not thought to be exploitable.
7428 *mancha <mancha1@zoho.com>*
7430 * Fix for the attack described in the paper "Recovering OpenSSL
7431 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7432 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7433 <http://eprint.iacr.org/2014/140>
7435 Thanks to Yuval Yarom and Naomi Benger for discovering this
7436 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7438 *Yuval Yarom and Naomi Benger*
7440 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7442 * Keep original DTLS digest and encryption contexts in retransmission
7443 structures so we can use the previous session parameters if they need
7444 to be resent. [CVE-2013-6450][]
7448 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7449 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7450 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7451 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7452 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7453 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7455 *Rob Stradling, Adam Langley*
7457 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7459 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7461 This addresses the flaw in CBC record processing discovered by
7462 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7463 at: <http://www.isg.rhul.ac.uk/tls/>
7465 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7466 Security Group at Royal Holloway, University of London
7467 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7468 Emilia Käsper for the initial patch.
7471 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7473 * Return an error when checking OCSP signatures when key is NULL.
7474 This fixes a DoS attack. [CVE-2013-0166][]
7478 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7479 the right response is stapled. Also change SSL_get_certificate()
7480 so it returns the certificate actually sent.
7481 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7482 (This is a backport)
7484 *Rob Stradling <rob.stradling@comodo.com>*
7486 * Fix possible deadlock when decoding public keys.
7490 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7492 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7495 * Sanity check record length before skipping explicit IV in DTLS
7498 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7499 fuzzing as a service testing platform.
7504 * Initialise tkeylen properly when encrypting CMS messages.
7505 Thanks to Solar Designer of Openwall for reporting this issue.
7509 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7511 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7512 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7513 in CRYPTO_realloc_clean.
7515 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7516 issue and to Adam Langley <agl@chromium.org> for fixing it.
7519 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7521 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7523 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7524 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7525 content decryption and always return the same error. Note: this attack
7526 needs on average 2^20 messages so it only affects automated senders. The
7527 old behaviour can be re-enabled in the CMS code by setting the
7528 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7529 an MMA defence is not necessary.
7530 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7531 this issue. [CVE-2012-0884][]
7535 * Fix CVE-2011-4619: make sure we really are receiving a
7536 client hello before rejecting multiple SGC restarts. Thanks to
7537 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7541 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7543 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7544 Thanks to Antonio Martin, Enterprise Secure Access Research and
7545 Development, Cisco Systems, Inc. for discovering this bug and
7546 preparing a fix. [CVE-2012-0050][]
7550 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7552 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7553 of the Vaudenay padding oracle attack on CBC mode encryption
7554 which enables an efficient plaintext recovery attack against
7555 the OpenSSL implementation of DTLS. Their attack exploits timing
7556 differences arising during decryption processing. A research
7557 paper describing this attack can be found at:
7558 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7559 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7560 Security Group at Royal Holloway, University of London
7561 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7562 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7563 for preparing the fix. [CVE-2011-4108][]
7565 *Robin Seggelmann, Michael Tuexen*
7567 * Clear bytes used for block padding of SSL 3.0 records.
7570 *Adam Langley (Google)*
7572 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7573 Kadianakis <desnacked@gmail.com> for discovering this issue and
7574 Adam Langley for preparing the fix. [CVE-2011-4619][]
7576 *Adam Langley (Google)*
7578 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7580 *Andrey Kulikov <amdeich@gmail.com>*
7582 * Prevent malformed RFC3779 data triggering an assertion failure.
7583 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7584 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7586 *Rob Austein <sra@hactrn.net>*
7588 * Improved PRNG seeding for VOS.
7590 *Paul Green <Paul.Green@stratus.com>*
7592 * Fix ssl_ciph.c set-up race.
7594 *Adam Langley (Google)*
7596 * Fix spurious failures in ecdsatest.c.
7598 *Emilia Käsper (Google)*
7600 * Fix the BIO_f_buffer() implementation (which was mixing different
7601 interpretations of the `..._len` fields).
7603 *Adam Langley (Google)*
7605 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7606 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7607 threads won't reuse the same blinding coefficients.
7609 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7610 lock to call BN_BLINDING_invert_ex, and avoids one use of
7611 BN_BLINDING_update for each BN_BLINDING structure (previously,
7612 the last update always remained unused).
7614 *Emilia Käsper (Google)*
7616 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7618 *Bob Buckholz (Google)*
7620 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7622 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7623 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7625 *Kaspar Brand <ossl@velox.ch>*
7627 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7628 for multi-threaded use of ECDH. [CVE-2011-3210][]
7630 *Adam Langley (Google)*
7632 * Fix x509_name_ex_d2i memory leak on bad inputs.
7636 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7637 signature public key algorithm by using OID xref utilities instead.
7638 Before this you could only use some ECC ciphersuites with SHA1 only.
7642 * Add protection against ECDSA timing attacks as mentioned in the paper
7643 by Billy Bob Brumley and Nicola Tuveri, see:
7644 <http://eprint.iacr.org/2011/232.pdf>
7646 *Billy Bob Brumley and Nicola Tuveri*
7648 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7650 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7652 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7654 * Fix bug in string printing code: if *any* escaping is enabled we must
7655 escape the escape character (backslash) or the resulting string is
7660 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7662 * Disable code workaround for ancient and obsolete Netscape browsers
7663 and servers: an attacker can use it in a ciphersuite downgrade attack.
7664 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7668 * Fixed J-PAKE implementation error, originally discovered by
7669 Sebastien Martini, further info and confirmation from Stefan
7670 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7674 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7676 * Fix extension code to avoid race conditions which can result in a buffer
7677 overrun vulnerability: resumed sessions must not be modified as they can
7678 be shared by multiple threads. CVE-2010-3864
7682 * Fix WIN32 build system to correctly link an ENGINE directory into
7687 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7689 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7692 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7694 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7696 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7697 context. The operation can be customised via the ctrl mechanism in
7698 case ENGINEs want to include additional functionality.
7702 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7706 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7707 output hashes compatible with older versions of OpenSSL.
7709 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7711 * Fix compression algorithm handling: if resuming a session use the
7712 compression algorithm of the resumed session instead of determining
7713 it from client hello again. Don't allow server to change algorithm.
7717 * Add load_crls() function to apps tidying load_certs() too. Add option
7718 to verify utility to allow additional CRLs to be included.
7722 * Update OCSP request code to permit adding custom headers to the request:
7723 some responders need this.
7727 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7730 *Julia Lawall <julia@diku.dk>*
7732 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7733 needlessly dereferenced structures, used obsolete functions and
7734 didn't handle all updated verify codes correctly.
7738 * Disable MD2 in the default configuration.
7742 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7743 indicate the initial BIO being pushed or popped. This makes it possible
7744 to determine whether the BIO is the one explicitly called or as a result
7745 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7746 it handles reference counts correctly and doesn't zero out the I/O bio
7747 when it is not being explicitly popped. WARNING: applications which
7748 included workarounds for the old buggy behaviour will need to be modified
7749 or they could free up already freed BIOs.
7753 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7754 renaming to all platforms (within the 0.9.8 branch, this was
7755 done conditionally on Netware platforms to avoid a name clash).
7757 *Guenter <lists@gknw.net>*
7759 * Add ECDHE and PSK support to DTLS.
7761 *Michael Tuexen <tuexen@fh-muenster.de>*
7763 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7768 * Add "missing" function EVP_MD_flags() (without this the only way to
7769 retrieve a digest flags is by accessing the structure directly. Update
7770 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7771 or cipher is registered as in the "from" argument. Print out all
7772 registered digests in the dgst usage message instead of manually
7773 attempting to work them out.
7777 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7778 this allows the use of compression and extensions. Change default cipher
7779 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7780 by default unless an application cipher string requests it.
7784 * Alter match criteria in PKCS12_parse(). It used to try to use local
7785 key ids to find matching certificates and keys but some PKCS#12 files
7786 don't follow the (somewhat unwritten) rules and this strategy fails.
7787 Now just gather all certificates together and the first private key
7788 then look for the first certificate that matches the key.
7792 * Support use of registered digest and cipher names for dgst and cipher
7793 commands instead of having to add each one as a special case. So now
7800 openssl dgst -sha256 foo
7802 and this works for ENGINE based algorithms too.
7806 * Update Gost ENGINE to support parameter files.
7808 *Victor B. Wagner <vitus@cryptocom.ru>*
7810 * Support GeneralizedTime in ca utility.
7812 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7814 * Enhance the hash format used for certificate directory links. The new
7815 form uses the canonical encoding (meaning equivalent names will work
7816 even if they aren't identical) and uses SHA1 instead of MD5. This form
7817 is incompatible with the older format and as a result c_rehash should
7818 be used to rebuild symbolic links.
7822 * Make PKCS#8 the default write format for private keys, replacing the
7823 traditional format. This form is standardised, more secure and doesn't
7824 include an implicit MD5 dependency.
7828 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7829 committed to OpenSSL should pass this lot as a minimum.
7833 * Add session ticket override functionality for use by EAP-FAST.
7835 *Jouni Malinen <j@w1.fi>*
7837 * Modify HMAC functions to return a value. Since these can be implemented
7838 in an ENGINE errors can occur.
7842 * Type-checked OBJ_bsearch_ex.
7846 * Type-checked OBJ_bsearch. Also some constification necessitated
7847 by type-checking. Still to come: TXT_DB, bsearch(?),
7848 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7853 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7854 seconds to a tm structure directly, instead of going through OS
7855 specific date routines. This avoids any issues with OS routines such
7856 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7857 and X509_time_adj_ex() to cover the extended range. The existing
7858 X509_time_adj() is still usable and will no longer have any date issues.
7862 * Delta CRL support. New use deltas option which will attempt to locate
7863 and search any appropriate delta CRLs available.
7865 This work was sponsored by Google.
7869 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7870 code and add additional score elements. Validate alternate CRL paths
7871 as part of the CRL checking and indicate a new error "CRL path validation
7872 error" in this case. Applications wanting additional details can use
7873 the verify callback and check the new "parent" field. If this is not
7874 NULL CRL path validation is taking place. Existing applications won't
7875 see this because it requires extended CRL support which is off by
7878 This work was sponsored by Google.
7882 * Support for freshest CRL extension.
7884 This work was sponsored by Google.
7888 * Initial indirect CRL support. Currently only supported in the CRLs
7889 passed directly and not via lookup. Process certificate issuer
7890 CRL entry extension and lookup CRL entries by bother issuer name
7891 and serial number. Check and process CRL issuer entry in IDP extension.
7893 This work was sponsored by Google.
7897 * Add support for distinct certificate and CRL paths. The CRL issuer
7898 certificate is validated separately in this case. Only enabled if
7899 an extended CRL support flag is set: this flag will enable additional
7900 CRL functionality in future.
7902 This work was sponsored by Google.
7906 * Add support for policy mappings extension.
7908 This work was sponsored by Google.
7912 * Fixes to pathlength constraint, self issued certificate handling,
7913 policy processing to align with RFC3280 and PKITS tests.
7915 This work was sponsored by Google.
7919 * Support for name constraints certificate extension. DN, email, DNS
7920 and URI types are currently supported.
7922 This work was sponsored by Google.
7926 * To cater for systems that provide a pointer-based thread ID rather
7927 than numeric, deprecate the current numeric thread ID mechanism and
7928 replace it with a structure and associated callback type. This
7929 mechanism allows a numeric "hash" to be extracted from a thread ID in
7930 either case, and on platforms where pointers are larger than 'long',
7931 mixing is done to help ensure the numeric 'hash' is usable even if it
7932 can't be guaranteed unique. The default mechanism is to use "&errno"
7933 as a pointer-based thread ID to distinguish between threads.
7935 Applications that want to provide their own thread IDs should now use
7936 CRYPTO_THREADID_set_callback() to register a callback that will call
7937 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7939 Note that ERR_remove_state() is now deprecated, because it is tied
7940 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7941 to free the current thread's error state should be replaced by
7942 ERR_remove_thread_state(NULL).
7944 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7945 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7946 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7947 application was previously providing a numeric thread callback that
7948 was inappropriate for distinguishing threads, then uniqueness might
7949 have been obtained with &errno that happened immediately in the
7950 intermediate development versions of OpenSSL; this is no longer the
7951 case, the numeric thread callback will now override the automatic use
7954 *Geoff Thorpe, with help from Bodo Moeller*
7956 * Initial support for different CRL issuing certificates. This covers a
7957 simple case where the self issued certificates in the chain exist and
7958 the real CRL issuer is higher in the existing chain.
7960 This work was sponsored by Google.
7964 * Removed effectively defunct crypto/store from the build.
7968 * Revamp of STACK to provide stronger type-checking. Still to come:
7969 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7970 ASN1_STRING, CONF_VALUE.
7974 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7975 RAM on SSL connections. This option can save about 34k per idle SSL.
7979 * Revamp of LHASH to provide stronger type-checking. Still to come:
7980 STACK, TXT_DB, bsearch, qsort.
7984 * Initial support for Cryptographic Message Syntax (aka CMS) based
7985 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7986 support for data, signedData, compressedData, digestedData and
7987 encryptedData, envelopedData types included. Scripts to check against
7988 RFC4134 examples draft and interop and consistency checks of many
7989 content types and variants.
7993 * Add options to enc utility to support use of zlib compression BIO.
7997 * Extend mk1mf to support importing of options and assembly language
7998 files from Configure script, currently only included in VC-WIN32.
7999 The assembly language rules can now optionally generate the source
8000 files from the associated perl scripts.
8004 * Implement remaining functionality needed to support GOST ciphersuites.
8005 Interop testing has been performed using CryptoPro implementations.
8007 *Victor B. Wagner <vitus@cryptocom.ru>*
8009 * s390x assembler pack.
8013 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8018 * Implement Opaque PRF Input TLS extension as specified in
8019 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8020 official specification yet and no extension type assignment by
8021 IANA exists, this extension (for now) will have to be explicitly
8022 enabled when building OpenSSL by providing the extension number
8023 to use. For example, specify an option
8025 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8027 to the "config" or "Configure" script to enable the extension,
8028 assuming extension number 0x9527 (which is a completely arbitrary
8029 and unofficial assignment based on the MD5 hash of the Internet
8030 Draft). Note that by doing so, you potentially lose
8031 interoperability with other TLS implementations since these might
8032 be using the same extension number for other purposes.
8034 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8035 opaque PRF input value to use in the handshake. This will create
8036 an internal copy of the length-'len' string at 'src', and will
8037 return non-zero for success.
8039 To get more control and flexibility, provide a callback function
8042 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8043 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8047 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8050 Callback function 'cb' will be called in handshakes, and is
8051 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8052 Argument 'arg' is for application purposes (the value as given to
8053 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8054 be provided to the callback function). The callback function
8055 has to return non-zero to report success: usually 1 to use opaque
8056 PRF input just if possible, or 2 to enforce use of the opaque PRF
8057 input. In the latter case, the library will abort the handshake
8058 if opaque PRF input is not successfully negotiated.
8060 Arguments 'peerinput' and 'len' given to the callback function
8061 will always be NULL and 0 in the case of a client. A server will
8062 see the client's opaque PRF input through these variables if
8063 available (NULL and 0 otherwise). Note that if the server
8064 provides an opaque PRF input, the length must be the same as the
8065 length of the client's opaque PRF input.
8067 Note that the callback function will only be called when creating
8068 a new session (session resumption can resume whatever was
8069 previously negotiated), and will not be called in SSL 2.0
8070 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8071 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8072 for applications that need to enforce opaque PRF input.
8076 * Update ssl code to support digests other than SHA1+MD5 for handshake
8079 *Victor B. Wagner <vitus@cryptocom.ru>*
8081 * Add RFC4507 support to OpenSSL. This includes the corrections in
8082 RFC4507bis. The encrypted ticket format is an encrypted encoded
8083 SSL_SESSION structure, that way new session features are automatically
8086 If a client application caches session in an SSL_SESSION structure
8087 support is transparent because tickets are now stored in the encoded
8090 The SSL_CTX structure automatically generates keys for ticket
8091 protection in servers so again support should be possible
8092 with no application modification.
8094 If a client or server wishes to disable RFC4507 support then the option
8095 SSL_OP_NO_TICKET can be set.
8097 Add a TLS extension debugging callback to allow the contents of any client
8098 or server extensions to be examined.
8100 This work was sponsored by Google.
8104 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8105 OpenSSL should now compile cleanly on gcc 4.2
8107 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8109 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8110 support including streaming MAC support: this is required for GOST
8111 ciphersuite support.
8113 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8115 * Add option -stream to use PKCS#7 streaming in smime utility. New
8116 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8117 to output in BER and PEM format.
8121 * Experimental support for use of HMAC via EVP_PKEY interface. This
8122 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8123 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8124 ENGINE support for HMAC keys which are unextractable. New -mac and
8125 -macopt options to dgst utility.
8129 * New option -sigopt to dgst utility. Update dgst to use
8130 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8131 alternative signing parameters such as X9.31 or PSS in the dgst
8136 * Change ssl_cipher_apply_rule(), the internal function that does
8137 the work each time a ciphersuite string requests enabling
8138 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8139 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8140 the order of disabled ciphersuites such that those ciphersuites
8141 that most recently went from enabled to disabled not only stay
8142 in order with respect to each other, but also have higher priority
8143 than other disabled ciphersuites the next time ciphersuites are
8146 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8147 the same ciphersuites as with "HIGH" alone, but in a specific
8148 order where the PSK ciphersuites come first (since they are the
8149 most recently disabled ciphersuites when "HIGH" is parsed).
8151 Also, change ssl_create_cipher_list() (using this new
8152 functionality) such that between otherwise identical
8153 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8158 * Change ssl_create_cipher_list() so that it automatically
8159 arranges the ciphersuites in reasonable order before starting
8160 to process the rule string. Thus, the definition for "DEFAULT"
8161 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8162 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8163 This makes it much easier to arrive at a reasonable default order
8164 in applications for which anonymous ciphers are OK (meaning
8165 that you can't actually use DEFAULT).
8167 *Bodo Moeller; suggested by Victor Duchovni*
8169 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8170 processing) into multiple integers instead of setting
8171 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8172 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8173 (These masks as well as the individual bit definitions are hidden
8174 away into the non-exported interface ssl/ssl_locl.h, so this
8175 change to the definition of the SSL_CIPHER structure shouldn't
8176 affect applications.) This give us more bits for each of these
8177 categories, so there is no longer a need to coagulate AES128 and
8178 AES256 into a single algorithm bit, and to coagulate Camellia128
8179 and Camellia256 into a single algorithm bit, which has led to all
8182 Thus, among other things, the kludge introduced in 0.9.7m and
8183 0.9.8e for masking out AES256 independently of AES128 or masking
8184 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8186 With the change, we also introduce new ciphersuite aliases that
8187 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8192 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8193 Use the leftmost N bytes of the signature input if the input is
8194 larger than the prime q (with N being the size in bytes of q).
8198 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8199 it yet and it is largely untested.
8203 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8207 * Initial incomplete changes to avoid need for function casts in OpenSSL
8208 some compilers (gcc 4.2 and later) reject their use. Safestack is
8209 reimplemented. Update ASN1 to avoid use of legacy functions.
8213 * Win32/64 targets are linked with Winsock2.
8217 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8218 to external functions. This can be used to increase CRL handling
8219 efficiency especially when CRLs are very large by (for example) storing
8220 the CRL revoked certificates in a database.
8224 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8225 new CRLs added to a directory can be used. New command line option
8226 -verify_return_error to s_client and s_server. This causes real errors
8227 to be returned by the verify callback instead of carrying on no matter
8228 what. This reflects the way a "real world" verify callback would behave.
8232 * GOST engine, supporting several GOST algorithms and public key formats.
8233 Kindly donated by Cryptocom.
8237 * Partial support for Issuing Distribution Point CRL extension. CRLs
8238 partitioned by DP are handled but no indirect CRL or reason partitioning
8239 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8240 selected via a scoring technique which handles IDP and AKID in CRLs.
8244 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8245 will ultimately be used for all verify operations: this will remove the
8246 X509_STORE dependency on certificate verification and allow alternative
8247 lookup methods. X509_STORE based implementations of these two callbacks.
8251 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8252 Modify get_crl() to find a valid (unexpired) CRL if possible.
8256 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8257 this would be called X509_CRL_cmp() but that name is already used by
8258 a function that just compares CRL issuer names. Cache several CRL
8259 extensions in X509_CRL structure and cache CRLDP in X509.
8263 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8264 this maps equivalent X509_NAME structures into a consistent structure.
8265 Name comparison can then be performed rapidly using memcmp().
8269 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8274 * Allow digests to supply their own micalg string for S/MIME type using
8275 the ctrl EVP_MD_CTRL_MICALG.
8279 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8280 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8281 ctrl. It can then customise the structure before and/or after signing
8286 * New function OBJ_add_sigid() to allow application defined signature OIDs
8287 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8288 to free up any added signature OIDs.
8292 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8293 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8294 digest and cipher tables. New options added to openssl utility:
8295 list-message-digest-algorithms and list-cipher-algorithms.
8299 * Change the array representation of binary polynomials: the list
8300 of degrees of non-zero coefficients is now terminated with -1.
8301 Previously it was terminated with 0, which was also part of the
8302 value; thus, the array representation was not applicable to
8303 polynomials where t^0 has coefficient zero. This change makes
8304 the array representation useful in a more general context.
8308 * Various modifications and fixes to SSL/TLS cipher string
8309 handling. For ECC, the code now distinguishes between fixed ECDH
8310 with RSA certificates on the one hand and with ECDSA certificates
8311 on the other hand, since these are separate ciphersuites. The
8312 unused code for Fortezza ciphersuites has been removed.
8314 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8315 (not "ECDHE"). For consistency with the code for DH
8316 certificates, use of ECDH certificates is now considered ECDH
8317 authentication, not RSA or ECDSA authentication (the latter is
8318 merely the CA's signing algorithm and not actively used in the
8321 The temporary ciphersuite alias "ECCdraft" is no longer
8322 available, and ECC ciphersuites are no longer excluded from "ALL"
8323 and "DEFAULT". The following aliases now exist for RFC 4492
8324 ciphersuites, most of these by analogy with the DH case:
8326 kECDHr - ECDH cert, signed with RSA
8327 kECDHe - ECDH cert, signed with ECDSA
8328 kECDH - ECDH cert (signed with either RSA or ECDSA)
8329 kEECDH - ephemeral ECDH
8330 ECDH - ECDH cert or ephemeral ECDH
8336 AECDH - anonymous ECDH
8337 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8341 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8342 Use correct micalg parameters depending on digest(s) in signed message.
8346 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8347 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8351 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8352 an engine to register a method. Add ENGINE lookups for methods and
8353 functional reference processing.
8357 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8358 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8363 * New -resign option to smime utility. This adds one or more signers
8364 to an existing PKCS#7 signedData structure. Also -md option to use an
8365 alternative message digest algorithm for signing.
8369 * Tidy up PKCS#7 routines and add new functions to make it easier to
8370 create PKCS7 structures containing multiple signers. Update smime
8371 application to support multiple signers.
8375 * New -macalg option to pkcs12 utility to allow setting of an alternative
8380 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8381 Reorganize PBE internals to lookup from a static table using NIDs,
8382 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8383 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8384 PRF which will be automatically used with PBES2.
8388 * Replace the algorithm specific calls to generate keys in "req" with the
8393 * Update PKCS#7 enveloped data routines to use new API. This is now
8394 supported by any public key method supporting the encrypt operation. A
8395 ctrl is added to allow the public key algorithm to examine or modify
8396 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8401 * Add a ctrl to asn1 method to allow a public key algorithm to express
8402 a default digest type to use. In most cases this will be SHA1 but some
8403 algorithms (such as GOST) need to specify an alternative digest. The
8404 return value indicates how strong the preference is 1 means optional and
8405 2 is mandatory (that is it is the only supported type). Modify
8406 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8407 use the default md. Update openssl utilities to use the default digest
8408 type for signing if it is not explicitly indicated.
8412 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8413 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8414 signing method from the key type. This effectively removes the link
8415 between digests and public key types.
8419 * Add an OID cross reference table and utility functions. Its purpose is to
8420 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8421 rsaEncryption. This will allow some of the algorithm specific hackery
8422 needed to use the correct OID to be removed.
8426 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8427 structures for PKCS7_sign(). They are now set up by the relevant public
8432 * Add provisional EC pkey method with support for ECDSA and ECDH.
8436 * Add support for key derivation (agreement) in the API, DH method and
8441 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8442 public and private key formats. As a side effect these add additional
8443 command line functionality not previously available: DSA signatures can be
8444 generated and verified using pkeyutl and DH key support and generation in
8451 *Oliver Tappe <zooey@hirschkaefer.de>*
8453 * New make target "install_html_docs" installs HTML renditions of the
8456 *Oliver Tappe <zooey@hirschkaefer.de>*
8458 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8459 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8460 support key and parameter generation and add initial key generation
8461 functionality for RSA.
8465 * Add functions for main EVP_PKEY_method operations. The undocumented
8466 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8467 `EVP_PKEY_{encrypt,decrypt}_old`.
8471 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8472 key API, doesn't do much yet.
8476 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8477 public key algorithms. New option to openssl utility:
8478 "list-public-key-algorithms" to print out info.
8482 * Implement the Supported Elliptic Curves Extension for
8483 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8487 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8488 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8492 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8493 utilities such as rsa, dsa, dsaparam etc except they process any key
8498 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8499 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8500 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8505 * Initial support for pluggable public key ASN1.
8506 De-spaghettify the public key ASN1 handling. Move public and private
8507 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8508 algorithm specific handling to a single module within the relevant
8509 algorithm directory. Add functions to allow (near) opaque processing
8510 of public and private key structures.
8514 * Implement the Supported Point Formats Extension for
8515 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8519 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8520 for the psk identity [hint] and the psk callback functions to the
8521 SSL_SESSION, SSL and SSL_CTX structure.
8524 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8528 SSL_CTX_use_psk_identity_hint
8529 SSL_get_psk_identity_hint
8530 SSL_get_psk_identity
8531 SSL_use_psk_identity_hint
8533 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8535 * Add RFC 3161 compliant time stamp request creation, response generation
8536 and response verification functionality.
8538 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8540 * Add initial support for TLS extensions, specifically for the server_name
8541 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8542 have new members for a host name. The SSL data structure has an
8543 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8544 stored in that context to allow for session resumption, even after the
8545 SSL has been switched to a new SSL_CTX in reaction to a client's
8546 server_name extension.
8548 New functions (subject to change):
8550 SSL_get_servername()
8551 SSL_get_servername_type()
8554 New CTRL codes and macros (subject to change):
8556 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8557 - SSL_CTX_set_tlsext_servername_callback()
8558 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8559 - SSL_CTX_set_tlsext_servername_arg()
8560 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8562 openssl s_client has a new '-servername ...' option.
8564 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8565 '-key2 ...', '-servername_fatal' (subject to change). This allows
8566 testing the HostName extension for a specific single host name ('-cert'
8567 and '-key' remain fallbacks for handshakes without HostName
8568 negotiation). If the unrecognized_name alert has to be sent, this by
8569 default is a warning; it becomes fatal with the '-servername_fatal'
8572 *Peter Sylvester, Remy Allais, Christophe Renou*
8574 * Whirlpool hash implementation is added.
8578 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8579 bn(64,32). Because of instruction set limitations it doesn't have
8580 any negative impact on performance. This was done mostly in order
8581 to make it possible to share assembler modules, such as bn_mul_mont
8582 implementations, between 32- and 64-bit builds without hassle.
8586 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8587 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8592 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8593 dedicated Montgomery multiplication procedure, is introduced.
8594 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8595 "64-bit" performance on certain 32-bit targets.
8599 * New option SSL_OP_NO_COMP to disable use of compression selectively
8600 in SSL structures. New SSL ctrl to set maximum send fragment size.
8601 Save memory by setting the I/O buffer sizes dynamically instead of
8602 using the maximum available value.
8606 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8607 in addition to the text details.
8611 * Very, very preliminary EXPERIMENTAL support for printing of general
8612 ASN1 structures. This currently produces rather ugly output and doesn't
8613 handle several customised structures at all.
8617 * Integrated support for PVK file format and some related formats such
8618 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8619 these in the 'rsa' and 'dsa' utilities.
8623 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8627 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8628 place for the (very old) "NETSCAPE" format certificates which are now
8629 handled using new ASN1 code equivalents.
8633 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8634 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8635 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8639 * Modify CRL distribution points extension code to print out previously
8640 unsupported fields. Enhance extension setting code to allow setting of
8645 * Add print and set support for Issuing Distribution Point CRL extension.
8649 * Change 'Configure' script to enable Camellia by default.
8656 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8658 * When rejecting SSL/TLS records due to an incorrect version number, never
8659 update s->server with a new major version number. As of
8660 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8661 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8662 the previous behavior could result in a read attempt at NULL when
8663 receiving specific incorrect SSL/TLS records once record payload
8664 protection is active. [CVE-2010-0740][]
8666 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8668 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8669 could be crashed if the relevant tables were not present (e.g. chrooted).
8671 *Tomas Hoger <thoger@redhat.com>*
8673 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8675 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8677 *Martin Olsson, Neel Mehta*
8679 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8680 accommodate for stack sorting, always a write lock!).
8684 * On some versions of WIN32 Heap32Next is very slow. This can cause
8685 excessive delays in the RAND_poll(): over a minute. As a workaround
8686 include a time check in the inner Heap32Next loop too.
8690 * The code that handled flushing of data in SSL/TLS originally used the
8691 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8692 the problem outlined in PR#1949. The fix suggested there however can
8693 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8694 of Apache). So instead simplify the code to flush unconditionally.
8695 This should be fine since flushing with no data to flush is a no op.
8699 * Handle TLS versions 2.0 and later properly and correctly use the
8700 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8701 off ancient servers have a habit of sticking around for a while...
8705 * Modify compression code so it frees up structures without using the
8706 ex_data callbacks. This works around a problem where some applications
8707 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8708 restarting) then use compression (e.g. SSL with compression) later.
8709 This results in significant per-connection memory leaks and
8710 has caused some security issues including CVE-2008-1678 and
8715 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8716 change when encrypting or decrypting.
8720 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8721 connect and renegotiate with servers which do not support RI.
8722 Until RI is more widely deployed this option is enabled by default.
8726 * Add "missing" ssl ctrls to clear options and mode.
8730 * If client attempts to renegotiate and doesn't support RI respond with
8731 a no_renegotiation alert as required by RFC5746. Some renegotiating
8732 TLS clients will continue a connection gracefully when they receive
8733 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8734 waiting for a server hello which it will never receive. Now we treat a
8735 received no_renegotiation alert as a fatal error. This is because
8736 applications requesting a renegotiation might well expect it to succeed
8737 and would have no code in place to handle the server denying it so the
8738 only safe thing to do is to terminate the connection.
8742 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8743 peer supports secure renegotiation and 0 otherwise. Print out peer
8744 renegotiation support in s_client/s_server.
8748 * Replace the highly broken and deprecated SPKAC certification method with
8749 the updated NID creation version. This should correctly handle UTF8.
8753 * Implement RFC5746. Re-enable renegotiation but require the extension
8754 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8755 turns out to be a bad idea. It has been replaced by
8756 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8757 SSL_CTX_set_options(). This is really not recommended unless you
8758 know what you are doing.
8760 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8762 * Fixes to stateless session resumption handling. Use initial_ctx when
8763 issuing and attempting to decrypt tickets in case it has changed during
8764 servername handling. Use a non-zero length session ID when attempting
8765 stateless session resumption: this makes it possible to determine if
8766 a resumption has occurred immediately after receiving server hello
8767 (several places in OpenSSL subtly assume this) instead of later in
8772 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8773 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8774 fixes for a few places where the return code is not checked
8777 *Julia Lawall <julia@diku.dk>*
8779 * Add --strict-warnings option to Configure script to include devteam
8780 warnings in other configurations.
8784 * Add support for --libdir option and LIBDIR variable in makefiles. This
8785 makes it possible to install openssl libraries in locations which
8786 have names other than "lib", for example "/usr/lib64" which some
8789 *Steve Henson, based on patch from Jeremy Utley*
8791 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8792 X690 8.9.12 and can produce some misleading textual output of OIDs.
8794 *Steve Henson, reported by Dan Kaminsky*
8796 * Delete MD2 from algorithm tables. This follows the recommendation in
8797 several standards that it is not used in new applications due to
8798 several cryptographic weaknesses. For binary compatibility reasons
8799 the MD2 API is still compiled in by default.
8803 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8808 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8809 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8812 *Guenter <lists@gknw.net>*
8814 * Fix the server certificate chain building code to use X509_verify_cert(),
8815 it used to have an ad-hoc builder which was unable to cope with anything
8816 other than a simple chain.
8818 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8820 * Don't check self signed certificate signatures in X509_verify_cert()
8821 by default (a flag can override this): it just wastes time without
8822 adding any security. As a useful side effect self signed root CAs
8823 with non-FIPS digests are now usable in FIPS mode.
8827 * In dtls1_process_out_of_seq_message() the check if the current message
8828 is already buffered was missing. For every new message was memory
8829 allocated, allowing an attacker to perform an denial of service attack
8830 with sending out of seq handshake messages until there is no memory
8831 left. Additionally every future message was buffered, even if the
8832 sequence number made no sense and would be part of another handshake.
8833 So only messages with sequence numbers less than 10 in advance will be
8834 buffered. [CVE-2009-1378][]
8836 *Robin Seggelmann, discovered by Daniel Mentz*
8838 * Records are buffered if they arrive with a future epoch to be
8839 processed after finishing the corresponding handshake. There is
8840 currently no limitation to this buffer allowing an attacker to perform
8841 a DOS attack with sending records with future epochs until there is no
8842 memory left. This patch adds the pqueue_size() function to determine
8843 the size of a buffer and limits the record buffer to 100 entries.
8846 *Robin Seggelmann, discovered by Daniel Mentz*
8848 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8849 parent structure is freed. [CVE-2009-1379][]
8853 * Handle non-blocking I/O properly in SSL_shutdown() call.
8855 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8857 * Add `2.5.4.*` OIDs
8859 *Ilya O. <vrghost@gmail.com>*
8861 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8863 * Disable renegotiation completely - this fixes a severe security
8864 problem [CVE-2009-3555][] at the cost of breaking all
8865 renegotiation. Renegotiation can be re-enabled by setting
8866 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8867 run-time. This is really not recommended unless you know what
8872 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8874 * Don't set val to NULL when freeing up structures, it is freed up by
8875 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8876 zeroing past the valid field. [CVE-2009-0789][]
8878 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8880 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8881 checked correctly. This would allow some invalid signed attributes to
8882 appear to verify correctly. [CVE-2009-0591][]
8884 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8886 * Reject UniversalString and BMPString types with invalid lengths. This
8887 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8888 a legal length. [CVE-2009-0590][]
8892 * Set S/MIME signing as the default purpose rather than setting it
8893 unconditionally. This allows applications to override it at the store
8898 * Permit restricted recursion of ASN1 strings. This is needed in practice
8899 to handle some structures.
8903 * Improve efficiency of mem_gets: don't search whole buffer each time
8906 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8908 * New -hex option for openssl rand.
8912 * Print out UTF8String and NumericString when parsing ASN1.
8916 * Support NumericString type for name components.
8920 * Allow CC in the environment to override the automatically chosen
8921 compiler. Note that nothing is done to ensure flags work with the
8926 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8928 * Properly check EVP_VerifyFinal() and similar return values
8931 *Ben Laurie, Bodo Moeller, Google Security Team*
8933 * Enable TLS extensions by default.
8937 * Allow the CHIL engine to be loaded, whether the application is
8938 multithreaded or not. (This does not release the developer from the
8939 obligation to set up the dynamic locking callbacks.)
8941 *Sander Temme <sander@temme.net>*
8943 * Use correct exit code if there is an error in dgst command.
8945 *Steve Henson; problem pointed out by Roland Dirlewanger*
8947 * Tweak Configure so that you need to say "experimental-jpake" to enable
8948 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8952 * Add experimental JPAKE support, including demo authentication in
8953 s_client and s_server.
8957 * Set the comparison function in v3_addr_canonize().
8959 *Rob Austein <sra@hactrn.net>*
8961 * Add support for XMPP STARTTLS in s_client.
8963 *Philip Paeps <philip@freebsd.org>*
8965 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8966 to ensure that even with this option, only ciphersuites in the
8967 server's preference list will be accepted. (Note that the option
8968 applies only when resuming a session, so the earlier behavior was
8969 just about the algorithm choice for symmetric cryptography.)
8973 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8975 * Fix NULL pointer dereference if a DTLS server received
8976 ChangeCipherSpec as first record [CVE-2009-1386][].
8980 * Fix a state transition in s3_srvr.c and d1_srvr.c
8981 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8985 * The fix in 0.9.8c that supposedly got rid of unsafe
8986 double-checked locking was incomplete for RSA blinding,
8987 addressing just one layer of what turns out to have been
8988 doubly unsafe triple-checked locking.
8990 So now fix this for real by retiring the MONT_HELPER macro
8991 in crypto/rsa/rsa_eay.c.
8993 *Bodo Moeller; problem pointed out by Marius Schilder*
8995 * Various precautionary measures:
8997 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8999 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9000 (NB: This would require knowledge of the secret session ticket key
9001 to exploit, in which case you'd be SOL either way.)
9003 - Change bn_nist.c so that it will properly handle input BIGNUMs
9004 outside the expected range.
9006 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9009 *Neel Mehta, Bodo Moeller*
9011 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9012 the load fails. Useful for distros.
9014 *Ben Laurie and the FreeBSD team*
9016 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9020 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9024 * Expand ENGINE to support engine supplied SSL client certificate functions.
9026 This work was sponsored by Logica.
9030 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9031 keystores. Support for SSL/TLS client authentication too.
9032 Not compiled unless enable-capieng specified to Configure.
9034 This work was sponsored by Logica.
9038 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9039 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9040 attribute creation routines such as certificate requests and PKCS#12
9045 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9047 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9048 handshake which could lead to a client crash as found using the
9049 Codenomicon TLS test suite [CVE-2008-1672][]
9051 *Steve Henson, Mark Cox*
9053 * Fix double free in TLS server name extensions which could lead to
9054 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
9058 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9060 Clear the error queue to ensure that error entries left from
9061 older function calls do not interfere with the correct operation.
9063 *Lutz Jaenicke, Erik de Castro Lopo*
9065 * Remove root CA certificates of commercial CAs:
9067 The OpenSSL project does not recommend any specific CA and does not
9068 have any policy with respect to including or excluding any CA.
9069 Therefore it does not make any sense to ship an arbitrary selection
9070 of root CA certificates with the OpenSSL software.
9074 * RSA OAEP patches to fix two separate invalid memory reads.
9075 The first one involves inputs when 'lzero' is greater than
9076 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9077 before the beginning of from). The second one involves inputs where
9078 the 'db' section contains nothing but zeroes (there is a one-byte
9079 invalid read after the end of 'db').
9081 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9083 * Partial backport from 0.9.9-dev:
9085 Introduce bn_mul_mont (dedicated Montgomery multiplication
9086 procedure) as a candidate for BIGNUM assembler implementation.
9087 While 0.9.9-dev uses assembler for various architectures, only
9088 x86_64 is available by default here in the 0.9.8 branch, and
9089 32-bit x86 is available through a compile-time setting.
9091 To try the 32-bit x86 assembler implementation, use Configure
9092 option "enable-montasm" (which exists only for this backport).
9094 As "enable-montasm" for 32-bit x86 disclaims code stability
9095 anyway, in this constellation we activate additional code
9096 backported from 0.9.9-dev for further performance improvements,
9097 namely BN_from_montgomery_word. (To enable this otherwise,
9098 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9100 *Andy Polyakov (backport partially by Bodo Moeller)*
9102 * Add TLS session ticket callback. This allows an application to set
9103 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9104 values. This is useful for key rollover for example where several key
9105 sets may exist with different names.
9109 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9110 This was broken until now in 0.9.8 releases, such that the only way
9111 a registered ENGINE could be used (assuming it initialises
9112 successfully on the host) was to explicitly set it as the default
9113 for the relevant algorithms. This is in contradiction with 0.9.7
9114 behaviour and the documentation. With this fix, when an ENGINE is
9115 registered into a given algorithm's table of implementations, the
9116 'uptodate' flag is reset so that auto-discovery will be used next
9117 time a new context for that algorithm attempts to select an
9120 *Ian Lister (tweaked by Geoff Thorpe)*
9122 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9123 implementation in the following ways:
9125 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9128 Lack of BER streaming support means one pass streaming processing is
9129 only supported if data is detached: setting the streaming flag is
9130 ignored for embedded content.
9132 CMS support is disabled by default and must be explicitly enabled
9133 with the enable-cms configuration option.
9137 * Update the GMP engine glue to do direct copies between BIGNUM and
9138 mpz_t when openssl and GMP use the same limb size. Otherwise the
9139 existing "conversion via a text string export" trick is still used.
9141 *Paul Sheer <paulsheer@gmail.com>*
9143 * Zlib compression BIO. This is a filter BIO which compressed and
9144 uncompresses any data passed through it.
9148 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9149 RFC3394 compatible AES key wrapping.
9153 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9154 sets string data without copying. X509_ALGOR_set0() and
9155 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9156 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9157 from an X509_ATTRIBUTE structure optionally checking it occurs only
9158 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9163 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9164 to get the expected BN_FLG_CONSTTIME behavior.
9166 *Bodo Moeller (Google)*
9170 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9171 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9172 - added some more tests to do_tests.pl
9173 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9174 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9175 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9176 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9177 - various changes to netware.pl to enable gcc-cross builds on Win32
9179 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9180 - various changes to fix missing prototype warnings
9181 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9182 - added AES, WHIRLPOOL and CPUID assembler code to build files
9183 - added missing AES assembler make rules to mk1mf.pl
9184 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9186 *Guenter Knauf <eflash@gmx.net>*
9188 * Implement certificate status request TLS extension defined in RFC3546.
9189 A client can set the appropriate parameters and receive the encoded
9190 OCSP response via a callback. A server can query the supplied parameters
9191 and set the encoded OCSP response in the callback. Add simplified examples
9192 to s_client and s_server.
9196 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9199 + Binary incompatibility of ssl_ctx_st structure
9200 + DTLS interoperation with non-compliant servers
9201 + Don't call get_session_cb() without proposed session
9202 + Fix ia64 assembler code
9204 *Andy Polyakov, Steve Henson*
9206 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9208 * DTLS Handshake overhaul. There were longstanding issues with
9209 OpenSSL DTLS implementation, which were making it impossible for
9210 RFC 4347 compliant client to communicate with OpenSSL server.
9211 Unfortunately just fixing these incompatibilities would "cut off"
9212 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9213 server keeps tolerating non RFC compliant syntax. The opposite is
9214 not true, 0.9.8f client can not communicate with earlier server.
9215 This update even addresses CVE-2007-4995.
9219 * Changes to avoid need for function casts in OpenSSL: some compilers
9220 (gcc 4.2 and later) reject their use.
9221 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9224 * Add RFC4507 support to OpenSSL. This includes the corrections in
9225 RFC4507bis. The encrypted ticket format is an encrypted encoded
9226 SSL_SESSION structure, that way new session features are automatically
9229 If a client application caches session in an SSL_SESSION structure
9230 support is transparent because tickets are now stored in the encoded
9233 The SSL_CTX structure automatically generates keys for ticket
9234 protection in servers so again support should be possible
9235 with no application modification.
9237 If a client or server wishes to disable RFC4507 support then the option
9238 SSL_OP_NO_TICKET can be set.
9240 Add a TLS extension debugging callback to allow the contents of any client
9241 or server extensions to be examined.
9243 This work was sponsored by Google.
9247 * Add initial support for TLS extensions, specifically for the server_name
9248 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9249 have new members for a host name. The SSL data structure has an
9250 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9251 stored in that context to allow for session resumption, even after the
9252 SSL has been switched to a new SSL_CTX in reaction to a client's
9253 server_name extension.
9255 New functions (subject to change):
9257 SSL_get_servername()
9258 SSL_get_servername_type()
9261 New CTRL codes and macros (subject to change):
9263 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9264 - SSL_CTX_set_tlsext_servername_callback()
9265 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9266 - SSL_CTX_set_tlsext_servername_arg()
9267 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9269 openssl s_client has a new '-servername ...' option.
9271 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9272 '-key2 ...', '-servername_fatal' (subject to change). This allows
9273 testing the HostName extension for a specific single host name ('-cert'
9274 and '-key' remain fallbacks for handshakes without HostName
9275 negotiation). If the unrecognized_name alert has to be sent, this by
9276 default is a warning; it becomes fatal with the '-servername_fatal'
9279 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9281 * Add AES and SSE2 assembly language support to VC++ build.
9285 * Mitigate attack on final subtraction in Montgomery reduction.
9289 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9290 (which previously caused an internal error).
9294 * Squeeze another 10% out of IGE mode when in != out.
9298 * AES IGE mode speedup.
9300 *Dean Gaudet (Google)*
9302 * Add the Korean symmetric 128-bit cipher SEED (see
9303 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9304 add SEED ciphersuites from RFC 4162:
9306 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9307 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9308 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9309 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9311 To minimize changes between patchlevels in the OpenSSL 0.9.8
9312 series, SEED remains excluded from compilation unless OpenSSL
9313 is configured with 'enable-seed'.
9315 *KISA, Bodo Moeller*
9317 * Mitigate branch prediction attacks, which can be practical if a
9318 single processor is shared, allowing a spy process to extract
9319 information. For detailed background information, see
9320 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9321 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9322 and Necessary Software Countermeasures"). The core of the change
9323 are new versions BN_div_no_branch() and
9324 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9325 respectively, which are slower, but avoid the security-relevant
9326 conditional branches. These are automatically called by BN_div()
9327 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9328 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9329 remove a conditional branch.
9331 BN_FLG_CONSTTIME is the new name for the previous
9332 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9333 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9334 in the exponent causes BN_mod_exp_mont() to use the alternative
9335 implementation in BN_mod_exp_mont_consttime().) The old name
9336 remains as a deprecated alias.
9338 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9339 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9340 constant-time implementations for more than just exponentiation.
9341 Here too the old name is kept as a deprecated alias.
9343 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9344 the BN_BLINDING structure gets an independent copy of the
9345 modulus. This means that the previous `BIGNUM *m` argument to
9346 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9347 essentially becomes `const BIGNUM *m`, although we can't actually
9348 change this in the header file before 0.9.9. It allows
9349 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9350 enable BN_FLG_CONSTTIME.
9352 *Matthew D Wood (Intel Corp)*
9354 * In the SSL/TLS server implementation, be strict about session ID
9355 context matching (which matters if an application uses a single
9356 external cache for different purposes). Previously,
9357 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9358 set. This did ensure strict client verification, but meant that,
9359 with applications using a single external cache for quite
9360 different requirements, clients could circumvent ciphersuite
9361 restrictions for a given session ID context by starting a session
9362 in a different context.
9366 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9367 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9368 authentication-only ciphersuites.
9372 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9373 not complete and could lead to a possible single byte overflow
9374 [CVE-2007-5135][] [Ben Laurie]
9376 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9378 * Since AES128 and AES256 (and similarly Camellia128 and
9379 Camellia256) share a single mask bit in the logic of
9380 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9381 kludge to work properly if AES128 is available and AES256 isn't
9382 (or if Camellia128 is available and Camellia256 isn't).
9386 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9387 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9388 When a point or a seed is encoded in a BIT STRING, we need to
9389 prevent the removal of trailing zero bits to get the proper DER
9390 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9391 of a NamedBitList, for which trailing 0 bits need to be removed.)
9395 * Have SSL/TLS server implementation tolerate "mismatched" record
9396 protocol version while receiving ClientHello even if the
9397 ClientHello is fragmented. (The server can't insist on the
9398 particular protocol version it has chosen before the ServerHello
9399 message has informed the client about his choice.)
9403 * Add RFC 3779 support.
9405 *Rob Austein for ARIN, Ben Laurie*
9407 * Load error codes if they are not already present instead of using a
9408 static variable. This allows them to be cleanly unloaded and reloaded.
9409 Improve header file function name parsing.
9413 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9414 or CAPABILITY handshake as required by RFCs.
9418 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9420 * Introduce limits to prevent malicious keys being able to
9421 cause a denial of service. [CVE-2006-2940][]
9423 *Steve Henson, Bodo Moeller*
9425 * Fix ASN.1 parsing of certain invalid structures that can result
9426 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9428 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9429 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9431 * Fix SSL client code which could crash if connecting to a
9432 malicious SSLv2 server. [CVE-2006-4343][]
9434 *Tavis Ormandy and Will Drewry, Google Security Team*
9436 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9437 match only those. Before that, "AES256-SHA" would be interpreted
9438 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9439 the same strength classification in 0.9.7h) as we currently only
9440 have a single AES bit in the ciphersuite description bitmap.
9441 That change, however, also applied to ciphersuite strings such as
9442 "RC4-MD5" that intentionally matched multiple ciphersuites --
9443 namely, SSL 2.0 ciphersuites in addition to the more common ones
9444 from SSL 3.0/TLS 1.0.
9446 So we change the selection algorithm again: Naming an explicit
9447 ciphersuite selects this one ciphersuite, and any other similar
9448 ciphersuite (same bitmap) from *other* protocol versions.
9449 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9450 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9452 Since SSL 2.0 does not have any ciphersuites for which the
9453 128/256 bit distinction would be relevant, this works for now.
9454 The proper fix will be to use different bits for AES128 and
9455 AES256, which would have avoided the problems from the beginning;
9456 however, bits are scarce, so we can only do this in a new release
9457 (not just a patchlevel) when we can change the SSL_CIPHER
9458 definition to split the single 'unsigned long mask' bitmap into
9459 multiple values to extend the available space.
9463 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9465 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9466 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9468 * Add AES IGE and biIGE modes.
9472 * Change the Unix randomness entropy gathering to use poll() when
9473 possible instead of select(), since the latter has some
9474 undesirable limitations.
9476 *Darryl Miles via Richard Levitte and Bodo Moeller*
9478 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9479 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9480 cannot be implicitly activated as part of, e.g., the "AES" alias.
9481 However, please upgrade to OpenSSL 0.9.9[-dev] for
9482 non-experimental use of the ECC ciphersuites to get TLS extension
9483 support, which is required for curve and point format negotiation
9484 to avoid potential handshake problems.
9488 * Disable rogue ciphersuites:
9490 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9491 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9492 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9494 The latter two were purportedly from
9495 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9498 Also deactivate the remaining ciphersuites from
9499 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9500 unofficial, and the ID has long expired.
9504 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9505 dual-core machines) and other potential thread-safety issues.
9509 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9510 versions), which is now available for royalty-free use
9511 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9512 Also, add Camellia TLS ciphersuites from RFC 4132.
9514 To minimize changes between patchlevels in the OpenSSL 0.9.8
9515 series, Camellia remains excluded from compilation unless OpenSSL
9516 is configured with 'enable-camellia'.
9520 * Disable the padding bug check when compression is in use. The padding
9521 bug check assumes the first packet is of even length, this is not
9522 necessarily true if compression is enabled and can result in false
9523 positives causing handshake failure. The actual bug test is ancient
9524 code so it is hoped that implementations will either have fixed it by
9525 now or any which still have the bug do not support compression.
9529 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9531 * When applying a cipher rule check to see if string match is an explicit
9532 cipher suite and only match that one cipher suite if it is.
9536 * Link in manifests for VC++ if needed.
9538 *Austin Ziegler <halostatue@gmail.com>*
9540 * Update support for ECC-based TLS ciphersuites according to
9541 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9542 TLS extensions, which are supported starting with the 0.9.9
9543 branch, not in the OpenSSL 0.9.8 branch).
9547 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9548 opaque EVP_CIPHER_CTX handling.
9552 * Fixes and enhancements to zlib compression code. We now only use
9553 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9554 to conform with the standards mentioned here:
9555 <http://www.zlib.net/DLL_FAQ.txt>
9556 Static zlib linking now works on Windows and the new --with-zlib-include
9557 --with-zlib-lib options to Configure can be used to supply the location
9558 of the headers and library. Gracefully handle case where zlib library
9563 * Several fixes and enhancements to the OID generation code. The old code
9564 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9565 handle numbers larger than ULONG_MAX, truncated printing and had a
9566 non standard OBJ_obj2txt() behaviour.
9570 * Add support for building of engines under engine/ as shared libraries
9571 under VC++ build system.
9575 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9576 Hopefully, we will not see any false combination of paths any more.
9580 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9582 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9583 (part of SSL_OP_ALL). This option used to disable the
9584 countermeasure against man-in-the-middle protocol-version
9585 rollback in the SSL 2.0 server implementation, which is a bad
9586 idea. [CVE-2005-2969][]
9588 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9589 for Information Security, National Institute of Advanced Industrial
9590 Science and Technology [AIST], Japan)*
9592 * Add two function to clear and return the verify parameter flags.
9596 * Keep cipherlists sorted in the source instead of sorting them at
9597 runtime, thus removing the need for a lock.
9601 * Avoid some small subgroup attacks in Diffie-Hellman.
9603 *Nick Mathewson and Ben Laurie*
9605 * Add functions for well-known primes.
9609 * Extended Windows CE support.
9611 *Satoshi Nakamura and Andy Polyakov*
9613 * Initialize SSL_METHOD structures at compile time instead of during
9614 runtime, thus removing the need for a lock.
9618 * Make PKCS7_decrypt() work even if no certificate is supplied by
9619 attempting to decrypt each encrypted key in turn. Add support to
9624 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9626 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9629 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9633 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9634 key into the same file any more.
9638 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9642 * Add -utf8 command line and config file option to 'ca'.
9644 *Stefan <stf@udoma.org*
9646 * Removed the macro des_crypt(), as it seems to conflict with some
9647 libraries. Use DES_crypt().
9651 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9652 involves renaming the source and generated shared-libs for
9653 both. The engines will accept the corrected or legacy ids
9654 ('ncipher' and '4758_cca' respectively) when binding. NB,
9655 this only applies when building 'shared'.
9657 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9659 * Add attribute functions to EVP_PKEY structure. Modify
9660 PKCS12_create() to recognize a CSP name attribute and
9661 use it. Make -CSP option work again in pkcs12 utility.
9665 * Add new functionality to the bn blinding code:
9666 - automatic re-creation of the BN_BLINDING parameters after
9667 a fixed number of uses (currently 32)
9668 - add new function for parameter creation
9669 - introduce flags to control the update behaviour of the
9670 BN_BLINDING parameters
9671 - hide BN_BLINDING structure
9672 Add a second BN_BLINDING slot to the RSA structure to improve
9673 performance when a single RSA object is shared among several
9678 * Add support for DTLS.
9680 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9682 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9683 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9687 * Remove buggy and incomplete DH cert support from
9688 ssl/ssl_rsa.c and ssl/s3_both.c
9692 * Use SHA-1 instead of MD5 as the default digest algorithm for
9693 the apps/openssl applications.
9697 * Compile clean with "-Wall -Wmissing-prototypes
9698 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9699 DEBUG_SAFESTACK must also be set.
9703 * Change ./Configure so that certain algorithms can be disabled by default.
9704 The new counterpiece to "no-xxx" is "enable-xxx".
9706 The patented RC5 and MDC2 algorithms will now be disabled unless
9707 "enable-rc5" and "enable-mdc2", respectively, are specified.
9709 (IDEA remains enabled despite being patented. This is because IDEA
9710 is frequently required for interoperability, and there is no license
9711 fee for non-commercial use. As before, "no-idea" can be used to
9712 avoid this algorithm.)
9716 * Add processing of proxy certificates (see RFC 3820). This work was
9717 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9718 EGEE (Enabling Grids for E-science in Europe).
9722 * RC4 performance overhaul on modern architectures/implementations, such
9723 as Intel P4, IA-64 and AMD64.
9727 * New utility extract-section.pl. This can be used specify an alternative
9728 section number in a pod file instead of having to treat each file as
9729 a separate case in Makefile. This can be done by adding two lines to the
9732 =for comment openssl_section:XXX
9734 The blank line is mandatory.
9738 * New arguments -certform, -keyform and -pass for s_client and s_server
9739 to allow alternative format key and certificate files and passphrase
9744 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9745 update associated structures and add various utility functions.
9747 Add new policy related verify parameters, include policy checking in
9748 standard verify code. Enhance 'smime' application with extra parameters
9749 to support policy checking and print out.
9753 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9754 Nehemiah processors. These extensions support AES encryption in hardware
9755 as well as RNG (though RNG support is currently disabled).
9757 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9759 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9763 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9765 *Andy Polyakov and a number of other people*
9767 * Improved PowerPC platform support. Most notably BIGNUM assembler
9768 implementation contributed by IBM.
9770 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9772 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9773 exponent rather than 'unsigned long'. There is a corresponding change to
9774 the new 'rsa_keygen' element of the RSA_METHOD structure.
9776 *Jelte Jansen, Geoff Thorpe*
9778 * Functionality for creating the initial serial number file is now
9779 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9781 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9782 number file to 1, which is bound to cause problems. To avoid
9783 the problems while respecting compatibility between different 0.9.7
9784 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9785 CA.pl for serial number initialization. With the new release 0.9.8,
9786 we can fix the problem directly in the 'ca' utility.)
9790 * Reduced header interdependencies by declaring more opaque objects in
9791 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9792 give fewer recursive includes, which could break lazy source code - so
9793 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9794 developers should define this symbol when building and using openssl to
9795 ensure they track the recommended behaviour, interfaces, [etc], but
9796 backwards-compatible behaviour prevails when this isn't defined.
9800 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9804 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9805 This will generate a random key of the appropriate length based on the
9806 cipher context. The EVP_CIPHER can provide its own random key generation
9807 routine to support keys of a specific form. This is used in the des and
9808 3des routines to generate a key of the correct parity. Update S/MIME
9809 code to use new functions and hence generate correct parity DES keys.
9810 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9811 valid (weak or incorrect parity).
9815 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9816 as looking them up. This is useful when the verified structure may contain
9817 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9818 present unless the new PKCS7_NO_CRL flag is asserted.
9822 * Extend ASN1 oid configuration module. It now additionally accepts the
9825 shortName = some long name, 1.2.3.4
9829 * Reimplemented the BN_CTX implementation. There is now no more static
9830 limitation on the number of variables it can handle nor the depth of the
9831 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9832 information can now expand as required, and rather than having a single
9833 static array of bignums, BN_CTX now uses a linked-list of such arrays
9834 allowing it to expand on demand whilst maintaining the usefulness of
9835 BN_CTX's "bundling".
9839 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9840 to allow all RSA operations to function using a single BN_CTX.
9844 * Preliminary support for certificate policy evaluation and checking. This
9845 is initially intended to pass the tests outlined in "Conformance Testing
9846 of Relying Party Client Certificate Path Processing Logic" v1.07.
9850 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9851 remained unused and not that useful. A variety of other little bignum
9852 tweaks and fixes have also been made continuing on from the audit (see
9857 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9858 associated ASN1, EVP and SSL functions and old ASN1 macros.
9862 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9863 and this should never fail. So the return value from the use of
9864 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9865 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9869 * BN_CTX_get() should return zero-valued bignums, providing the same
9870 initialised value as BN_new().
9872 *Geoff Thorpe, suggested by Ulf Möller*
9874 * Support for inhibitAnyPolicy certificate extension.
9878 * An audit of the BIGNUM code is underway, for which debugging code is
9879 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9880 is considered valid when processing BIGNUMs, and causes execution to
9881 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9882 further steps are taken to deliberately pollute unused data in BIGNUM
9883 structures to try and expose faulty code further on. For now, openssl will
9884 (in its default mode of operation) continue to tolerate the inconsistent
9885 forms that it has tolerated in the past, but authors and packagers should
9886 consider trying openssl and their own applications when compiled with
9887 these debugging symbols defined. It will help highlight potential bugs in
9888 their own code, and will improve the test coverage for OpenSSL itself. At
9889 some point, these tighter rules will become openssl's default to improve
9890 maintainability, though the assert()s and other overheads will remain only
9891 in debugging configurations. See bn.h for more details.
9893 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9895 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9896 that can only be obtained through BN_CTX_new() (which implicitly
9897 initialises it). The presence of this function only made it possible
9898 to overwrite an existing structure (and cause memory leaks).
9902 * Because of the callback-based approach for implementing LHASH as a
9903 template type, lh_insert() adds opaque objects to hash-tables and
9904 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9905 to clean up those corresponding objects before destroying the hash table
9906 (and losing the object pointers). So some over-zealous constifications in
9907 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9908 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9909 prototyped to have "const" restrictions on the object pointers they are
9910 given (and so aren't required to cast them away any more).
9914 * The tmdiff.h API was so ugly and minimal that our own timing utility
9915 (speed) prefers to use its own implementation. The two implementations
9916 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9917 its object type properly exposed (MS_TM) instead of casting to/from
9918 `char *`. This may still change yet if someone realises MS_TM and
9920 aren't necessarily the greatest nomenclatures - but this is what was used
9921 internally to the implementation so I've used that for now.
9925 * Ensure that deprecated functions do not get compiled when
9926 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9927 the self-tests were still using deprecated key-generation functions so
9928 these have been updated also.
9932 * Reorganise PKCS#7 code to separate the digest location functionality
9933 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9934 New function PKCS7_set_digest() to set the digest type for PKCS#7
9935 digestedData type. Add additional code to correctly generate the
9936 digestedData type and add support for this type in PKCS7 initialization
9941 * New function PKCS7_set0_type_other() this initializes a PKCS7
9942 structure of type "other".
9946 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9947 sure the loop does correctly stop and breaking ("division by zero")
9948 modulus operations are not performed. The (pre-generated) prime
9949 table crypto/bn/bn_prime.h was already correct, but it could not be
9950 re-generated on some platforms because of the "division by zero"
9951 situation in the script.
9953 *Ralf S. Engelschall*
9955 * Update support for ECC-based TLS ciphersuites according to
9956 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9957 SHA-1 now is only used for "small" curves (where the
9958 representation of a field element takes up to 24 bytes); for
9959 larger curves, the field element resulting from ECDH is directly
9960 used as premaster secret.
9962 *Douglas Stebila (Sun Microsystems Laboratories)*
9964 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9965 curve secp160r1 to the tests.
9967 *Douglas Stebila (Sun Microsystems Laboratories)*
9969 * Add the possibility to load symbols globally with DSO.
9971 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9973 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9974 control of the error stack.
9978 * Add support for STORE in ENGINE.
9982 * Add the STORE type. The intention is to provide a common interface
9983 to certificate and key stores, be they simple file-based stores, or
9984 HSM-type store, or LDAP stores, or...
9985 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9989 * Add a generic structure called OPENSSL_ITEM. This can be used to
9990 pass a list of arguments to any function as well as provide a way
9991 for a function to pass data back to the caller.
9995 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9996 works like BUF_strdup() but can be used to duplicate a portion of
9997 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10002 * Add the function sk_find_ex() which works like sk_find(), but will
10003 return an index to an element even if an exact match couldn't be
10004 found. The index is guaranteed to point at the element where the
10005 searched-for key would be inserted to preserve sorting order.
10009 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10010 takes an extra flags argument for optional functionality. Currently,
10011 the following flags are defined:
10013 OBJ_BSEARCH_VALUE_ON_NOMATCH
10014 This one gets OBJ_bsearch_ex() to return a pointer to the first
10015 element where the comparing function returns a negative or zero
10018 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10019 This one gets OBJ_bsearch_ex() to return a pointer to the first
10020 element where the comparing function returns zero. This is useful
10021 if there are more than one element where the comparing function
10026 * Make it possible to create self-signed certificates with 'openssl ca'
10027 in such a way that the self-signed certificate becomes part of the
10028 CA database and uses the same mechanisms for serial number generation
10029 as all other certificate signing. The new flag '-selfsign' enables
10030 this functionality. Adapt CA.sh and CA.pl.in.
10034 * Add functionality to check the public key of a certificate request
10035 against a given private. This is useful to check that a certificate
10036 request can be signed by that key (self-signing).
10040 * Make it possible to have multiple active certificates with the same
10041 subject in the CA index file. This is done only if the keyword
10042 'unique_subject' is set to 'no' in the main CA section (default
10043 if 'CA_default') of the configuration file. The value is saved
10044 with the database itself in a separate index attribute file,
10045 named like the index file with '.attr' appended to the name.
10049 * Generate multi-valued AVAs using '+' notation in config files for
10054 * Support for nameConstraints certificate extension.
10058 * Support for policyConstraints certificate extension.
10062 * Support for policyMappings certificate extension.
10066 * Make sure the default DSA_METHOD implementation only uses its
10067 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10068 and change its own handlers to be NULL so as to remove unnecessary
10069 indirection. This lets alternative implementations fallback to the
10070 default implementation more easily.
10074 * Support for directoryName in GeneralName related extensions
10079 * Make it possible to link applications using Makefile.shared.
10080 Make that possible even when linking against static libraries!
10084 * Support for single pass processing for S/MIME signing. This now
10085 means that S/MIME signing can be done from a pipe, in addition
10086 cleartext signing (multipart/signed type) is effectively streaming
10087 and the signed data does not need to be all held in memory.
10089 This is done with a new flag PKCS7_STREAM. When this flag is set
10090 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10091 is done after the data is output (and digests calculated) in
10092 SMIME_write_PKCS7().
10096 * Add full support for -rpath/-R, both in shared libraries and
10097 applications, at least on the platforms where it's known how
10102 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10103 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10104 will now compute a table of multiples of the generator that
10105 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10106 faster (notably in the case of a single point multiplication,
10107 scalar * generator).
10109 *Nils Larsch, Bodo Moeller*
10111 * IPv6 support for certificate extensions. The various extensions
10112 which use the IP:a.b.c.d can now take IPv6 addresses using the
10113 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10118 * Added an ENGINE that implements RSA by performing private key
10119 exponentiations with the GMP library. The conversions to and from
10120 GMP's mpz_t format aren't optimised nor are any montgomery forms
10121 cached, and on x86 it appears OpenSSL's own performance has caught up.
10122 However there are likely to be other architectures where GMP could
10123 provide a boost. This ENGINE is not built in by default, but it can be
10124 specified at Configure time and should be accompanied by the necessary
10125 linker additions, eg;
10126 ./config -DOPENSSL_USE_GMP -lgmp
10130 * "openssl engine" will not display ENGINE/DSO load failure errors when
10131 testing availability of engines with "-t" - the old behaviour is
10132 produced by increasing the feature's verbosity with "-tt".
10136 * ECDSA routines: under certain error conditions uninitialized BN objects
10137 could be freed. Solution: make sure initialization is performed early
10138 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10143 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10144 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10145 software implementations. For DSA and DH, parameter generation can
10146 also be overridden by providing the appropriate method callbacks.
10150 * Change the "progress" mechanism used in key-generation and
10151 primality testing to functions that take a new BN_GENCB pointer in
10152 place of callback/argument pairs. The new API functions have `_ex`
10153 postfixes and the older functions are reimplemented as wrappers for
10154 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10155 declarations of the old functions to help (graceful) attempts to
10156 migrate to the new functions. Also, the new key-generation API
10157 functions operate on a caller-supplied key-structure and return
10158 success/failure rather than returning a key or NULL - this is to
10159 help make "keygen" another member function of RSA_METHOD etc.
10161 Example for using the new callback interface:
10163 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10164 void *my_arg = ...;
10167 BN_GENCB_set(&my_cb, my_callback, my_arg);
10169 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10170 /* For the meaning of a, b in calls to my_callback(), see the
10171 * documentation of the function that calls the callback.
10172 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10173 * my_callback should return 1 if it wants BN_is_prime_ex()
10174 * to continue, or 0 to stop.
10179 * Change the ZLIB compression method to be stateful, and make it
10180 available to TLS with the number defined in
10181 draft-ietf-tls-compression-04.txt.
10185 * Add the ASN.1 structures and functions for CertificatePair, which
10186 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10188 CertificatePair ::= SEQUENCE {
10189 forward [0] Certificate OPTIONAL,
10190 reverse [1] Certificate OPTIONAL,
10191 -- at least one of the pair shall be present -- }
10193 Also implement the PEM functions to read and write certificate
10194 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10196 This needed to be defined, mostly for the sake of the LDAP
10197 attribute crossCertificatePair, but may prove useful elsewhere as
10202 * Make it possible to inhibit symlinking of shared libraries in
10203 Makefile.shared, for Cygwin's sake.
10207 * Extend the BIGNUM API by creating a function
10208 void BN_set_negative(BIGNUM *a, int neg);
10209 and a macro that behave like
10210 int BN_is_negative(const BIGNUM *a);
10212 to avoid the need to access 'a->neg' directly in applications.
10216 * Implement fast modular reduction for pseudo-Mersenne primes
10217 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10218 EC_GROUP_new_curve_GFp() will now automatically use this
10221 *Nils Larsch <nla@trustcenter.de>*
10223 * Add new lock type (CRYPTO_LOCK_BN).
10227 * Change the ENGINE framework to automatically load engines
10228 dynamically from specific directories unless they could be
10229 found to already be built in or loaded. Move all the
10230 current engines except for the cryptodev one to a new
10231 directory engines/.
10232 The engines in engines/ are built as shared libraries if
10233 the "shared" options was given to ./Configure or ./config.
10234 Otherwise, they are inserted in libcrypto.a.
10235 /usr/local/ssl/engines is the default directory for dynamic
10236 engines, but that can be overridden at configure time through
10237 the usual use of --prefix and/or --openssldir, and at run
10238 time with the environment variable OPENSSL_ENGINES.
10240 *Geoff Thorpe and Richard Levitte*
10242 * Add Makefile.shared, a helper makefile to build shared
10243 libraries. Adapt Makefile.org.
10247 * Add version info to Win32 DLLs.
10249 *Peter 'Luna' Runestig" <peter@runestig.com>*
10251 * Add new 'medium level' PKCS#12 API. Certificates and keys
10252 can be added using this API to created arbitrary PKCS#12
10253 files while avoiding the low level API.
10255 New options to PKCS12_create(), key or cert can be NULL and
10256 will then be omitted from the output file. The encryption
10257 algorithm NIDs can be set to -1 for no encryption, the mac
10258 iteration count can be set to 0 to omit the mac.
10260 Enhance pkcs12 utility by making the -nokeys and -nocerts
10261 options work when creating a PKCS#12 file. New option -nomac
10262 to omit the mac, NONE can be set for an encryption algorithm.
10263 New code is modified to use the enhanced PKCS12_create()
10264 instead of the low level API.
10268 * Extend ASN1 encoder to support indefinite length constructed
10269 encoding. This can output sequences tags and octet strings in
10270 this form. Modify pk7_asn1.c to support indefinite length
10271 encoding. This is experimental and needs additional code to
10272 be useful, such as an ASN1 bio and some enhanced streaming
10275 Extend template encode functionality so that tagging is passed
10276 down to the template encoder.
10280 * Let 'openssl req' fail if an argument to '-newkey' is not
10281 recognized instead of using RSA as a default.
10285 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10286 As these are not official, they are not included in "ALL";
10287 the "ECCdraft" ciphersuite group alias can be used to select them.
10289 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10291 * Add ECDH engine support.
10293 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10295 * Add ECDH in new directory crypto/ecdh/.
10297 *Douglas Stebila (Sun Microsystems Laboratories)*
10299 * Let BN_rand_range() abort with an error after 100 iterations
10300 without success (which indicates a broken PRNG).
10304 * Change BN_mod_sqrt() so that it verifies that the input value
10305 is really the square of the return value. (Previously,
10306 BN_mod_sqrt would show GIGO behaviour.)
10310 * Add named elliptic curves over binary fields from X9.62, SECG,
10311 and WAP/WTLS; add OIDs that were still missing.
10313 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10315 * Extend the EC library for elliptic curves over binary fields
10316 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10319 EC_GF2m_simple_method
10323 EC_GROUP_new_curve_GF2m
10324 EC_GROUP_set_curve_GF2m
10325 EC_GROUP_get_curve_GF2m
10326 EC_POINT_set_affine_coordinates_GF2m
10327 EC_POINT_get_affine_coordinates_GF2m
10328 EC_POINT_set_compressed_coordinates_GF2m
10330 Point compression for binary fields is disabled by default for
10331 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10334 As binary polynomials are represented as BIGNUMs, various members
10335 of the EC_GROUP and EC_POINT data structures can be shared
10336 between the implementations for prime fields and binary fields;
10337 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10338 are essentially identical to their `..._GFp` counterparts.
10339 (For simplicity, the `..._GFp` prefix has been dropped from
10340 various internal method names.)
10342 An internal 'field_div' method (similar to 'field_mul' and
10343 'field_sqr') has been added; this is used only for binary fields.
10345 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10347 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10348 through methods ('mul', 'precompute_mult').
10350 The generic implementations (now internally called 'ec_wNAF_mul'
10351 and 'ec_wNAF_precomputed_mult') remain the default if these
10352 methods are undefined.
10354 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10356 * New function EC_GROUP_get_degree, which is defined through
10357 EC_METHOD. For curves over prime fields, this returns the bit
10358 length of the modulus.
10360 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10362 * New functions EC_GROUP_dup, EC_POINT_dup.
10363 (These simply call ..._new and ..._copy).
10365 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10367 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10368 Polynomials are represented as BIGNUMs (where the sign bit is not
10369 used) in the following functions [macros]:
10372 BN_GF2m_sub [= BN_GF2m_add]
10373 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10374 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10375 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10377 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10378 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10379 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10380 BN_GF2m_cmp [= BN_ucmp]
10382 (Note that only the 'mod' functions are actually for fields GF(2^m).
10383 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10385 For some functions, an the irreducible polynomial defining a
10386 field can be given as an 'unsigned int[]' with strictly
10387 decreasing elements giving the indices of those bits that are set;
10388 i.e., p[] represents the polynomial
10389 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10391 p[0] > p[1] > ... > p[k] = 0.
10392 This applies to the following functions:
10395 BN_GF2m_mod_mul_arr
10396 BN_GF2m_mod_sqr_arr
10397 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10398 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10399 BN_GF2m_mod_exp_arr
10400 BN_GF2m_mod_sqrt_arr
10401 BN_GF2m_mod_solve_quad_arr
10405 Conversion can be performed by the following functions:
10410 bntest.c has additional tests for binary polynomial arithmetic.
10412 Two implementations for BN_GF2m_mod_div() are available.
10413 The default algorithm simply uses BN_GF2m_mod_inv() and
10414 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10415 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10416 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10418 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10420 * Add new error code 'ERR_R_DISABLED' that can be used when some
10421 functionality is disabled at compile-time.
10423 *Douglas Stebila <douglas.stebila@sun.com>*
10425 * Change default behaviour of 'openssl asn1parse' so that more
10426 information is visible when viewing, e.g., a certificate:
10428 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10429 mode the content of non-printable OCTET STRINGs is output in a
10430 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10431 avoid the appearance of a printable string.
10433 *Nils Larsch <nla@trustcenter.de>*
10435 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10437 EC_GROUP_set_asn1_flag()
10438 EC_GROUP_get_asn1_flag()
10439 EC_GROUP_set_point_conversion_form()
10440 EC_GROUP_get_point_conversion_form()
10441 These control ASN1 encoding details:
10442 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10443 has been set to OPENSSL_EC_NAMED_CURVE.
10444 - Points are encoded in uncompressed form by default; options for
10445 asn1_for are as for point2oct, namely
10446 POINT_CONVERSION_COMPRESSED
10447 POINT_CONVERSION_UNCOMPRESSED
10448 POINT_CONVERSION_HYBRID
10450 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10452 EC_GROUP_set_seed()
10453 EC_GROUP_get0_seed()
10454 EC_GROUP_get_seed_len()
10455 This is used only for ASN1 purposes (so far).
10457 *Nils Larsch <nla@trustcenter.de>*
10459 * Add 'field_type' member to EC_METHOD, which holds the NID
10460 of the appropriate field type OID. The new function
10461 EC_METHOD_get_field_type() returns this value.
10463 *Nils Larsch <nla@trustcenter.de>*
10466 EC_POINT_point2bn()
10467 EC_POINT_bn2point()
10468 EC_POINT_point2hex()
10469 EC_POINT_hex2point()
10470 providing useful interfaces to EC_POINT_point2oct() and
10471 EC_POINT_oct2point().
10473 *Nils Larsch <nla@trustcenter.de>*
10475 * Change internals of the EC library so that the functions
10476 EC_GROUP_set_generator()
10477 EC_GROUP_get_generator()
10478 EC_GROUP_get_order()
10479 EC_GROUP_get_cofactor()
10480 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10481 to methods, which would lead to unnecessary code duplication when
10482 adding different types of curves.
10484 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10486 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10487 arithmetic, and such that modified wNAFs are generated
10488 (which avoid length expansion in many cases).
10492 * Add a function EC_GROUP_check_discriminant() (defined via
10493 EC_METHOD) that verifies that the curve discriminant is non-zero.
10495 Add a function EC_GROUP_check() that makes some sanity tests
10496 on a EC_GROUP, its generator and order. This includes
10497 EC_GROUP_check_discriminant().
10499 *Nils Larsch <nla@trustcenter.de>*
10501 * Add ECDSA in new directory crypto/ecdsa/.
10503 Add applications 'openssl ecparam' and 'openssl ecdsa'
10504 (these are based on 'openssl dsaparam' and 'openssl dsa').
10506 ECDSA support is also included in various other files across the
10507 library. Most notably,
10508 - 'openssl req' now has a '-newkey ecdsa:file' option;
10509 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10510 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10511 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10512 them suitable for ECDSA where domain parameters must be
10513 extracted before the specific public key;
10514 - ECDSA engine support has been added.
10516 *Nils Larsch <nla@trustcenter.de>*
10518 * Include some named elliptic curves, and add OIDs from X9.62,
10519 SECG, and WAP/WTLS. Each curve can be obtained from the new
10521 EC_GROUP_new_by_curve_name(),
10522 and the list of available named curves can be obtained with
10523 EC_get_builtin_curves().
10524 Also add a 'curve_name' member to EC_GROUP objects, which can be
10526 EC_GROUP_set_curve_name()
10527 EC_GROUP_get_curve_name()
10529 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10531 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10532 was actually never needed) and in BN_mul(). The removal in BN_mul()
10533 required a small change in bn_mul_part_recursive() and the addition
10534 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10535 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10536 bn_sub_words() and bn_add_words() except they take arrays with
10541 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10543 * Cleanse PEM buffers before freeing them since they may contain
10546 *Benjamin Bennett <ben@psc.edu>*
10548 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10549 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10550 authentication-only ciphersuites.
10554 * Since AES128 and AES256 share a single mask bit in the logic of
10555 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10556 kludge to work properly if AES128 is available and AES256 isn't.
10560 * Expand security boundary to match 1.1.1 module.
10564 * Remove redundant features: hash file source, editing of test vectors
10565 modify fipsld to use external fips_premain.c signature.
10569 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10570 run algorithm test programs.
10574 * Make algorithm test programs more tolerant of whitespace.
10578 * Have SSL/TLS server implementation tolerate "mismatched" record
10579 protocol version while receiving ClientHello even if the
10580 ClientHello is fragmented. (The server can't insist on the
10581 particular protocol version it has chosen before the ServerHello
10582 message has informed the client about his choice.)
10586 * Load error codes if they are not already present instead of using a
10587 static variable. This allows them to be cleanly unloaded and reloaded.
10591 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10593 * Introduce limits to prevent malicious keys being able to
10594 cause a denial of service. [CVE-2006-2940][]
10596 *Steve Henson, Bodo Moeller*
10598 * Fix ASN.1 parsing of certain invalid structures that can result
10599 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10601 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10602 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10604 * Fix SSL client code which could crash if connecting to a
10605 malicious SSLv2 server. [CVE-2006-4343][]
10607 *Tavis Ormandy and Will Drewry, Google Security Team*
10609 * Change ciphersuite string processing so that an explicit
10610 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10611 will no longer include "AES128-SHA"), and any other similar
10612 ciphersuite (same bitmap) from *other* protocol versions (so that
10613 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10614 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10615 changes from 0.9.8b and 0.9.8d.
10619 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10621 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10622 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10624 * Change the Unix randomness entropy gathering to use poll() when
10625 possible instead of select(), since the latter has some
10626 undesirable limitations.
10628 *Darryl Miles via Richard Levitte and Bodo Moeller*
10630 * Disable rogue ciphersuites:
10632 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10633 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10634 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10636 The latter two were purportedly from
10637 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10640 Also deactivate the remaining ciphersuites from
10641 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10642 unofficial, and the ID has long expired.
10646 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10647 dual-core machines) and other potential thread-safety issues.
10651 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10653 * Adapt fipsld and the build system to link against the validated FIPS
10654 module in FIPS mode.
10658 * Fixes for VC++ 2005 build under Windows.
10662 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10663 from a Windows bash shell such as MSYS. It is autodetected from the
10664 "config" script when run from a VC++ environment. Modify standard VC++
10665 build to use fipscanister.o from the GNU make build.
10669 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10671 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10672 The value now differs depending on if you build for FIPS or not.
10673 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10674 safely run with a non-FIPSed libcrypto, as it may crash because of
10675 the difference induced by this change.
10679 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10681 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10682 (part of SSL_OP_ALL). This option used to disable the
10683 countermeasure against man-in-the-middle protocol-version
10684 rollback in the SSL 2.0 server implementation, which is a bad
10685 idea. [CVE-2005-2969][]
10687 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10688 for Information Security, National Institute of Advanced Industrial
10689 Science and Technology [AIST, Japan)]*
10691 * Minimal support for X9.31 signatures and PSS padding modes. This is
10692 mainly for FIPS compliance and not fully integrated at this stage.
10696 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10697 the exponentiation using a fixed-length exponent. (Otherwise,
10698 the information leaked through timing could expose the secret key
10699 after many signatures; cf. Bleichenbacher's attack on DSA with
10704 * Make a new fixed-window mod_exp implementation the default for
10705 RSA, DSA, and DH private-key operations so that the sequence of
10706 squares and multiplies and the memory access pattern are
10707 independent of the particular secret key. This will mitigate
10708 cache-timing and potential related attacks.
10710 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10711 and this is automatically used by BN_mod_exp_mont() if the new flag
10712 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10713 will use this BN flag for private exponents unless the flag
10714 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10715 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10717 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10719 * Change the client implementation for SSLv23_method() and
10720 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10721 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10722 (Previously, the SSL 2.0 backwards compatible Client Hello
10723 message format would be used even with SSL_OP_NO_SSLv2.)
10727 * Add support for smime-type MIME parameter in S/MIME messages which some
10732 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10733 a threadsafe manner. Modify rsa code to use new function and add calls
10734 to dsa and dh code (which had race conditions before).
10738 * Include the fixed error library code in the C error file definitions
10739 instead of fixing them up at runtime. This keeps the error code
10740 structures constant.
10744 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10746 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10749 * Fixes for newer kerberos headers. NB: the casts are needed because
10750 the 'length' field is signed on one version and unsigned on another
10751 with no (?) obvious way to tell the difference, without these VC++
10752 complains. Also the "definition" of FAR (blank) is no longer included
10753 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10754 some needed definitions.
10758 * Undo Cygwin change.
10762 * Added support for proxy certificates according to RFC 3820.
10763 Because they may be a security thread to unaware applications,
10764 they must be explicitly allowed in run-time. See
10765 docs/HOWTO/proxy_certificates.txt for further information.
10769 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10771 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10772 server and client random values. Previously
10773 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10774 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10776 This change has negligible security impact because:
10778 1. Server and client random values still have 24 bytes of pseudo random
10781 2. Server and client random values are sent in the clear in the initial
10784 3. The master secret is derived using the premaster secret (48 bytes in
10785 size for static RSA ciphersuites) as well as client server and random
10788 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10791 *Stephen Henson, reported by UK NISCC*
10793 * Use Windows randomness collection on Cygwin.
10797 * Fix hang in EGD/PRNGD query when communication socket is closed
10798 prematurely by EGD/PRNGD.
10800 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10802 * Prompt for pass phrases when appropriate for PKCS12 input format.
10806 * Back-port of selected performance improvements from development
10807 branch, as well as improved support for PowerPC platforms.
10811 * Add lots of checks for memory allocation failure, error codes to indicate
10812 failure and freeing up memory if a failure occurs.
10814 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10816 * Add new -passin argument to dgst.
10820 * Perform some character comparisons of different types in X509_NAME_cmp:
10821 this is needed for some certificates that re-encode DNs into UTF8Strings
10822 (in violation of RFC3280) and can't or won't issue name rollover
10827 * Make an explicit check during certificate validation to see that
10828 the CA setting in each certificate on the chain is correct. As a
10829 side effect always do the following basic checks on extensions,
10830 not just when there's an associated purpose to the check:
10832 - if there is an unhandled critical extension (unless the user
10833 has chosen to ignore this fault)
10834 - if the path length has been exceeded (if one is set at all)
10835 - that certain extensions fit the associated purpose (if one has
10840 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10842 * Avoid a race condition when CRLs are checked in a multi threaded
10843 environment. This would happen due to the reordering of the revoked
10844 entries during signature checking and serial number lookup. Now the
10845 encoding is cached and the serial number sort performed under a lock.
10846 Add new STACK function sk_is_sorted().
10850 * Add Delta CRL to the extension code.
10854 * Various fixes to s3_pkt.c so alerts are sent properly.
10856 *David Holmes <d.holmes@f5.com>*
10858 * Reduce the chances of duplicate issuer name and serial numbers (in
10859 violation of RFC3280) using the OpenSSL certificate creation utilities.
10860 This is done by creating a random 64 bit value for the initial serial
10861 number when a serial number file is created or when a self signed
10862 certificate is created using 'openssl req -x509'. The initial serial
10863 number file is created using 'openssl x509 -next_serial' in CA.pl
10864 rather than being initialized to 1.
10868 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10870 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10871 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10873 *Joe Orton, Steve Henson*
10875 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10878 *Joe Orton, Steve Henson*
10880 * Make it possible to have multiple active certificates with the same
10881 subject in the CA index file. This is done only if the keyword
10882 'unique_subject' is set to 'no' in the main CA section (default
10883 if 'CA_default') of the configuration file. The value is saved
10884 with the database itself in a separate index attribute file,
10885 named like the index file with '.attr' appended to the name.
10889 * X509 verify fixes. Disable broken certificate workarounds when
10890 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10891 keyUsage extension present. Don't accept CRLs with unhandled critical
10892 extensions: since verify currently doesn't process CRL extensions this
10893 rejects a CRL with *any* critical extensions. Add new verify error codes
10898 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10899 A clarification of RFC2560 will require the use of OCTET STRINGs and
10900 some implementations cannot handle the current raw format. Since OpenSSL
10901 copies and compares OCSP nonces as opaque blobs without any attempt at
10902 parsing them this should not create any compatibility issues.
10906 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10907 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10908 this HMAC (and other) operations are several times slower than OpenSSL
10913 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10915 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10917 * Use the correct content when signing type "other".
10921 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10923 * Fix various bugs revealed by running the NISCC test suite:
10925 Stop out of bounds reads in the ASN1 code when presented with
10926 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10928 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10930 If verify callback ignores invalid public key errors don't try to check
10931 certificate signature with the NULL public key.
10935 * New -ignore_err option in ocsp application to stop the server
10936 exiting on the first error in a request.
10940 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10941 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10946 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10947 extra data after the compression methods not only for TLS 1.0
10948 but also for SSL 3.0 (as required by the specification).
10950 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10952 * Change X509_certificate_type() to mark the key as exported/exportable
10953 when it's 512 *bits* long, not 512 bytes.
10957 * Change AES_cbc_encrypt() so it outputs exact multiple of
10958 blocks during encryption.
10962 * Various fixes to base64 BIO and non blocking I/O. On write
10963 flushes were not handled properly if the BIO retried. On read
10964 data was not being buffered properly and had various logic bugs.
10965 This also affects blocking I/O when the data being decoded is a
10970 * Various S/MIME bugfixes and compatibility changes:
10971 output correct application/pkcs7 MIME type if
10972 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10973 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10974 of files as .eml work). Correctly handle very long lines in MIME
10979 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10981 * Countermeasure against the Klima-Pokorny-Rosa extension of
10982 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10983 a protocol version number mismatch like a decryption error
10984 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10988 * Turn on RSA blinding by default in the default implementation
10989 to avoid a timing attack. Applications that don't want it can call
10990 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10991 They would be ill-advised to do so in most cases.
10993 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10995 * Change RSA blinding code so that it works when the PRNG is not
10996 seeded (in this case, the secret RSA exponent is abused as
10997 an unpredictable seed -- if it is not unpredictable, there
10998 is no point in blinding anyway). Make RSA blinding thread-safe
10999 by remembering the creator's thread ID in rsa->blinding and
11000 having all other threads use local one-time blinding factors
11001 (this requires more computation than sharing rsa->blinding, but
11002 avoids excessive locking; and if an RSA object is not shared
11003 between threads, blinding will still be very fast).
11007 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11008 ENGINE as defaults for all supported algorithms irrespective of
11009 the 'flags' parameter. 'flags' is now honoured, so applications
11010 should make sure they are passing it correctly.
11014 * Target "mingw" now allows native Windows code to be generated in
11015 the Cygwin environment as well as with the MinGW compiler.
11019 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11021 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11022 via timing by performing a MAC computation even if incorrect
11023 block cipher padding has been found. This is a countermeasure
11024 against active attacks where the attacker has to distinguish
11025 between bad padding and a MAC verification error. [CVE-2003-0078][]
11027 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11028 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11029 Martin Vuagnoux (EPFL, Ilion)*
11031 * Make the no-err option work as intended. The intention with no-err
11032 is not to have the whole error stack handling routines removed from
11033 libcrypto, it's only intended to remove all the function name and
11034 reason texts, thereby removing some of the footprint that may not
11035 be interesting if those errors aren't displayed anyway.
11037 NOTE: it's still possible for any application or module to have its
11038 own set of error texts inserted. The routines are there, just not
11039 used by default when no-err is given.
11043 * Add support for FreeBSD on IA64.
11045 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11047 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11048 Kerberos function mit_des_cbc_cksum(). Before this change,
11049 the value returned by DES_cbc_cksum() was like the one from
11050 mit_des_cbc_cksum(), except the bytes were swapped.
11052 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11054 * Allow an application to disable the automatic SSL chain building.
11055 Before this a rather primitive chain build was always performed in
11056 ssl3_output_cert_chain(): an application had no way to send the
11057 correct chain if the automatic operation produced an incorrect result.
11059 Now the chain builder is disabled if either:
11061 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11063 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11065 The reasoning behind this is that an application would not want the
11066 auto chain building to take place if extra chain certificates are
11067 present and it might also want a means of sending no additional
11068 certificates (for example the chain has two certificates and the
11073 * Add the possibility to build without the ENGINE framework.
11075 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11077 * Under Win32 gmtime() can return NULL: check return value in
11078 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11082 * DSA routines: under certain error conditions uninitialized BN objects
11083 could be freed. Solution: make sure initialization is performed early
11084 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11085 Nils Larsch <nla@trustcenter.de> via PR#459)
11089 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11090 checked on reconnect on the client side, therefore session resumption
11091 could still fail with a "ssl session id is different" error. This
11092 behaviour is masked when SSL_OP_ALL is used due to
11093 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11094 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11095 followup to PR #377.
11099 * IA-32 assembler support enhancements: unified ELF targets, support
11100 for SCO/Caldera platforms, fix for Cygwin shared build.
11104 * Add support for FreeBSD on sparc64. As a consequence, support for
11105 FreeBSD on non-x86 processors is separate from x86 processors on
11106 the config script, much like the NetBSD support.
11108 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11110 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11112 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11115 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11116 code (06) was taken as the first octet of the session ID and the last
11117 octet was ignored consequently. As a result SSLv2 client side session
11118 caching could not have worked due to the session ID mismatch between
11120 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11125 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11126 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11131 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11132 seems that in spite of existing for more than a year, many application
11133 author have done nothing to provide the necessary callbacks, which
11134 means that this particular engine will not work properly anywhere.
11135 This is a very unfortunate situation which forces us, in the name
11136 of usability, to give the hw_ncipher.c a static lock, which is part
11138 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11139 appear in 0.9.8 or later. We EXPECT application authors to have
11140 dealt properly with this when 0.9.8 is released (unless we actually
11141 make such changes in the libcrypto locking code that changes will
11142 have to be made anyway).
11146 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11147 octets have been read, EOF or an error occurs. Without this change
11148 some truncated ASN1 structures will not produce an error.
11152 * Disable Heimdal support, since it hasn't been fully implemented.
11153 Still give the possibility to force the use of Heimdal, but with
11154 warnings and a request that patches get sent to openssl-dev.
11158 * Add the VC-CE target, introduce the WINCE sysname, and add
11159 INSTALL.WCE and appropriate conditionals to make it build.
11161 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11163 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11164 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11165 edit numbers of the version.
11167 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11169 * Introduce safe string copy and catenation functions
11170 (BUF_strlcpy() and BUF_strlcat()).
11172 *Ben Laurie (CHATS) and Richard Levitte*
11174 * Avoid using fixed-size buffers for one-line DNs.
11176 *Ben Laurie (CHATS)*
11178 * Add BUF_MEM_grow_clean() to avoid information leakage when
11179 resizing buffers containing secrets, and use where appropriate.
11181 *Ben Laurie (CHATS)*
11183 * Avoid using fixed size buffers for configuration file location.
11185 *Ben Laurie (CHATS)*
11187 * Avoid filename truncation for various CA files.
11189 *Ben Laurie (CHATS)*
11191 * Use sizeof in preference to magic numbers.
11193 *Ben Laurie (CHATS)*
11195 * Avoid filename truncation in cert requests.
11197 *Ben Laurie (CHATS)*
11199 * Add assertions to check for (supposedly impossible) buffer
11202 *Ben Laurie (CHATS)*
11204 * Don't cache truncated DNS entries in the local cache (this could
11205 potentially lead to a spoofing attack).
11207 *Ben Laurie (CHATS)*
11209 * Fix various buffers to be large enough for hex/decimal
11210 representations in a platform independent manner.
11212 *Ben Laurie (CHATS)*
11214 * Add CRYPTO_realloc_clean() to avoid information leakage when
11215 resizing buffers containing secrets, and use where appropriate.
11217 *Ben Laurie (CHATS)*
11219 * Add BIO_indent() to avoid much slightly worrying code to do
11222 *Ben Laurie (CHATS)*
11224 * Convert sprintf()/BIO_puts() to BIO_printf().
11226 *Ben Laurie (CHATS)*
11228 * buffer_gets() could terminate with the buffer only half
11231 *Ben Laurie (CHATS)*
11233 * Add assertions to prevent user-supplied crypto functions from
11234 overflowing internal buffers by having large block sizes, etc.
11236 *Ben Laurie (CHATS)*
11238 * New OPENSSL_assert() macro (similar to assert(), but enabled
11241 *Ben Laurie (CHATS)*
11243 * Eliminate unused copy of key in RC4.
11245 *Ben Laurie (CHATS)*
11247 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11249 *Ben Laurie (CHATS)*
11251 * Fix off-by-one error in EGD path.
11253 *Ben Laurie (CHATS)*
11255 * If RANDFILE path is too long, ignore instead of truncating.
11257 *Ben Laurie (CHATS)*
11259 * Eliminate unused and incorrectly sized X.509 structure
11262 *Ben Laurie (CHATS)*
11264 * Eliminate unused and dangerous function knumber().
11266 *Ben Laurie (CHATS)*
11268 * Eliminate unused and dangerous structure, KSSL_ERR.
11270 *Ben Laurie (CHATS)*
11272 * Protect against overlong session ID context length in an encoded
11273 session object. Since these are local, this does not appear to be
11276 *Ben Laurie (CHATS)*
11278 * Change from security patch (see 0.9.6e below) that did not affect
11279 the 0.9.6 release series:
11281 Remote buffer overflow in SSL3 protocol - an attacker could
11282 supply an oversized master key in Kerberos-enabled versions.
11285 *Ben Laurie (CHATS)*
11287 * Change the SSL kerb5 codes to match RFC 2712.
11291 * Make -nameopt work fully for req and add -reqopt switch.
11293 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11295 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11297 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11299 * Make sure tests can be performed even if the corresponding algorithms
11300 have been removed entirely. This was also the last step to make
11301 OpenSSL compilable with DJGPP under all reasonable conditions.
11303 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11305 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11306 to allow version independent disabling of normally unselected ciphers,
11307 which may be activated as a side-effect of selecting a single cipher.
11309 (E.g., cipher list string "RSA" enables ciphersuites that are left
11310 out of "ALL" because they do not provide symmetric encryption.
11311 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11313 *Lutz Jaenicke, Bodo Moeller*
11315 * Add appropriate support for separate platform-dependent build
11316 directories. The recommended way to make a platform-dependent
11317 build directory is the following (tested on Linux), maybe with
11320 # Place yourself outside of the OpenSSL source tree. In
11321 # this example, the environment variable OPENSSL_SOURCE
11322 # is assumed to contain the absolute OpenSSL source directory.
11323 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11324 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11325 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11326 mkdir -p `dirname $F`
11327 ln -s $OPENSSL_SOURCE/$F $F
11330 To be absolutely sure not to disturb the source tree, a "make clean"
11331 is a good thing. If it isn't successful, don't worry about it,
11332 it probably means the source directory is very clean.
11336 * Make sure any ENGINE control commands make local copies of string
11337 pointers passed to them whenever necessary. Otherwise it is possible
11338 the caller may have overwritten (or deallocated) the original string
11339 data when a later ENGINE operation tries to use the stored values.
11341 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11343 * Improve diagnostics in file reading and command-line digests.
11345 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11347 * Add AES modes CFB and OFB to the object database. Correct an
11348 error in AES-CFB decryption.
11352 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11353 allows existing EVP_CIPHER_CTX structures to be reused after
11354 calling `EVP_*Final()`. This behaviour is used by encryption
11355 BIOs and some applications. This has the side effect that
11356 applications must explicitly clean up cipher contexts with
11357 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11361 * Check the values of dna and dnb in bn_mul_recursive before calling
11362 bn_mul_comba (a non zero value means the a or b arrays do not contain
11363 n2 elements) and fallback to bn_mul_normal if either is not zero.
11367 * Fix escaping of non-ASCII characters when using the -subj option
11368 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11372 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11373 form for "surname", serialNumber has no short form.
11374 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11375 therefore remove "mail" short name for "internet 7".
11376 The OID for unique identifiers in X509 certificates is
11377 x500UniqueIdentifier, not uniqueIdentifier.
11378 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11382 * Add an "init" command to the ENGINE config module and auto initialize
11383 ENGINEs. Without any "init" command the ENGINE will be initialized
11384 after all ctrl commands have been executed on it. If init=1 the
11385 ENGINE is initialized at that point (ctrls before that point are run
11386 on the uninitialized ENGINE and after on the initialized one). If
11387 init=0 then the ENGINE will not be initialized at all.
11391 * Fix the 'app_verify_callback' interface so that the user-defined
11392 argument is actually passed to the callback: In the
11393 SSL_CTX_set_cert_verify_callback() prototype, the callback
11394 declaration has been changed from
11397 int (*cb)(X509_STORE_CTX *,void *);
11398 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11399 i=s->ctx->app_verify_callback(&ctx)
11400 has been changed into
11401 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11403 To update applications using SSL_CTX_set_cert_verify_callback(),
11404 a dummy argument can be added to their callback functions.
11406 *D. K. Smetters <smetters@parc.xerox.com>*
11408 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11410 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11412 * Add and OPENSSL_LOAD_CONF define which will cause
11413 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11414 This allows older applications to transparently support certain
11415 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11416 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11417 load the config file and OPENSSL_add_all_algorithms_conf() which will
11418 always load it have also been added.
11422 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11423 Adjust NIDs and EVP layer.
11425 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11427 * Config modules support in openssl utility.
11429 Most commands now load modules from the config file,
11430 though in a few (such as version) this isn't done
11431 because it couldn't be used for anything.
11433 In the case of ca and req the config file used is
11434 the same as the utility itself: that is the -config
11435 command line option can be used to specify an
11440 * Move default behaviour from OPENSSL_config(). If appname is NULL
11441 use "openssl_conf" if filename is NULL use default openssl config file.
11445 * Add an argument to OPENSSL_config() to allow the use of an alternative
11446 config section name. Add a new flag to tolerate a missing config file
11447 and move code to CONF_modules_load_file().
11451 * Support for crypto accelerator cards from Accelerated Encryption
11452 Processing, www.aep.ie. (Use engine 'aep')
11453 The support was copied from 0.9.6c [engine] and adapted/corrected
11454 to work with the new engine framework.
11456 *AEP Inc. and Richard Levitte*
11458 * Support for SureWare crypto accelerator cards from Baltimore
11459 Technologies. (Use engine 'sureware')
11460 The support was copied from 0.9.6c [engine] and adapted
11461 to work with the new engine framework.
11465 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11466 make the newer ENGINE framework commands for the CHIL engine work.
11468 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11470 * Make it possible to produce shared libraries on ReliantUNIX.
11472 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11474 * Add the configuration target debug-linux-ppro.
11475 Make 'openssl rsa' use the general key loading routines
11476 implemented in apps.c, and make those routines able to
11477 handle the key format FORMAT_NETSCAPE and the variant
11480 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11482 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11484 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11486 * Add -keyform to rsautl, and document -engine.
11488 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11490 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11491 BIO_R_NO_SUCH_FILE error code rather than the generic
11492 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11496 * Add new functions
11497 ERR_peek_last_error
11498 ERR_peek_last_error_line
11499 ERR_peek_last_error_line_data.
11500 These are similar to
11502 ERR_peek_error_line
11503 ERR_peek_error_line_data,
11504 but report on the latest error recorded rather than the first one
11505 still in the error queue.
11507 *Ben Laurie, Bodo Moeller*
11509 * default_algorithms option in ENGINE config module. This allows things
11511 default_algorithms = ALL
11512 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11516 * Preliminary ENGINE config module.
11520 * New experimental application configuration code.
11524 * Change the AES code to follow the same name structure as all other
11525 symmetric ciphers, and behave the same way. Move everything to
11526 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11528 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11530 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11532 *Ben Laurie and Theo de Raadt*
11534 * Add option to output public keys in req command.
11536 *Massimiliano Pala madwolf@openca.org*
11538 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11539 (up to about 10% better than before for P-192 and P-224).
11543 * New functions/macros
11545 SSL_CTX_set_msg_callback(ctx, cb)
11546 SSL_CTX_set_msg_callback_arg(ctx, arg)
11547 SSL_set_msg_callback(ssl, cb)
11548 SSL_set_msg_callback_arg(ssl, arg)
11550 to request calling a callback function
11552 void cb(int write_p, int version, int content_type,
11553 const void *buf, size_t len, SSL *ssl, void *arg)
11555 whenever a protocol message has been completely received
11556 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11557 protocol version according to which the SSL library interprets
11558 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11559 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11560 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11561 specification (change_cipher_spec(20), alert(21), handshake(22)).
11562 'buf' and 'len' point to the actual message, 'ssl' to the
11563 SSL object, and 'arg' is the application-defined value set by
11564 SSL[_CTX]_set_msg_callback_arg().
11566 'openssl s_client' and 'openssl s_server' have new '-msg' options
11567 to enable a callback that displays all protocol messages.
11571 * Change the shared library support so shared libraries are built as
11572 soon as the corresponding static library is finished, and thereby get
11573 openssl and the test programs linked against the shared library.
11574 This still only happens when the keyword "shard" has been given to
11575 the configuration scripts.
11577 NOTE: shared library support is still an experimental thing, and
11578 backward binary compatibility is still not guaranteed.
11580 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11582 * Add support for Subject Information Access extension.
11584 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11586 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11587 additional bytes when new memory had to be allocated, not just
11588 when reusing an existing buffer.
11592 * New command line and configuration option 'utf8' for the req command.
11593 This allows field values to be specified as UTF8 strings.
11597 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11598 runs for the former and machine-readable output for the latter.
11602 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11603 of the e-mail address in the DN (i.e., it will go into a certificate
11604 extension only). The new configuration file option 'email_in_dn = no'
11605 has the same effect.
11607 *Massimiliano Pala madwolf@openca.org*
11609 * Change all functions with names starting with `des_` to be starting
11610 with `DES_` instead. Add wrappers that are compatible with libdes,
11611 but are named `_ossl_old_des_*`. Finally, add macros that map the
11612 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11613 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11614 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11617 Since we provide two compatibility mappings, the user needs to
11618 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11619 compatibility is desired. The default (i.e., when that macro
11620 isn't defined) is OpenSSL 0.9.6c compatibility.
11622 There are also macros that enable and disable the support of old
11623 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11624 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11625 are defined, the default will apply: to support the old des routines.
11627 In either case, one must include openssl/des.h to get the correct
11628 definitions. Do not try to just include openssl/des_old.h, that
11631 NOTE: This is a major break of an old API into a new one. Software
11632 authors are encouraged to switch to the `DES_` style functions. Some
11633 time in the future, des_old.h and the libdes compatibility functions
11634 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11635 default), and then completely removed.
11639 * Test for certificates which contain unsupported critical extensions.
11640 If such a certificate is found during a verify operation it is
11641 rejected by default: this behaviour can be overridden by either
11642 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11643 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11644 X509_supported_extension() has also been added which returns 1 if a
11645 particular extension is supported.
11649 * Modify the behaviour of EVP cipher functions in similar way to digests
11650 to retain compatibility with existing code.
11654 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11655 compatibility with existing code. In particular the 'ctx' parameter does
11656 not have to be to be initialized before the call to EVP_DigestInit() and
11657 it is tidied up after a call to EVP_DigestFinal(). New function
11658 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11659 EVP_MD_CTX_copy() changed to not require the destination to be
11660 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11661 requires the destination to be valid.
11663 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11664 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11668 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11669 so that complete 'Handshake' protocol structures are kept in memory
11670 instead of overwriting 'msg_type' and 'length' with 'body' data.
11674 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11676 *Massimo Santin via Richard Levitte*
11678 * Major restructuring to the underlying ENGINE code. This includes
11679 reduction of linker bloat, separation of pure "ENGINE" manipulation
11680 (initialisation, etc) from functionality dealing with implementations
11681 of specific crypto interfaces. This change also introduces integrated
11682 support for symmetric ciphers and digest implementations - so ENGINEs
11683 can now accelerate these by providing EVP_CIPHER and EVP_MD
11684 implementations of their own. This is detailed in
11685 [crypto/engine/README.md](crypto/engine/README.md)
11686 as it couldn't be adequately described here. However, there are a few
11687 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11688 were changed in the original introduction of ENGINE code have now
11689 reverted back - the hooking from this code to ENGINE is now a good
11690 deal more passive and at run-time, operations deal directly with
11691 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11692 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11693 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11694 they were not being used by the framework as there is no concept of a
11695 BIGNUM_METHOD and they could not be generalised to the new
11696 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11697 ENGINE_cpy() has been removed as it cannot be consistently defined in
11702 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11706 * Change mkdef.pl to sort symbols that get the same entry number,
11707 and make sure the automatically generated functions `ERR_load_*`
11708 become part of libeay.num as well.
11712 * New function SSL_renegotiate_pending(). This returns true once
11713 renegotiation has been requested (either SSL_renegotiate() call
11714 or HelloRequest/ClientHello received from the peer) and becomes
11715 false once a handshake has been completed.
11716 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11717 sends a HelloRequest, but does not ensure that a handshake takes
11718 place. SSL_renegotiate_pending() is useful for checking if the
11719 client has followed the request.)
11723 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11724 By default, clients may request session resumption even during
11725 renegotiation (if session ID contexts permit); with this option,
11726 session resumption is possible only in the first handshake.
11728 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11729 more bits available for options that should not be part of
11730 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11734 * Add some demos for certificate and certificate request creation.
11738 * Make maximum certificate chain size accepted from the peer application
11739 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11740 "Douglas E. Engert" <deengert@anl.gov>.
11744 * Add support for shared libraries for Unixware-7
11745 (Boyd Lynn Gerber <gerberb@zenez.com>).
11749 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11750 be done prior to destruction. Use this to unload error strings from
11751 ENGINEs that load their own error strings. NB: This adds two new API
11752 functions to "get" and "set" this destroy handler in an ENGINE.
11756 * Alter all existing ENGINE implementations (except "openssl" and
11757 "openbsd") to dynamically instantiate their own error strings. This
11758 makes them more flexible to be built both as statically-linked ENGINEs
11759 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11760 Also, add stub code to each that makes building them as self-contained
11761 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11765 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11766 implementations into applications that are completely implemented in
11767 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11768 commands that can be used to configure what shared-library to load and
11769 to control aspects of the way it is handled. Also, made an update to
11770 the [README-Engine.md](README-Engine.md) file
11771 that brings its information up-to-date and
11772 provides some information and instructions on the "dynamic" ENGINE
11773 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11777 * Make it possible to unload ranges of ERR strings with a new
11778 "ERR_unload_strings" function.
11782 * Add a copy() function to EVP_MD.
11786 * Make EVP_MD routines take a context pointer instead of just the
11787 md_data void pointer.
11791 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11792 that the digest can only process a single chunk of data
11793 (typically because it is provided by a piece of
11794 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11795 is only going to provide a single chunk of data, and hence the
11796 framework needn't accumulate the data for oneshot drivers.
11800 * As with "ERR", make it possible to replace the underlying "ex_data"
11801 functions. This change also alters the storage and management of global
11802 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11803 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11804 index counters. The API functions that use this state have been changed
11805 to take a "class_index" rather than pointers to the class's local STACK
11806 and counter, and there is now an API function to dynamically create new
11807 classes. This centralisation allows us to (a) plug a lot of the
11808 thread-safety problems that existed, and (b) makes it possible to clean
11809 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11810 such data would previously have always leaked in application code and
11811 workarounds were in place to make the memory debugging turn a blind eye
11812 to it. Application code that doesn't use this new function will still
11813 leak as before, but their memory debugging output will announce it now
11814 rather than letting it slide.
11816 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11817 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11818 has a return value to indicate success or failure.
11822 * Make it possible to replace the underlying "ERR" functions such that the
11823 global state (2 LHASH tables and 2 locks) is only used by the "default"
11824 implementation. This change also adds two functions to "get" and "set"
11825 the implementation prior to it being automatically set the first time
11826 any other ERR function takes place. Ie. an application can call "get",
11827 pass the return value to a module it has just loaded, and that module
11828 can call its own "set" function using that value. This means the
11829 module's "ERR" operations will use (and modify) the error state in the
11830 application and not in its own statically linked copy of OpenSSL code.
11834 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11835 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11836 the operation, and provides a more encapsulated way for external code
11837 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11838 to use these functions rather than manually incrementing the counts.
11840 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11844 * Add EVP test program.
11848 * Add symmetric cipher support to ENGINE. Expect the API to change!
11852 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11853 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11854 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11855 These allow a CRL to be built without having to access X509_CRL fields
11856 directly. Modify 'ca' application to use new functions.
11860 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11861 bug workarounds. Rollback attack detection is a security feature.
11862 The problem will only arise on OpenSSL servers when TLSv1 is not
11863 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11864 Software authors not wanting to support TLSv1 will have special reasons
11865 for their choice and can explicitly enable this option.
11867 *Bodo Moeller, Lutz Jaenicke*
11869 * Rationalise EVP so it can be extended: don't include a union of
11870 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11871 (similar to those existing for EVP_CIPHER_CTX).
11876 EVP_MD_CTX_init(&md); /* new function call */
11877 EVP_DigestInit(&md, EVP_sha1());
11878 EVP_DigestUpdate(&md, in, len);
11879 EVP_DigestFinal(&md, out, NULL);
11880 EVP_MD_CTX_cleanup(&md); /* new function call */
11884 * Make DES key schedule conform to the usual scheme, as well as
11885 correcting its structure. This means that calls to DES functions
11886 now have to pass a pointer to a des_key_schedule instead of a
11887 plain des_key_schedule (which was actually always a pointer
11890 des_key_schedule ks;
11892 des_set_key_checked(..., &ks);
11893 des_ncbc_encrypt(..., &ks, ...);
11895 (Note that a later change renames 'des_...' into 'DES_...'.)
11899 * Initial reduction of linker bloat: the use of some functions, such as
11900 PEM causes large amounts of unused functions to be linked in due to
11901 poor organisation. For example pem_all.c contains every PEM function
11902 which has a knock on effect of linking in large amounts of (unused)
11903 ASN1 code. Grouping together similar functions and splitting unrelated
11904 functions prevents this.
11908 * Cleanup of EVP macros.
11912 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11913 correct `_ecb suffix`.
11917 * Add initial OCSP responder support to ocsp application. The
11918 revocation information is handled using the text based index
11919 use by the ca application. The responder can either handle
11920 requests generated internally, supplied in files (for example
11921 via a CGI script) or using an internal minimal server.
11925 * Add configuration choices to get zlib compression for TLS.
11929 * Changes to Kerberos SSL for RFC 2712 compliance:
11930 1. Implemented real KerberosWrapper, instead of just using
11931 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11932 2. Implemented optional authenticator field of KerberosWrapper.
11934 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11935 and authenticator structs; see crypto/krb5/.
11937 Generalized Kerberos calls to support multiple Kerberos libraries.
11938 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11939 via Richard Levitte*
11941 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11942 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11943 values for each of the key sizes rather than having just
11944 parameters (and 'speed' generating keys each time).
11948 * Speed up EVP routines.
11951 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11952 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11953 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11954 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11956 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11957 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11958 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11961 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11963 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11967 * Added the OS2-EMX target.
11969 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11971 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11972 to support NCONF routines in extension code. New function CONF_set_nconf()
11973 to allow functions which take an NCONF to also handle the old LHASH
11974 structure: this means that the old CONF compatible routines can be
11975 retained (in particular wrt extensions) without having to duplicate the
11976 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11980 * Enhance the general user interface with mechanisms for inner control
11981 and with possibilities to have yes/no kind of prompts.
11985 * Change all calls to low level digest routines in the library and
11986 applications to use EVP. Add missing calls to HMAC_cleanup() and
11987 don't assume HMAC_CTX can be copied using memcpy().
11989 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11991 * Add the possibility to control engines through control names but with
11992 arbitrary arguments instead of just a string.
11993 Change the key loaders to take a UI_METHOD instead of a callback
11994 function pointer. NOTE: this breaks binary compatibility with earlier
11995 versions of OpenSSL [engine].
11996 Adapt the nCipher code for these new conditions and add a card insertion
12001 * Enhance the general user interface with mechanisms to better support
12002 dialog box interfaces, application-defined prompts, the possibility
12003 to use defaults (for example default passwords from somewhere else)
12004 and interrupts/cancellations.
12008 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12009 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12013 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12014 tidy up some unnecessarily weird code in 'sk_new()').
12016 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12018 * Change the key loading routines for ENGINEs to use the same kind
12019 callback (pem_password_cb) as all other routines that need this
12024 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12025 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12026 than this minimum value is recommended.
12030 * New random seeder for OpenVMS, using the system process statistics
12031 that are easily reachable.
12035 * Windows apparently can't transparently handle global
12036 variables defined in DLLs. Initialisations such as:
12038 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12040 won't compile. This is used by the any applications that need to
12041 declare their own ASN1 modules. This was fixed by adding the option
12042 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12043 needed for static libraries under Win32.
12047 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12048 setting of purpose and trust fields. New X509_STORE trust and
12049 purpose functions and tidy up setting in other SSL functions.
12053 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12054 structure. These are inherited by X509_STORE_CTX when it is
12055 initialised. This allows various defaults to be set in the
12056 X509_STORE structure (such as flags for CRL checking and custom
12057 purpose or trust settings) for functions which only use X509_STORE_CTX
12058 internally such as S/MIME.
12060 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12061 trust settings if they are not set in X509_STORE. This allows X509_STORE
12062 purposes and trust (in S/MIME for example) to override any set by default.
12064 Add command line options for CRL checking to smime, s_client and s_server
12069 * Initial CRL based revocation checking. If the CRL checking flag(s)
12070 are set then the CRL is looked up in the X509_STORE structure and
12071 its validity and signature checked, then if the certificate is found
12072 in the CRL the verify fails with a revoked error.
12074 Various new CRL related callbacks added to X509_STORE_CTX structure.
12076 Command line options added to 'verify' application to support this.
12078 This needs some additional work, such as being able to handle multiple
12079 CRLs with different times, extension based lookup (rather than just
12080 by subject name) and ultimately more complete V2 CRL extension
12085 * Add a general user interface API (crypto/ui/). This is designed
12086 to replace things like des_read_password and friends (backward
12087 compatibility functions using this new API are provided).
12088 The purpose is to remove prompting functions from the DES code
12089 section as well as provide for prompting through dialog boxes in
12090 a window system and the like.
12094 * Add "ex_data" support to ENGINE so implementations can add state at a
12095 per-structure level rather than having to store it globally.
12099 * Make it possible for ENGINE structures to be copied when retrieved by
12100 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12101 This causes the "original" ENGINE structure to act like a template,
12102 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12103 operational state can be localised to each ENGINE structure, despite the
12104 fact they all share the same "methods". New ENGINE structures returned in
12105 this case have no functional references and the return value is the single
12106 structural reference. This matches the single structural reference returned
12107 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12112 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12113 needs to match any other type at all we need to manually clear the
12118 * Changes to the "openssl engine" utility to include;
12119 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12120 about an ENGINE's available control commands.
12121 - executing control commands from command line arguments using the
12122 '-pre' and '-post' switches. '-post' is only used if '-t' is
12123 specified and the ENGINE is successfully initialised. The syntax for
12124 the individual commands are colon-separated, for example;
12125 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12129 * New dynamic control command support for ENGINEs. ENGINEs can now
12130 declare their own commands (numbers), names (strings), descriptions,
12131 and input types for run-time discovery by calling applications. A
12132 subset of these commands are implicitly classed as "executable"
12133 depending on their input type, and only these can be invoked through
12134 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12135 can be based on user input, config files, etc). The distinction is
12136 that "executable" commands cannot return anything other than a boolean
12137 result and can only support numeric or string input, whereas some
12138 discoverable commands may only be for direct use through
12139 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12140 pointers, or other custom uses. The "executable" commands are to
12141 support parameterisations of ENGINE behaviour that can be
12142 unambiguously defined by ENGINEs and used consistently across any
12143 OpenSSL-based application. Commands have been added to all the
12144 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12145 control over shared-library paths without source code alterations.
12149 * Changed all ENGINE implementations to dynamically allocate their
12150 ENGINEs rather than declaring them statically. Apart from this being
12151 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12152 this also allows the implementations to compile without using the
12153 internal engine_int.h header.
12157 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12158 'const' value. Any code that should be able to modify a RAND_METHOD
12159 should already have non-const pointers to it (ie. they should only
12160 modify their own ones).
12164 * Made a variety of little tweaks to the ENGINE code.
12165 - "atalla" and "ubsec" string definitions were moved from header files
12166 to C code. "nuron" string definitions were placed in variables
12167 rather than hard-coded - allowing parameterisation of these values
12168 later on via ctrl() commands.
12169 - Removed unused "#if 0"'d code.
12170 - Fixed engine list iteration code so it uses ENGINE_free() to release
12171 structural references.
12172 - Constified the RAND_METHOD element of ENGINE structures.
12173 - Constified various get/set functions as appropriate and added
12174 missing functions (including a catch-all ENGINE_cpy that duplicates
12175 all ENGINE values onto a new ENGINE except reference counts/state).
12176 - Removed NULL parameter checks in get/set functions. Setting a method
12177 or function to NULL is a way of cancelling out a previously set
12178 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12179 and doesn't justify the extra error symbols and code.
12180 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12181 flags from engine_int.h to engine.h.
12182 - Changed prototypes for ENGINE handler functions (init(), finish(),
12183 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12187 * Implement binary inversion algorithm for BN_mod_inverse in addition
12188 to the algorithm using long division. The binary algorithm can be
12189 used only if the modulus is odd. On 32-bit systems, it is faster
12190 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12191 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12192 up to 450 bits. In 64-bit environments, the binary algorithm
12193 appears to be advantageous for much longer moduli; here we use it
12194 for moduli up to 2048 bits.
12198 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12199 could not support the combine flag in choice fields.
12203 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12204 extensions from a certificate request to the certificate.
12208 * Allow multiple 'certopt' and 'nameopt' options to be separated
12209 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12210 file: this allows the display of the certificate about to be
12211 signed to be customised, to allow certain fields to be included
12212 or excluded and extension details. The old system didn't display
12213 multicharacter strings properly, omitted fields not in the policy
12214 and couldn't display additional details such as extensions.
12218 * Function EC_POINTs_mul for multiple scalar multiplication
12219 of an arbitrary number of elliptic curve points
12220 \sum scalars[i]*points[i],
12221 optionally including the generator defined for the EC_GROUP:
12222 scalar*generator + \sum scalars[i]*points[i].
12224 EC_POINT_mul is a simple wrapper function for the typical case
12225 that the point list has just one item (besides the optional
12230 * First EC_METHODs for curves over GF(p):
12232 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12233 operations and provides various method functions that can also
12234 operate with faster implementations of modular arithmetic.
12236 EC_GFp_mont_method() reuses most functions that are part of
12237 EC_GFp_simple_method, but uses Montgomery arithmetic.
12239 *Bodo Moeller; point addition and point doubling
12240 implementation directly derived from source code provided by
12241 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12243 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12244 crypto/ec/ec_lib.c):
12246 Curves are EC_GROUP objects (with an optional group generator)
12247 based on EC_METHODs that are built into the library.
12249 Points are EC_POINT objects based on EC_GROUP objects.
12251 Most of the framework would be able to handle curves over arbitrary
12252 finite fields, but as there are no obvious types for fields other
12253 than GF(p), some functions are limited to that for now.
12257 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12258 that the file contains a complete HTTP response.
12262 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12263 change the def and num file printf format specifier from "%-40sXXX"
12264 to "%-39s XXX". The latter will always guarantee a space after the
12265 field while the former will cause them to run together if the field
12266 is 40 of more characters long.
12270 * Constify the cipher and digest 'method' functions and structures
12271 and modify related functions to take constant EVP_MD and EVP_CIPHER
12276 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12277 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12281 * Modify `EVP_Digest*()` routines so they now return values. Although the
12282 internal software routines can never fail additional hardware versions
12287 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12289 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12290 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12293 ERR_R_NESTED_ASN1_ERROR
12295 ERR_R_MISSING_ASN1_EOS
12296 were 4 .. 9, conflicting with
12297 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12299 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12300 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12302 Add new error code 'ERR_R_INTERNAL_ERROR'.
12306 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12311 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12312 sets the subject name for a new request or supersedes the
12313 subject name in a given request. Formats that can be parsed are
12314 'CN=Some Name, OU=myOU, C=IT'
12316 'CN=Some Name/OU=myOU/C=IT'.
12318 Add options '-batch' and '-verbose' to 'openssl req'.
12320 *Massimiliano Pala <madwolf@hackmasters.net>*
12322 * Introduce the possibility to access global variables through
12323 functions on platform were that's the best way to handle exporting
12324 global variables in shared libraries. To enable this functionality,
12325 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12326 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12327 is normally done by Configure or something similar).
12329 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12330 in the source file (foo.c) like this:
12332 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12333 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12335 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12336 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12338 OPENSSL_DECLARE_GLOBAL(int,foo);
12339 #define foo OPENSSL_GLOBAL_REF(foo)
12340 OPENSSL_DECLARE_GLOBAL(double,bar);
12341 #define bar OPENSSL_GLOBAL_REF(bar)
12343 The #defines are very important, and therefore so is including the
12344 header file everywhere where the defined globals are used.
12346 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12347 of ASN.1 items, but that structure is a bit different.
12349 The largest change is in util/mkdef.pl which has been enhanced with
12350 better and easier to understand logic to choose which symbols should
12351 go into the Windows .def files as well as a number of fixes and code
12352 cleanup (among others, algorithm keywords are now sorted
12353 lexicographically to avoid constant rewrites).
12357 * In BN_div() keep a copy of the sign of 'num' before writing the
12358 result to 'rm' because if rm==num the value will be overwritten
12359 and produce the wrong result if 'num' is negative: this caused
12360 problems with BN_mod() and BN_nnmod().
12364 * Function OCSP_request_verify(). This checks the signature on an
12365 OCSP request and verifies the signer certificate. The signer
12366 certificate is just checked for a generic purpose and OCSP request
12371 * Add OCSP_check_validity() function to check the validity of OCSP
12372 responses. OCSP responses are prepared in real time and may only
12373 be a few seconds old. Simply checking that the current time lies
12374 between thisUpdate and nextUpdate max reject otherwise valid responses
12375 caused by either OCSP responder or client clock inaccuracy. Instead
12376 we allow thisUpdate and nextUpdate to fall within a certain period of
12377 the current time. The age of the response can also optionally be
12378 checked. Two new options -validity_period and -status_age added to
12383 * If signature or public key algorithm is unrecognized print out its
12384 OID rather that just UNKNOWN.
12388 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12389 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12390 ID to be generated from the issuer certificate alone which can then be
12391 passed to OCSP_id_issuer_cmp().
12395 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12396 ASN1 modules to export functions returning ASN1_ITEM pointers
12397 instead of the ASN1_ITEM structures themselves. This adds several
12398 new macros which allow the underlying ASN1 function/structure to
12399 be accessed transparently. As a result code should not use ASN1_ITEM
12400 references directly (such as &X509_it) but instead use the relevant
12401 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12402 use of the new ASN1 code on platforms where exporting structures
12403 is problematical (for example in shared libraries) but exporting
12404 functions returning pointers to structures is not.
12408 * Add support for overriding the generation of SSL/TLS session IDs.
12409 These callbacks can be registered either in an SSL_CTX or per SSL.
12410 The purpose of this is to allow applications to control, if they wish,
12411 the arbitrary values chosen for use as session IDs, particularly as it
12412 can be useful for session caching in multiple-server environments. A
12413 command-line switch for testing this (and any client code that wishes
12414 to use such a feature) has been added to "s_server".
12416 *Geoff Thorpe, Lutz Jaenicke*
12418 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12419 of the form `#if defined(...) || defined(...) || ...` and
12420 `#if !defined(...) && !defined(...) && ...`. This also avoids
12421 the growing number of special cases it was previously handling.
12425 * Make all configuration macros available for application by making
12426 sure they are available in opensslconf.h, by giving them names starting
12427 with `OPENSSL_` to avoid conflicts with other packages and by making
12428 sure e_os2.h will cover all platform-specific cases together with
12430 Additionally, it is now possible to define configuration/platform-
12431 specific names (called "system identities"). In the C code, these
12432 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12433 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12434 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12439 * New option -set_serial to 'req' and 'x509' this allows the serial
12440 number to use to be specified on the command line. Previously self
12441 signed certificates were hard coded with serial number 0 and the
12442 CA options of 'x509' had to use a serial number in a file which was
12447 * New options to 'ca' utility to support V2 CRL entry extensions.
12448 Currently CRL reason, invalidity date and hold instruction are
12449 supported. Add new CRL extensions to V3 code and some new objects.
12453 * New function EVP_CIPHER_CTX_set_padding() this is used to
12454 disable standard block padding (aka PKCS#5 padding) in the EVP
12455 API, which was previously mandatory. This means that the data is
12456 not padded in any way and so the total length much be a multiple
12457 of the block size, otherwise an error occurs.
12461 * Initial (incomplete) OCSP SSL support.
12465 * New function OCSP_parse_url(). This splits up a URL into its host,
12466 port and path components: primarily to parse OCSP URLs. New -url
12467 option to ocsp utility.
12471 * New nonce behavior. The return value of OCSP_check_nonce() now
12472 reflects the various checks performed. Applications can decide
12473 whether to tolerate certain situations such as an absent nonce
12474 in a response when one was present in a request: the ocsp application
12475 just prints out a warning. New function OCSP_add1_basic_nonce()
12476 this is to allow responders to include a nonce in a response even if
12477 the request is nonce-less.
12481 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12482 skipped when using openssl x509 multiple times on a single input file,
12483 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12487 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12488 set string type: to handle setting ASN1_TIME structures. Fix ca
12489 utility to correctly initialize revocation date of CRLs.
12493 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12494 the clients preferred ciphersuites and rather use its own preferences.
12495 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12496 Internet Explorer by ensuring unchanged hash method during stepup.
12497 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12501 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12502 to aes and add a new 'exist' option to print out symbols that don't
12507 * Additional options to ocsp utility to allow flags to be set and
12508 additional certificates supplied.
12512 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12513 OCSP client a number of certificate to only verify the response
12518 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12519 handle the new API. Currently only ECB, CBC modes supported. Add new
12522 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12523 Encryption Standard (AES) Ciphersuites for Transport Layer
12524 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12525 not enabled by default and were not part of the "ALL" ciphersuite
12526 alias because they were not yet official; they could be
12527 explicitly requested by specifying the "AESdraft" ciphersuite
12528 group alias. In the final release of OpenSSL 0.9.7, the group
12529 alias is called "AES" and is part of "ALL".)
12531 *Ben Laurie, Steve Henson, Bodo Moeller*
12533 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12534 request to response.
12538 * Functions for OCSP responders. OCSP_request_onereq_count(),
12539 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12540 extract information from a certificate request. OCSP_response_create()
12541 creates a response and optionally adds a basic response structure.
12542 OCSP_basic_add1_status() adds a complete single response to a basic
12543 response and returns the OCSP_SINGLERESP structure just added (to allow
12544 extensions to be included for example). OCSP_basic_add1_cert() adds a
12545 certificate to a basic response and OCSP_basic_sign() signs a basic
12546 response with various flags. New helper functions ASN1_TIME_check()
12547 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12548 (converts ASN1_TIME to GeneralizedTime).
12552 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12553 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12554 structure from a certificate. X509_pubkey_digest() digests the public_key
12555 contents: this is used in various key identifiers.
12559 * Make sk_sort() tolerate a NULL argument.
12561 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12563 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12564 passed by the function are trusted implicitly. If any of them signed the
12565 response then it is assumed to be valid and is not verified.
12569 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12570 to data. This was previously part of the PKCS7 ASN1 code. This
12571 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12572 *Steve Henson, reported by Kenneth R. Robinette
12573 <support@securenetterm.com>*
12575 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12576 routines: without these tracing memory leaks is very painful.
12577 Fix leaks in PKCS12 and PKCS7 routines.
12581 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12582 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12583 effectively meant GeneralizedTime would never be used. Now it
12584 is initialised to -1 but X509_time_adj() now has to check the value
12585 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12586 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12587 *Steve Henson, reported by Kenneth R. Robinette
12588 <support@securenetterm.com>*
12590 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12591 result in a zero length in the ASN1_INTEGER structure which was
12592 not consistent with the structure when d2i_ASN1_INTEGER() was used
12593 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12594 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12595 where it did not print out a minus for negative ASN1_INTEGER.
12599 * Add summary printout to ocsp utility. The various functions which
12600 convert status values to strings have been renamed to:
12601 OCSP_response_status_str(), OCSP_cert_status_str() and
12602 OCSP_crl_reason_str() and are no longer static. New options
12603 to verify nonce values and to disable verification. OCSP response
12604 printout format cleaned up.
12608 * Add additional OCSP certificate checks. These are those specified
12609 in RFC2560. This consists of two separate checks: the CA of the
12610 certificate being checked must either be the OCSP signer certificate
12611 or the issuer of the OCSP signer certificate. In the latter case the
12612 OCSP signer certificate must contain the OCSP signing extended key
12613 usage. This check is performed by attempting to match the OCSP
12614 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12615 in the OCSP_CERTID structures of the response.
12619 * Initial OCSP certificate verification added to OCSP_basic_verify()
12620 and related routines. This uses the standard OpenSSL certificate
12621 verify routines to perform initial checks (just CA validity) and
12622 to obtain the certificate chain. Then additional checks will be
12623 performed on the chain. Currently the root CA is checked to see
12624 if it is explicitly trusted for OCSP signing. This is used to set
12625 a root CA as a global signing root: that is any certificate that
12626 chains to that CA is an acceptable OCSP signing certificate.
12630 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12631 extensions from a separate configuration file.
12632 As when reading extensions from the main configuration file,
12633 the '-extensions ...' option may be used for specifying the
12636 *Massimiliano Pala <madwolf@comune.modena.it>*
12638 * New OCSP utility. Allows OCSP requests to be generated or
12639 read. The request can be sent to a responder and the output
12640 parsed, outputted or printed in text form. Not complete yet:
12641 still needs to check the OCSP response validity.
12645 * New subcommands for 'openssl ca':
12646 `openssl ca -status <serial>` prints the status of the cert with
12647 the given serial number (according to the index file).
12648 `openssl ca -updatedb` updates the expiry status of certificates
12651 *Massimiliano Pala <madwolf@comune.modena.it>*
12653 * New '-newreq-nodes' command option to CA.pl. This is like
12654 '-newreq', but calls 'openssl req' with the '-nodes' option
12655 so that the resulting key is not encrypted.
12657 *Damien Miller <djm@mindrot.org>*
12659 * New configuration for the GNU Hurd.
12661 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12663 * Initial code to implement OCSP basic response verify. This
12664 is currently incomplete. Currently just finds the signer's
12665 certificate and verifies the signature on the response.
12669 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12670 value of OPENSSLDIR. This is available via the new '-d' option
12671 to 'openssl version', and is also included in 'openssl version -a'.
12675 * Allowing defining memory allocation callbacks that will be given
12676 file name and line number information in additional arguments
12677 (a `const char*` and an int). The basic functionality remains, as
12678 well as the original possibility to just replace malloc(),
12679 realloc() and free() by functions that do not know about these
12680 additional arguments. To register and find out the current
12681 settings for extended allocation functions, the following
12682 functions are provided:
12684 CRYPTO_set_mem_ex_functions
12685 CRYPTO_set_locked_mem_ex_functions
12686 CRYPTO_get_mem_ex_functions
12687 CRYPTO_get_locked_mem_ex_functions
12689 These work the same way as CRYPTO_set_mem_functions and friends.
12690 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12691 extended allocation function is enabled.
12692 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12693 a conventional allocation function is enabled.
12695 *Richard Levitte, Bodo Moeller*
12697 * Finish off removing the remaining LHASH function pointer casts.
12698 There should no longer be any prototype-casting required when using
12699 the LHASH abstraction, and any casts that remain are "bugs". See
12700 the callback types and macros at the head of lhash.h for details
12701 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12705 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12706 If /dev/[u]random devices are not available or do not return enough
12707 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12709 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12710 /etc/entropy will be queried once each in this sequence, querying stops
12711 when enough entropy was collected without querying more sockets.
12715 * Change the Unix RAND_poll() variant to be able to poll several
12716 random devices, as specified by DEVRANDOM, until a sufficient amount
12717 of data has been collected. We spend at most 10 ms on each file
12718 (select timeout) and read in non-blocking mode. DEVRANDOM now
12719 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12720 (previously it was just the string "/dev/urandom"), so on typical
12721 platforms the 10 ms delay will never occur.
12722 Also separate out the Unix variant to its own file, rand_unix.c.
12723 For VMS, there's a currently-empty rand_vms.c.
12727 * Move OCSP client related routines to ocsp_cl.c. These
12728 provide utility functions which an application needing
12729 to issue a request to an OCSP responder and analyse the
12730 response will typically need: as opposed to those which an
12731 OCSP responder itself would need which will be added later.
12733 OCSP_request_sign() signs an OCSP request with an API similar
12734 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12735 response. OCSP_response_get1_basic() extracts basic response
12736 from response. OCSP_resp_find_status(): finds and extracts status
12737 information from an OCSP_CERTID structure (which will be created
12738 when the request structure is built). These are built from lower
12739 level functions which work on OCSP_SINGLERESP structures but
12740 won't normally be used unless the application wishes to examine
12741 extensions in the OCSP response for example.
12743 Replace nonce routines with a pair of functions.
12744 OCSP_request_add1_nonce() adds a nonce value and optionally
12745 generates a random value. OCSP_check_nonce() checks the
12746 validity of the nonce in an OCSP response.
12750 * Change function OCSP_request_add() to OCSP_request_add0_id().
12751 This doesn't copy the supplied OCSP_CERTID and avoids the
12752 need to free up the newly created id. Change return type
12753 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12754 This can then be used to add extensions to the request.
12755 Deleted OCSP_request_new(), since most of its functionality
12756 is now in OCSP_REQUEST_new() (and the case insensitive name
12757 clash) apart from the ability to set the request name which
12758 will be added elsewhere.
12762 * Update OCSP API. Remove obsolete extensions argument from
12763 various functions. Extensions are now handled using the new
12764 OCSP extension code. New simple OCSP HTTP function which
12765 can be used to send requests and parse the response.
12769 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12770 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12771 uses the special reorder version of SET OF to sort the attributes
12772 and reorder them to match the encoded order. This resolves a long
12773 standing problem: a verify on a PKCS7 structure just after signing
12774 it used to fail because the attribute order did not match the
12775 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12776 it uses the received order. This is necessary to tolerate some broken
12777 software that does not order SET OF. This is handled by encoding
12778 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12779 to produce the required SET OF.
12783 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12784 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12785 files to get correct declarations of the ASN.1 item variables.
12789 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12790 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12791 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12792 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12793 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12794 ASN1_ITEM and no wrapper functions.
12798 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12799 replace the old function pointer based I/O routines. Change most of
12800 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12804 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12805 lines, recognize more "algorithms" that can be deselected, and make
12806 it complain about algorithm deselection that isn't recognised.
12810 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12811 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12812 to use new functions. Add NO_ASN1_OLD which can be set to remove
12813 some old style ASN1 functions: this can be used to determine if old
12814 code will still work when these eventually go away.
12818 * New extension functions for OCSP structures, these follow the
12819 same conventions as certificates and CRLs.
12823 * New function X509V3_add1_i2d(). This automatically encodes and
12824 adds an extension. Its behaviour can be customised with various
12825 flags to append, replace or delete. Various wrappers added for
12826 certificates and CRLs.
12830 * Fix to avoid calling the underlying ASN1 print routine when
12831 an extension cannot be parsed. Correct a typo in the
12832 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12836 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12837 entries for variables.
12841 * Add functionality to apps/openssl.c for detecting locking
12842 problems: As the program is single-threaded, all we have
12843 to do is register a locking callback using an array for
12844 storing which locks are currently held by the program.
12848 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12849 SSL_get_ex_data_X509_STORE_idx(), which is used in
12850 ssl_verify_cert_chain() and thus can be called at any time
12851 during TLS/SSL handshakes so that thread-safety is essential.
12852 Unfortunately, the ex_data design is not at all suited
12853 for multi-threaded use, so it probably should be abolished.
12857 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12859 *Broadcom, tweaked and integrated by Geoff Thorpe*
12861 * Move common extension printing code to new function
12862 X509V3_print_extensions(). Reorganise OCSP print routines and
12863 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12867 * New function X509_signature_print() to remove duplication in some
12872 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12873 set (this was treated exactly the same as SET OF previously). This
12874 is used to reorder the STACK representing the structure to match the
12875 encoding. This will be used to get round a problem where a PKCS7
12876 structure which was signed could not be verified because the STACK
12877 order did not reflect the encoded order.
12881 * Reimplement the OCSP ASN1 module using the new code.
12885 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12886 for its ASN1 operations. The old style function pointers still exist
12887 for now but they will eventually go away.
12891 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12892 completely replaces the old ASN1 functionality with a table driven
12893 encoder and decoder which interprets an ASN1_ITEM structure describing
12894 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12895 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12896 has also been converted to the new form.
12900 * Change BN_mod_exp_recp so that negative moduli are tolerated
12901 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12902 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12903 for negative moduli.
12907 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12908 of not touching the result's sign bit.
12912 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12917 * Changed the LHASH code to use prototypes for callbacks, and created
12918 macros to declare and implement thin (optionally static) functions
12919 that provide type-safety and avoid function pointer casting for the
12920 type-specific callbacks.
12924 * Added Kerberos Cipher Suites to be used with TLS, as written in
12926 *Veers Staats <staatsvr@asc.hpc.mil>,
12927 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12929 * Reformat the FAQ so the different questions and answers can be divided
12930 in sections depending on the subject.
12934 * Have the zlib compression code load ZLIB.DLL dynamically under
12939 * New function BN_mod_sqrt for computing square roots modulo a prime
12940 (using the probabilistic Tonelli-Shanks algorithm unless
12941 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12942 be handled deterministically).
12944 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12946 * Make BN_mod_inverse faster by explicitly handling small quotients
12947 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12948 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12952 * New function BN_kronecker.
12956 * Fix BN_gcd so that it works on negative inputs; the result is
12957 positive unless both parameters are zero.
12958 Previously something reasonably close to an infinite loop was
12959 possible because numbers could be growing instead of shrinking
12960 in the implementation of Euclid's algorithm.
12964 * Fix BN_is_word() and BN_is_one() macros to take into account the
12965 sign of the number in question.
12967 Fix BN_is_word(a,w) to work correctly for w == 0.
12969 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12970 because its test if the absolute value of 'a' equals 'w'.
12971 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12972 it exists mostly for use in the implementations of BN_is_zero(),
12973 BN_is_one(), and BN_is_word().
12977 * New function BN_swap.
12981 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12982 the exponentiation functions are more likely to produce reasonable
12983 results on negative inputs.
12987 * Change BN_mod_mul so that the result is always non-negative.
12988 Previously, it could be negative if one of the factors was negative;
12989 I don't think anyone really wanted that behaviour.
12993 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
12994 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
12995 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
12996 and add new functions:
13005 BN_mod_lshift1_quick
13007 BN_mod_lshift_quick
13009 These functions always generate non-negative results.
13011 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13012 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13014 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13015 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13016 be reduced modulo `m`.
13018 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13021 The following entry accidentally appeared in the CHANGES file
13022 distributed with OpenSSL 0.9.7. The modifications described in
13023 it do *not* apply to OpenSSL 0.9.7.
13025 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13026 was actually never needed) and in BN_mul(). The removal in BN_mul()
13027 required a small change in bn_mul_part_recursive() and the addition
13028 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13029 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13030 bn_sub_words() and bn_add_words() except they take arrays with
13036 * In 'openssl passwd', verify passwords read from the terminal
13037 unless the '-salt' option is used (which usually means that
13038 verification would just waste user's time since the resulting
13039 hash is going to be compared with some given password hash)
13040 or the new '-noverify' option is used.
13042 This is an incompatible change, but it does not affect
13043 non-interactive use of 'openssl passwd' (passwords on the command
13044 line, '-stdin' option, '-in ...' option) and thus should not
13045 cause any problems.
13049 * Remove all references to RSAref, since there's no more need for it.
13053 * Make DSO load along a path given through an environment variable
13054 (SHLIB_PATH) with shl_load().
13058 * Constify the ENGINE code as a result of BIGNUM constification.
13059 Also constify the RSA code and most things related to it. In a
13060 few places, most notable in the depth of the ASN.1 code, ugly
13061 casts back to non-const were required (to be solved at a later
13066 * Make it so the openssl application has all engines loaded by default.
13070 * Constify the BIGNUM routines a little more.
13074 * Add the following functions:
13076 ENGINE_load_cswift()
13078 ENGINE_load_atalla()
13079 ENGINE_load_nuron()
13080 ENGINE_load_builtin_engines()
13082 That way, an application can itself choose if external engines that
13083 are built-in in OpenSSL shall ever be used or not. The benefit is
13084 that applications won't have to be linked with libdl or other dso
13085 libraries unless it's really needed.
13087 Changed 'openssl engine' to load all engines on demand.
13088 Changed the engine header files to avoid the duplication of some
13089 declarations (they differed!).
13093 * 'openssl engine' can now list capabilities.
13097 * Better error reporting in 'openssl engine'.
13101 * Never call load_dh_param(NULL) in s_server.
13105 * Add engine application. It can currently list engines by name and
13106 identity, and test if they are actually available.
13110 * Improve RPM specification file by forcing symbolic linking and making
13111 sure the installed documentation is also owned by root.root.
13113 *Damien Miller <djm@mindrot.org>*
13115 * Give the OpenSSL applications more possibilities to make use of
13116 keys (public as well as private) handled by engines.
13120 * Add OCSP code that comes from CertCo.
13124 * Add VMS support for the Rijndael code.
13128 * Added untested support for Nuron crypto accelerator.
13132 * Add support for external cryptographic devices. This code was
13133 previously distributed separately as the "engine" branch.
13135 *Geoff Thorpe, Richard Levitte*
13137 * Rework the filename-translation in the DSO code. It is now possible to
13138 have far greater control over how a "name" is turned into a filename
13139 depending on the operating environment and any oddities about the
13140 different shared library filenames on each system.
13144 * Support threads on FreeBSD-elf in Configure.
13148 * Fix for SHA1 assembly problem with MASM: it produces
13149 warnings about corrupt line number information when assembling
13150 with debugging information. This is caused by the overlapping
13153 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13156 NCONF_get_number() has no error checking at all. As a replacement,
13157 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13158 promoted strongly. The old NCONF_get_number is kept around for
13159 binary backward compatibility.
13160 Make it possible for methods to load from something other than a BIO,
13161 by providing a function pointer that is given a name instead of a BIO.
13162 For example, this could be used to load configuration data from an
13167 * Fix for non blocking accept BIOs. Added new I/O special reason
13168 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13169 with non blocking I/O was not possible because no retry code was
13170 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13175 * Added the beginnings of Rijndael support.
13179 * Fix for bug in DirectoryString mask setting. Add support for
13180 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13181 to allow certificate printing to more controllable, additional
13182 'certopt' option to 'x509' to allow new printing options to be
13187 * Clean old EAY MD5 hack from e_os.h.
13191 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13193 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13194 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13196 *Joe Orton, Steve Henson*
13198 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13200 * Fix additional bug revealed by the NISCC test suite:
13202 Stop bug triggering large recursion when presented with
13203 certain ASN.1 tags [CVE-2003-0851][]
13207 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13209 * Fix various bugs revealed by running the NISCC test suite:
13211 Stop out of bounds reads in the ASN1 code when presented with
13212 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13214 If verify callback ignores invalid public key errors don't try to check
13215 certificate signature with the NULL public key.
13219 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13220 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13225 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13226 extra data after the compression methods not only for TLS 1.0
13227 but also for SSL 3.0 (as required by the specification).
13229 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13231 * Change X509_certificate_type() to mark the key as exported/exportable
13232 when it's 512 *bits* long, not 512 bytes.
13236 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13238 * Countermeasure against the Klima-Pokorny-Rosa extension of
13239 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13240 a protocol version number mismatch like a decryption error
13241 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13245 * Turn on RSA blinding by default in the default implementation
13246 to avoid a timing attack. Applications that don't want it can call
13247 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13248 They would be ill-advised to do so in most cases.
13250 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13252 * Change RSA blinding code so that it works when the PRNG is not
13253 seeded (in this case, the secret RSA exponent is abused as
13254 an unpredictable seed -- if it is not unpredictable, there
13255 is no point in blinding anyway). Make RSA blinding thread-safe
13256 by remembering the creator's thread ID in rsa->blinding and
13257 having all other threads use local one-time blinding factors
13258 (this requires more computation than sharing rsa->blinding, but
13259 avoids excessive locking; and if an RSA object is not shared
13260 between threads, blinding will still be very fast).
13264 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13266 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13267 via timing by performing a MAC computation even if incorrect
13268 block cipher padding has been found. This is a countermeasure
13269 against active attacks where the attacker has to distinguish
13270 between bad padding and a MAC verification error. [CVE-2003-0078][]
13272 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13273 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13274 Martin Vuagnoux (EPFL, Ilion)*
13276 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13278 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13279 memory from its contents. This is done with a counter that will
13280 place alternating values in each byte. This can be used to solve
13281 two issues: 1) the removal of calls to memset() by highly optimizing
13282 compilers, and 2) cleansing with other values than 0, since those can
13283 be read through on certain media, for example a swap space on disk.
13287 * Bugfix: client side session caching did not work with external caching,
13288 because the session->cipher setting was not restored when reloading
13289 from the external cache. This problem was masked, when
13290 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13291 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13295 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13296 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13298 *Zeev Lieber <zeev-l@yahoo.com>*
13300 * Undo an undocumented change introduced in 0.9.6e which caused
13301 repeated calls to OpenSSL_add_all_ciphers() and
13302 OpenSSL_add_all_digests() to be ignored, even after calling
13307 * Change the default configuration reader to deal with last line not
13308 being properly terminated.
13312 * Change X509_NAME_cmp() so it applies the special rules on handling
13313 DN values that are of type PrintableString, as well as RDNs of type
13314 emailAddress where the value has the type ia5String.
13316 *stefank@valicert.com via Richard Levitte*
13318 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13319 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13320 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13321 the bitwise-OR of the two for use by the majority of applications
13322 wanting this behaviour, and update the docs. The documented
13323 behaviour and actual behaviour were inconsistent and had been
13324 changing anyway, so this is more a bug-fix than a behavioural
13327 *Geoff Thorpe, diagnosed by Nadav Har'El*
13329 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13330 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13334 * Fix initialization code race conditions in
13335 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13336 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13337 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13338 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13339 ssl2_get_cipher_by_char(),
13340 ssl3_get_cipher_by_char().
13342 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13344 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13345 the cached sessions are flushed, as the remove_cb() might use ex_data
13346 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13347 (see [openssl.org #212]).
13349 *Geoff Thorpe, Lutz Jaenicke*
13351 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13352 length, instead of the encoding length to d2i_ASN1_OBJECT.
13356 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13358 * [In 0.9.6g-engine release:]
13359 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13361 *Lynn Gazis <lgazis@rainbow.com>*
13363 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13365 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13366 and get fix the header length calculation.
13367 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13368 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13370 * Use proper error handling instead of 'assertions' in buffer
13371 overflow checks added in 0.9.6e. This prevents DoS (the
13372 assertions could call abort()).
13374 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13376 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13378 * Add various sanity checks to asn1_get_length() to reject
13379 the ASN1 length bytes if they exceed sizeof(long), will appear
13380 negative or the content length exceeds the length of the
13383 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13385 * Fix cipher selection routines: ciphers without encryption had no flags
13386 for the cipher strength set and where therefore not handled correctly
13387 by the selection routines (PR #130).
13391 * Fix EVP_dsa_sha macro.
13396 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13397 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13398 that was added in OpenSSL 0.9.6d.
13400 As the countermeasure turned out to be incompatible with some
13401 broken SSL implementations, the new option is part of SSL_OP_ALL.
13402 SSL_OP_ALL is usually employed when compatibility with weird SSL
13403 implementations is desired (e.g. '-bugs' option to 's_client' and
13404 's_server'), so the new option is automatically set in many
13409 * Changes in security patch:
13411 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13412 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13413 Air Force Materiel Command, USAF, under agreement number
13416 * Add various sanity checks to asn1_get_length() to reject
13417 the ASN1 length bytes if they exceed sizeof(long), will appear
13418 negative or the content length exceeds the length of the
13419 supplied buffer. [CVE-2002-0659][]
13421 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13423 * Assertions for various potential buffer overflows, not known to
13424 happen in practice.
13426 *Ben Laurie (CHATS)*
13428 * Various temporary buffers to hold ASCII versions of integers were
13429 too small for 64 bit platforms. [CVE-2002-0655][]
13430 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13432 * Remote buffer overflow in SSL3 protocol - an attacker could
13433 supply an oversized session ID to a client. [CVE-2002-0656][]
13435 *Ben Laurie (CHATS)*
13437 * Remote buffer overflow in SSL2 protocol - an attacker could
13438 supply an oversized client master key. [CVE-2002-0656][]
13440 *Ben Laurie (CHATS)*
13442 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13444 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13445 encoded as NULL) with id-dsa-with-sha1.
13447 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13449 * Check various `X509_...()` return values in apps/req.c.
13451 *Nils Larsch <nla@trustcenter.de>*
13453 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13454 an end-of-file condition would erroneously be flagged, when the CRLF
13455 was just at the end of a processed block. The bug was discovered when
13456 processing data through a buffering memory BIO handing the data to a
13457 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13458 <ptsekov@syntrex.com> and Nedelcho Stanev.
13462 * Implement a countermeasure against a vulnerability recently found
13463 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13464 before application data chunks to avoid the use of known IVs
13465 with data potentially chosen by the attacker.
13469 * Fix length checks in ssl3_get_client_hello().
13473 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13474 to prevent ssl3_read_internal() from incorrectly assuming that
13475 ssl3_read_bytes() found application data while handshake
13476 processing was enabled when in fact s->s3->in_read_app_data was
13477 merely automatically cleared during the initial handshake.
13479 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13481 * Fix object definitions for Private and Enterprise: they were not
13482 recognized in their shortname (=lowercase) representation. Extend
13483 obj_dat.pl to issue an error when using undefined keywords instead
13484 of silently ignoring the problem (Svenning Sorensen
13485 <sss@sss.dnsalias.net>).
13489 * Fix DH_generate_parameters() so that it works for 'non-standard'
13490 generators, i.e. generators other than 2 and 5. (Previously, the
13491 code did not properly initialise the 'add' and 'rem' values to
13492 BN_generate_prime().)
13494 In the new general case, we do not insist that 'generator' is
13495 actually a primitive root: This requirement is rather pointless;
13496 a generator of the order-q subgroup is just as good, if not
13501 * Map new X509 verification errors to alerts. Discovered and submitted by
13502 Tom Wu <tom@arcot.com>.
13506 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13507 returning non-zero before the data has been completely received
13508 when using non-blocking I/O.
13510 *Bodo Moeller; problem pointed out by John Hughes*
13512 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13514 *Ben Laurie, Lutz Jaenicke*
13516 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13517 Yoram Zahavi <YoramZ@gilian.com>).
13521 * Add information about CygWin 1.3 and on, and preserve proper
13522 configuration for the versions before that.
13524 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13526 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13527 check whether we deal with a copy of a session and do not delete from
13528 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13529 <izhar@checkpoint.com>.
13533 * Do not store session data into the internal session cache, if it
13534 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13535 flag is set). Proposed by Aslam <aslam@funk.com>.
13539 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13544 * [In 0.9.6d-engine release:]
13545 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13547 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13549 * Add the configuration target linux-s390x.
13551 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13553 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13554 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13555 variable as an indication that a ClientHello message has been
13556 received. As the flag value will be lost between multiple
13557 invocations of ssl3_accept when using non-blocking I/O, the
13558 function may not be aware that a handshake has actually taken
13559 place, thus preventing a new session from being added to the
13562 To avoid this problem, we now set s->new_session to 2 instead of
13563 using a local variable.
13565 *Lutz Jaenicke, Bodo Moeller*
13567 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13568 if the SSL_R_LENGTH_MISMATCH error is detected.
13570 *Geoff Thorpe, Bodo Moeller*
13572 * New 'shared_ldflag' column in Configure platform table.
13576 * Fix EVP_CIPHER_mode macro.
13578 *"Dan S. Camper" <dan@bti.net>*
13580 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13581 type, we must throw them away by setting rr->length to 0.
13583 *D P Chang <dpc@qualys.com>*
13585 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13587 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13588 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13589 worked incorrectly for those cases where range = `10..._2` and
13590 `3*range` is two bits longer than range.)
13594 * Only add signing time to PKCS7 structures if it is not already
13599 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13600 OBJ_ld_ce should be OBJ_id_ce.
13601 Also some ip-pda OIDs in crypto/objects/objects.txt were
13602 incorrect (cf. RFC 3039).
13604 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13606 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13607 returns early because it has nothing to do.
13609 *Andy Schneider <andy.schneider@bjss.co.uk>*
13611 * [In 0.9.6c-engine release:]
13612 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13614 *Andy Schneider <andy.schneider@bjss.co.uk>*
13616 * [In 0.9.6c-engine release:]
13617 Add support for Cryptographic Appliance's keyserver technology.
13618 (Use engine 'keyclient')
13620 *Cryptographic Appliances and Geoff Thorpe*
13622 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13623 is called via tools/c89.sh because arguments have to be
13624 rearranged (all '-L' options must appear before the first object
13627 *Richard Shapiro <rshapiro@abinitio.com>*
13629 * [In 0.9.6c-engine release:]
13630 Add support for Broadcom crypto accelerator cards, backported
13633 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13635 * [In 0.9.6c-engine release:]
13636 Add support for SureWare crypto accelerator cards from
13637 Baltimore Technologies. (Use engine 'sureware')
13639 *Baltimore Technologies and Mark Cox*
13641 * [In 0.9.6c-engine release:]
13642 Add support for crypto accelerator cards from Accelerated
13643 Encryption Processing, www.aep.ie. (Use engine 'aep')
13645 *AEP Inc. and Mark Cox*
13647 * Add a configuration entry for gcc on UnixWare.
13649 *Gary Benson <gbenson@redhat.com>*
13651 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13652 messages are stored in a single piece (fixed-length part and
13653 variable-length part combined) and fix various bugs found on the way.
13657 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13658 instead. BIO_gethostbyname() does not know what timeouts are
13659 appropriate, so entries would stay in cache even when they have
13661 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13663 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13664 faced with a pathologically small ClientHello fragment that does
13665 not contain client_version: Instead of aborting with an error,
13666 simply choose the highest available protocol version (i.e.,
13667 TLS 1.0 unless it is disabled). In practice, ClientHello
13668 messages are never sent like this, but this change gives us
13669 strictly correct behaviour at least for TLS.
13673 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13674 never resets s->method to s->ctx->method when called from within
13675 one of the SSL handshake functions.
13677 *Bodo Moeller; problem pointed out by Niko Baric*
13679 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13680 (sent using the client's version number) if client_version is
13681 smaller than the protocol version in use. Also change
13682 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13683 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13684 the client will at least see that alert.
13688 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13693 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13694 client receives HelloRequest while in a handshake.
13696 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13698 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13699 should end in 'break', not 'goto end' which circumvents various
13700 cleanups done in state SSL_ST_OK. But session related stuff
13701 must be disabled for SSL_ST_OK in the case that we just sent a
13704 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13705 before just sending a HelloRequest.
13707 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13709 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13710 reveal whether illegal block cipher padding was found or a MAC
13711 verification error occurred. (Neither SSLerr() codes nor alerts
13712 are directly visible to potential attackers, but the information
13713 may leak via logfiles.)
13715 Similar changes are not required for the SSL 2.0 implementation
13716 because the number of padding bytes is sent in clear for SSL 2.0,
13717 and the extra bytes are just ignored. However ssl/s2_pkt.c
13718 failed to verify that the purported number of padding bytes is in
13723 * Add OpenUNIX-8 support including shared libraries
13724 (Boyd Lynn Gerber <gerberb@zenez.com>).
13728 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13729 'wristwatch attack' using huge encoding parameters (cf.
13730 James H. Manger's CRYPTO 2001 paper). Note that the
13731 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13732 encoding parameters and hence was not vulnerable.
13736 * BN_sqr() bug fix.
13738 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13740 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13741 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13742 followed by modular reduction.
13744 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13746 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13747 equivalent based on BN_pseudo_rand() instead of BN_rand().
13751 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13752 This function was broken, as the check for a new client hello message
13753 to handle SGC did not allow these large messages.
13754 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13758 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13762 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13763 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13767 * Rework the configuration and shared library support for Tru64 Unix.
13768 The configuration part makes use of modern compiler features and
13769 still retains old compiler behavior for those that run older versions
13770 of the OS. The shared library support part includes a variant that
13771 uses the RPATH feature, and is available through the special
13772 configuration target "alpha-cc-rpath", which will never be selected
13775 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13777 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13778 with the same message size as in ssl3_get_certificate_request().
13779 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13780 messages might inadvertently be reject as too long.
13782 *Petr Lampa <lampa@fee.vutbr.cz>*
13784 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13788 * Modified SSL library such that the verify_callback that has been set
13789 specifically for an SSL object with SSL_set_verify() is actually being
13790 used. Before the change, a verify_callback set with this function was
13791 ignored and the verify_callback() set in the SSL_CTX at the time of
13792 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13793 to allow the necessary settings.
13797 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13798 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13799 done automatically (in contradiction to the requirements of the C
13800 standard). This made problems when used from OpenSSH.
13804 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13805 dh->length and always used
13807 BN_rand_range(priv_key, dh->p).
13809 BN_rand_range() is not necessary for Diffie-Hellman, and this
13810 specific range makes Diffie-Hellman unnecessarily inefficient if
13811 dh->length (recommended exponent length) is much smaller than the
13812 length of dh->p. We could use BN_rand_range() if the order of
13813 the subgroup was stored in the DH structure, but we only have
13818 BN_rand(priv_key, l, ...)
13820 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13827 RSA_eay_public_encrypt
13828 RSA_eay_private_decrypt
13829 RSA_eay_private_encrypt (signing)
13830 RSA_eay_public_decrypt (signature verification)
13832 (default implementations for RSA_public_encrypt,
13833 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13834 always reject numbers >= n.
13838 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13839 to synchronize access to 'locking_thread'. This is necessary on
13840 systems where access to 'locking_thread' (an 'unsigned long'
13841 variable) is not atomic.
13845 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13846 *before* setting the 'crypto_lock_rand' flag. The previous code had
13847 a race condition if 0 is a valid thread ID.
13849 *Travis Vitek <vitek@roguewave.com>*
13851 * Add support for shared libraries under Irix.
13853 *Albert Chin-A-Young <china@thewrittenword.com>*
13855 * Add configuration option to build on Linux on both big-endian and
13856 little-endian MIPS.
13858 *Ralf Baechle <ralf@uni-koblenz.de>*
13860 * Add the possibility to create shared libraries on HP-UX.
13864 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13866 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13867 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13868 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13869 PRNG state recovery was possible based on the output of
13870 one PRNG request appropriately sized to gain knowledge on
13871 'md' followed by enough consecutive 1-byte PRNG requests
13872 to traverse all of 'state'.
13874 1. When updating 'md_local' (the current thread's copy of 'md')
13875 during PRNG output generation, hash all of the previous
13876 'md_local' value, not just the half used for PRNG output.
13878 2. Make the number of bytes from 'state' included into the hash
13879 independent from the number of PRNG bytes requested.
13881 The first measure alone would be sufficient to avoid
13882 Markku-Juhani's attack. (Actually it had never occurred
13883 to me that the half of 'md_local' used for chaining was the
13884 half from which PRNG output bytes were taken -- I had always
13885 assumed that the secret half would be used.) The second
13886 measure makes sure that additional data from 'state' is never
13887 mixed into 'md_local' in small portions; this heuristically
13888 further strengthens the PRNG.
13892 * Fix crypto/bn/asm/mips3.s.
13896 * When only the key is given to "enc", the IV is undefined. Print out
13897 an error message in this case.
13901 * Handle special case when X509_NAME is empty in X509 printing routines.
13905 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13906 positive and less than q.
13910 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13911 used: it isn't thread safe and the add_lock_callback should handle
13914 *Paul Rose <Paul.Rose@bridge.com>*
13916 * Verify that incoming data obeys the block size in
13917 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13923 *Ulf Möller, Bodo Möller*
13925 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13926 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13927 when fixing the server behaviour for backwards-compatible 'client
13928 hello' messages. (Note that the attack is impractical against
13929 SSL 3.0 and TLS 1.0 anyway because length and version checking
13930 means that the probability of guessing a valid ciphertext is
13931 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13934 Before 0.9.5, the countermeasure (hide the error by generating a
13935 random 'decryption result') did not work properly because
13936 ERR_clear_error() was missing, meaning that SSL_get_error() would
13937 detect the supposedly ignored error.
13939 Both problems are now fixed.
13943 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13944 (previously it was 1024).
13948 * Fix for compatibility mode trust settings: ignore trust settings
13949 unless some valid trust or reject settings are present.
13953 * Fix for blowfish EVP: its a variable length cipher.
13957 * Fix various bugs related to DSA S/MIME verification. Handle missing
13958 parameters in DSA public key structures and return an error in the
13959 DSA routines if parameters are absent.
13963 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13964 in the current directory if neither $RANDFILE nor $HOME was set.
13965 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13966 caused some confusion to Windows users who haven't defined $HOME.
13967 Thus RAND_file_name() is changed again: e_os.h can define a
13968 DEFAULT_HOME, which will be used if $HOME is not set.
13969 For Windows, we use "C:"; on other platforms, we still require
13970 environment variables.
13972 * Move 'if (!initialized) RAND_poll()' into regions protected by
13973 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13974 having multiple threads call RAND_poll() concurrently.
13978 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13979 combination of a flag and a thread ID variable.
13980 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13981 flag), *other* threads can enter ssleay_add_bytes without obeying
13982 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13983 that they do not hold after the first thread unsets add_do_not_lock).
13987 * Change bctest again: '-x' expressions are not available in all
13988 versions of 'test'.
13992 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13994 * Fix a couple of memory leaks in PKCS7_dataDecode()
13996 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13998 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13999 the default extension for executables, if any. Also, make the perl
14000 scripts that use symlink() to test if it really exists and use "cp"
14001 if it doesn't. All this made OpenSSL compilable and installable in
14006 * Fix for asn1_GetSequence() for indefinite length constructed data.
14007 If SEQUENCE is length is indefinite just set c->slen to the total
14008 amount of data available.
14010 *Steve Henson, reported by shige@FreeBSD.org*
14012 *This change does not apply to 0.9.7.*
14014 * Change bctest to avoid here-documents inside command substitution
14015 (workaround for FreeBSD /bin/sh bug).
14016 For compatibility with Ultrix, avoid shell functions (introduced
14017 in the bctest version that searches along $PATH).
14021 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14022 with des_encrypt() defined on some operating systems, like Solaris
14027 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14028 On the Importance of Eliminating Errors in Cryptographic
14029 Computations, J. Cryptology 14 (2001) 2, 101-119,
14030 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14034 * MIPS assembler BIGNUM division bug fix.
14038 * Disabled incorrect Alpha assembler code.
14042 * Fix PKCS#7 decode routines so they correctly update the length
14043 after reading an EOC for the EXPLICIT tag.
14047 *This change does not apply to 0.9.7.*
14049 * Fix bug in PKCS#12 key generation routines. This was triggered
14050 if a 3DES key was generated with a 0 initial byte. Include
14051 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14052 (but broken) behaviour.
14056 * Enhance bctest to search for a working bc along $PATH and print
14059 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14061 * Fix memory leaks in err.c: free err_data string if necessary;
14062 don't write to the wrong index in ERR_set_error_data.
14066 * Implement ssl23_peek (analogous to ssl23_read), which previously
14071 * Replace rdtsc with `_emit` statements for VC++ version 5.
14073 *Jeremy Cooper <jeremy@baymoo.org>*
14075 * Make it possible to reuse SSLv2 sessions.
14079 * In copy_email() check for >= 0 as a return value for
14080 X509_NAME_get_index_by_NID() since 0 is a valid index.
14082 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14084 * Avoid coredump with unsupported or invalid public keys by checking if
14085 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14086 PKCS7_verify() fails with non detached data.
14090 * Don't use getenv in library functions when run as setuid/setgid.
14091 New function OPENSSL_issetugid().
14095 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14096 due to incorrect handling of multi-threading:
14098 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14100 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14102 3. Count how many times MemCheck_off() has been called so that
14103 nested use can be treated correctly. This also avoids
14104 inband-signalling in the previous code (which relied on the
14105 assumption that thread ID 0 is impossible).
14109 * Add "-rand" option also to s_client and s_server.
14113 * Fix CPU detection on Irix 6.x.
14114 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14115 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14117 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14122 *This change does not apply to 0.9.7.*
14124 * Use the cached encoding of an X509_NAME structure rather than
14125 copying it. This is apparently the reason for the libsafe "errors"
14126 but the code is actually correct.
14130 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14131 Bleichenbacher's DSA attack.
14132 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14133 to be set and top=0 forces the highest bit to be set; top=-1 is new
14134 and leaves the highest bit random.
14136 *Ulf Moeller, Bodo Moeller*
14138 * In the `NCONF_...`-based implementations for `CONF_...` queries
14139 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14140 a temporary CONF structure with the data component set to NULL
14141 (which gives segmentation faults in lh_retrieve).
14142 Instead, use NULL for the CONF pointer in CONF_get_string and
14143 CONF_get_number (which may use environment variables) and directly
14144 return NULL from CONF_get_section.
14148 * Fix potential buffer overrun for EBCDIC.
14152 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14153 keyUsage if basicConstraints absent for a CA.
14157 * Make SMIME_write_PKCS7() write mail header values with a format that
14158 is more generally accepted (no spaces before the semicolon), since
14159 some programs can't parse those values properly otherwise. Also make
14160 sure BIO's that break lines after each write do not create invalid
14165 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14166 macros previously used would not encode an empty SEQUENCE OF
14167 and break the signature.
14171 *This change does not apply to 0.9.7.*
14173 * Zero the premaster secret after deriving the master secret in
14178 * Add some EVP_add_digest_alias registrations (as found in
14179 OpenSSL_add_all_digests()) to SSL_library_init()
14180 aka OpenSSL_add_ssl_algorithms(). This provides improved
14181 compatibility with peers using X.509 certificates
14182 with unconventional AlgorithmIdentifier OIDs.
14186 * Fix for Irix with NO_ASM.
14188 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14190 * ./config script fixes.
14192 *Ulf Moeller, Richard Levitte*
14194 * Fix 'openssl passwd -1'.
14198 * Change PKCS12_key_gen_asc() so it can cope with non null
14199 terminated strings whose length is passed in the passlen
14200 parameter, for example from PEM callbacks. This was done
14201 by adding an extra length parameter to asc2uni().
14203 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14205 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14206 call failed, free the DSA structure.
14210 * Fix to uni2asc() to cope with zero length Unicode strings.
14211 These are present in some PKCS#12 files.
14215 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14216 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14217 when writing a 32767 byte record.
14219 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14221 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14222 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14224 (RSA objects have a reference count access to which is protected
14225 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14226 so they are meant to be shared between threads.)
14227 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14228 "Reddie, Steven" <Steven.Reddie@ca.com>*
14230 * Fix a deadlock in CRYPTO_mem_leaks().
14234 * Use better test patterns in bntest.
14238 * rand_win.c fix for Borland C.
14242 * BN_rshift bugfix for n == 0.
14246 * Add a 'bctest' script that checks for some known 'bc' bugs
14247 so that 'make test' does not abort just because 'bc' is broken.
14251 * Store verify_result within SSL_SESSION also for client side to
14252 avoid potential security hole. (Re-used sessions on the client side
14253 always resulted in verify_result==X509_V_OK, not using the original
14254 result of the server certificate verification.)
14258 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14259 SSL3_RT_APPLICATION_DATA, return 0.
14260 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14265 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14266 releases, have been re-implemented by renaming the previous
14267 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14268 and ssl3_read_internal, respectively, and adding 'peek' parameters
14269 to them. The new ssl[23]_{read,peek} functions are calls to
14270 ssl[23]_read_internal with the 'peek' flag set appropriately.
14271 A 'peek' parameter has also been added to ssl3_read_bytes, which
14272 does the actual work for ssl3_read_internal.
14276 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14277 the method-specific "init()" handler. Also clean up ex_data after
14278 calling the method-specific "finish()" handler. Previously, this was
14279 happening the other way round.
14283 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14284 The previous value, 12, was not always sufficient for BN_mod_exp().
14288 * Make sure that shared libraries get the internal name engine with
14289 the full version number and not just 0. This should mark the
14290 shared libraries as not backward compatible. Of course, this should
14291 be changed again when we can guarantee backward binary compatibility.
14295 * Fix typo in get_cert_by_subject() in by_dir.c
14297 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14299 * Rework the system to generate shared libraries:
14301 - Make note of the expected extension for the shared libraries and
14302 if there is a need for symbolic links from for example libcrypto.so.0
14303 to libcrypto.so.0.9.7. There is extended info in Configure for
14306 - Make as few rebuilds of the shared libraries as possible.
14308 - Still avoid linking the OpenSSL programs with the shared libraries.
14310 - When installing, install the shared libraries separately from the
14315 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14317 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14318 and not in SSL_clear because the latter is also used by the
14319 accept/connect functions; previously, the settings made by
14320 SSL_set_read_ahead would be lost during the handshake.
14322 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14324 * Correct util/mkdef.pl to be selective about disabled algorithms.
14325 Previously, it would create entries for disabled algorithms no
14330 * Added several new manual pages for SSL_* function.
14334 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14336 * In ssl23_get_client_hello, generate an error message when faced
14337 with an initial SSL 3.0/TLS record that is too small to contain the
14338 first two bytes of the ClientHello message, i.e. client_version.
14339 (Note that this is a pathologic case that probably has never happened
14340 in real life.) The previous approach was to use the version number
14341 from the record header as a substitute; but our protocol choice
14342 should not depend on that one because it is not authenticated
14343 by the Finished messages.
14347 * More robust randomness gathering functions for Windows.
14349 *Jeffrey Altman <jaltman@columbia.edu>*
14351 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14352 not set then we don't setup the error code for issuer check errors
14353 to avoid possibly overwriting other errors which the callback does
14354 handle. If an application does set the flag then we assume it knows
14355 what it is doing and can handle the new informational codes
14360 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14361 a general "ANY" type, as such it should be able to decode anything
14362 including tagged types. However it didn't check the class so it would
14363 wrongly interpret tagged types in the same way as their universal
14364 counterpart and unknown types were just rejected. Changed so that the
14365 tagged and unknown types are handled in the same way as a SEQUENCE:
14366 that is the encoding is stored intact. There is also a new type
14367 "V_ASN1_OTHER" which is used when the class is not universal, in this
14368 case we have no idea what the actual type is so we just lump them all
14373 * On VMS, stdout may very well lead to a file that is written to
14374 in a record-oriented fashion. That means that every write() will
14375 write a separate record, which will be read separately by the
14376 programs trying to read from it. This can be very confusing.
14378 The solution is to put a BIO filter in the way that will buffer
14379 text until a linefeed is reached, and then write everything a
14380 line at a time, so every record written will be an actual line,
14381 not chunks of lines and not (usually doesn't happen, but I've
14382 seen it once) several lines in one record. BIO_f_linebuffer() is
14385 Currently, it's a VMS-only method, because that's where it has
14386 been tested well enough.
14390 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14391 it can return incorrect results.
14392 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14393 but it was in 0.9.6-beta[12].)
14397 * Disable the check for content being present when verifying detached
14398 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14399 include zero length content when signing messages.
14403 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14404 BIO_ctrl (for BIO pairs).
14408 * Add DSO method for VMS.
14412 * Bug fix: Montgomery multiplication could produce results with the
14417 * Add RPM specification openssl.spec and modify it to build three
14418 packages. The default package contains applications, application
14419 documentation and run-time libraries. The devel package contains
14420 include files, static libraries and function documentation. The
14421 doc package contains the contents of the doc directory. The original
14422 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14426 * Add a large number of documentation files for many SSL routines.
14428 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14430 * Add a configuration entry for Sony News 4.
14432 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14434 * Don't set the two most significant bits to one when generating a
14435 random number < q in the DSA library.
14439 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14440 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14441 the underlying transport is blocking) if a handshake took place.
14442 (The default behaviour is needed by applications such as s_client
14443 and s_server that use select() to determine when to use SSL_read;
14444 but for applications that know in advance when to expect data, it
14445 just makes things more complicated.)
14449 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14454 * Add a few more EBCDIC conditionals that make `req` and `x509`
14455 work better on such systems.
14457 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14459 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14460 Update PKCS12_parse() so it copies the friendlyName and the
14461 keyid to the certificates aux info.
14465 * Fix bug in PKCS7_verify() which caused an infinite loop
14466 if there was more than one signature.
14468 *Sven Uszpelkat <su@celocom.de>*
14470 * Major change in util/mkdef.pl to include extra information
14471 about each symbol, as well as presenting variables as well
14472 as functions. This change means that there's n more need
14473 to rebuild the .num files when some algorithms are excluded.
14477 * Allow the verify time to be set by an application,
14478 rather than always using the current time.
14482 * Phase 2 verify code reorganisation. The certificate
14483 verify code now looks up an issuer certificate by a
14484 number of criteria: subject name, authority key id
14485 and key usage. It also verifies self signed certificates
14486 by the same criteria. The main comparison function is
14487 X509_check_issued() which performs these checks.
14489 Lot of changes were necessary in order to support this
14490 without completely rewriting the lookup code.
14492 Authority and subject key identifier are now cached.
14494 The LHASH 'certs' is X509_STORE has now been replaced
14495 by a STACK_OF(X509_OBJECT). This is mainly because an
14496 LHASH can't store or retrieve multiple objects with
14497 the same hash value.
14499 As a result various functions (which were all internal
14500 use only) have changed to handle the new X509_STORE
14501 structure. This will break anything that messed round
14502 with X509_STORE internally.
14504 The functions X509_STORE_add_cert() now checks for an
14505 exact match, rather than just subject name.
14507 The X509_STORE API doesn't directly support the retrieval
14508 of multiple certificates matching a given criteria, however
14509 this can be worked round by performing a lookup first
14510 (which will fill the cache with candidate certificates)
14511 and then examining the cache for matches. This is probably
14512 the best we can do without throwing out X509_LOOKUP
14513 entirely (maybe later...).
14515 The X509_VERIFY_CTX structure has been enhanced considerably.
14517 All certificate lookup operations now go via a get_issuer()
14518 callback. Although this currently uses an X509_STORE it
14519 can be replaced by custom lookups. This is a simple way
14520 to bypass the X509_STORE hackery necessary to make this
14521 work and makes it possible to use more efficient techniques
14522 in future. A very simple version which uses a simple
14523 STACK for its trusted certificate store is also provided
14524 using X509_STORE_CTX_trusted_stack().
14526 The verify_cb() and verify() callbacks now have equivalents
14527 in the X509_STORE_CTX structure.
14529 X509_STORE_CTX also has a 'flags' field which can be used
14530 to customise the verify behaviour.
14534 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14535 excludes S/MIME capabilities.
14539 * When a certificate request is read in keep a copy of the
14540 original encoding of the signed data and use it when outputting
14541 again. Signatures then use the original encoding rather than
14542 a decoded, encoded version which may cause problems if the
14543 request is improperly encoded.
14547 * For consistency with other BIO_puts implementations, call
14548 buffer_write(b, ...) directly in buffer_puts instead of calling
14551 In BIO_puts, increment b->num_write as in BIO_write.
14553 *Peter.Sylvester@EdelWeb.fr*
14555 * Fix BN_mul_word for the case where the word is 0. (We have to use
14556 BN_zero, we may not return a BIGNUM with an array consisting of
14557 words set to zero.)
14561 * Avoid calling abort() from within the library when problems are
14562 detected, except if preprocessor symbols have been defined
14563 (such as REF_CHECK, BN_DEBUG etc.).
14567 * New openssl application 'rsautl'. This utility can be
14568 used for low level RSA operations. DER public key
14569 BIO/fp routines also added.
14573 * New Configure entry and patches for compiling on QNX 4.
14575 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14577 * A demo state-machine implementation was sponsored by
14578 Nuron (<http://www.nuron.com/>) and is now available in
14579 demos/state_machine.
14583 * New options added to the 'dgst' utility for signature
14584 generation and verification.
14588 * Unrecognized PKCS#7 content types are now handled via a
14589 catch all ASN1_TYPE structure. This allows unsupported
14590 types to be stored as a "blob" and an application can
14591 encode and decode it manually.
14595 * Fix various signed/unsigned issues to make a_strex.c
14596 compile under VC++.
14598 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14600 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14601 length if passed a buffer. ASN1_INTEGER_to_BN failed
14602 if passed a NULL BN and its argument was negative.
14604 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14606 * Modification to PKCS#7 encoding routines to output definite
14607 length encoding. Since currently the whole structures are in
14608 memory there's not real point in using indefinite length
14609 constructed encoding. However if OpenSSL is compiled with
14610 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14614 * Added BIO_vprintf() and BIO_vsnprintf().
14618 * Added more prefixes to parse for in the strings written
14619 through a logging bio, to cover all the levels that are available
14620 through syslog. The prefixes are now:
14622 PANIC, EMERG, EMR => LOG_EMERG
14623 ALERT, ALR => LOG_ALERT
14624 CRIT, CRI => LOG_CRIT
14625 ERROR, ERR => LOG_ERR
14626 WARNING, WARN, WAR => LOG_WARNING
14627 NOTICE, NOTE, NOT => LOG_NOTICE
14628 INFO, INF => LOG_INFO
14629 DEBUG, DBG => LOG_DEBUG
14631 and as before, if none of those prefixes are present at the
14632 beginning of the string, LOG_ERR is chosen.
14634 On Win32, the `LOG_*` levels are mapped according to this:
14636 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14637 LOG_WARNING => EVENTLOG_WARNING_TYPE
14638 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14642 * Made it possible to reconfigure with just the configuration
14643 argument "reconf" or "reconfigure". The command line arguments
14644 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14645 and are retrieved from there when reconfiguring.
14651 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14653 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14657 * The obj_dat.pl script was messing up the sorting of object
14658 names. The reason was that it compared the quoted version
14659 of strings as a result "OCSP" > "OCSP Signing" because
14660 " > SPACE. Changed script to store unquoted versions of
14661 names and add quotes on output. It was also omitting some
14662 names from the lookup table if they were given a default
14663 value (that is if SN is missing it is given the same
14664 value as LN and vice versa), these are now added on the
14665 grounds that if an object has a name we should be able to
14666 look it up. Finally added warning output when duplicate
14667 short or long names are found.
14671 * Changes needed for Tandem NSK.
14673 *Scott Uroff <scott@xypro.com>*
14675 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14676 RSA_padding_check_SSLv23(), special padding was never detected
14677 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14678 version rollback attacks was not effective.
14680 In s23_clnt.c, don't use special rollback-attack detection padding
14681 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14682 client; similarly, in s23_srvr.c, don't do the rollback check if
14683 SSL 2.0 is the only protocol enabled in the server.
14687 * Make it possible to get hexdumps of unprintable data with 'openssl
14688 asn1parse'. By implication, the functions ASN1_parse_dump() and
14689 BIO_dump_indent() are added.
14693 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14694 these print out strings and name structures based on various
14695 flags including RFC2253 support and proper handling of
14696 multibyte characters. Added options to the 'x509' utility
14697 to allow the various flags to be set.
14701 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14702 Also change the functions X509_cmp_current_time() and
14703 X509_gmtime_adj() work with an ASN1_TIME structure,
14704 this will enable certificates using GeneralizedTime in validity
14705 dates to be checked.
14709 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14710 negative public key encodings) on by default,
14711 NO_NEG_PUBKEY_BUG can be set to disable it.
14715 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14716 content octets. An i2c_ASN1_OBJECT is unnecessary because
14717 the encoding can be trivially obtained from the structure.
14721 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14722 not read locks (`CRYPTO_r_[un]lock`).
14726 * A first attempt at creating official support for shared
14727 libraries through configuration. I've kept it so the
14728 default is static libraries only, and the OpenSSL programs
14729 are always statically linked for now, but there are
14730 preparations for dynamic linking in place.
14731 This has been tested on Linux and Tru64.
14735 * Randomness polling function for Win9x, as described in:
14736 Peter Gutmann, Software Generation of Practically Strong
14741 * Fix so PRNG is seeded in req if using an already existing
14746 * New options to smime application. -inform and -outform
14747 allow alternative formats for the S/MIME message including
14748 PEM and DER. The -content option allows the content to be
14749 specified separately. This should allow things like Netscape
14750 form signing output easier to verify.
14754 * Fix the ASN1 encoding of tags using the 'long form'.
14758 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14759 STRING types. These convert content octets to and from the
14760 underlying type. The actual tag and length octets are
14761 already assumed to have been read in and checked. These
14762 are needed because all other string types have virtually
14763 identical handling apart from the tag. By having versions
14764 of the ASN1 functions that just operate on content octets
14765 IMPLICIT tagging can be handled properly. It also allows
14766 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14767 and ASN1_INTEGER are identical apart from the tag.
14771 * Change the handling of OID objects as follows:
14773 - New object identifiers are inserted in objects.txt, following
14774 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14775 - objects.pl is used to process obj_mac.num and create a new
14777 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14780 This is currently kind of a hack, and the perl code in objects.pl
14781 isn't very elegant, but it works as I intended. The simplest way
14782 to check that it worked correctly is to look in obj_dat.h and
14783 check the array nid_objs and make sure the objects haven't moved
14784 around (this is important!). Additions are OK, as well as
14785 consistent name changes.
14789 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14793 * Addition of the command line parameter '-rand file' to 'openssl req'.
14794 The given file adds to whatever has already been seeded into the
14795 random pool through the RANDFILE configuration file option or
14796 environment variable, or the default random state file.
14800 * mkstack.pl now sorts each macro group into lexical order.
14801 Previously the output order depended on the order the files
14802 appeared in the directory, resulting in needless rewriting
14807 * Patches to make OpenSSL compile under Win32 again. Mostly
14808 work arounds for the VC++ problem that it treats func() as
14809 func(void). Also stripped out the parts of mkdef.pl that
14810 added extra typesafe functions: these no longer exist.
14814 * Reorganisation of the stack code. The macros are now all
14815 collected in safestack.h . Each macro is defined in terms of
14816 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14817 DEBUG_SAFESTACK is now handled in terms of function casts,
14818 this has the advantage of retaining type safety without the
14819 use of additional functions. If DEBUG_SAFESTACK is not defined
14820 then the non typesafe macros are used instead. Also modified the
14821 mkstack.pl script to handle the new form. Needs testing to see
14822 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14823 the default if no major problems. Similar behaviour for ASN1_SET_OF
14824 and PKCS12_STACK_OF.
14828 * When some versions of IIS use the 'NET' form of private key the
14829 key derivation algorithm is different. Normally MD5(password) is
14830 used as a 128 bit RC4 key. In the modified case
14831 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14832 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14833 as the old Netscape_RSA functions except they have an additional
14834 'sgckey' parameter which uses the modified algorithm. Also added
14835 an -sgckey command line option to the rsa utility. Thanks to
14836 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14837 algorithm to openssl-dev.
14841 * The evp_local.h macros were using 'c.##kname' which resulted in
14842 invalid expansion on some systems (SCO 5.0.5 for example).
14843 Corrected to 'c.kname'.
14845 *Phillip Porch <root@theporch.com>*
14847 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14848 a STACK of email addresses from a certificate or request, these look
14849 in the subject name and the subject alternative name extensions and
14850 omit any duplicate addresses.
14854 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14855 This makes DSA verification about 2 % faster.
14859 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14860 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14861 plus overhead for 1024 bit moduli).
14862 This makes exponentiations about 0.5 % faster for 1024 bit
14863 exponents (as measured by "openssl speed rsa2048").
14867 * Rename memory handling macros to avoid conflicts with other
14869 Malloc => OPENSSL_malloc
14870 Malloc_locked => OPENSSL_malloc_locked
14871 Realloc => OPENSSL_realloc
14872 Free => OPENSSL_free
14876 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14877 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14881 * CygWin32 support.
14883 *John Jarvie <jjarvie@newsguy.com>*
14885 * The type-safe stack code has been rejigged. It is now only compiled
14886 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14887 by default all type-specific stack functions are "#define"d back to
14888 standard stack functions. This results in more streamlined output
14889 but retains the type-safety checking possibilities of the original
14894 * The STACK code has been cleaned up, and certain type declarations
14895 that didn't make a lot of sense have been brought in line. This has
14896 also involved a cleanup of sorts in safestack.h to more correctly
14897 map type-safe stack functions onto their plain stack counterparts.
14898 This work has also resulted in a variety of "const"ifications of
14899 lots of the code, especially `_cmp` operations which should normally
14900 be prototyped with "const" parameters anyway.
14904 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14905 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14906 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14907 where all of 'md' is used each time the PRNG is used, but 'state'
14908 is used only indexed by a cyclic counter. As entropy may not be
14909 well distributed from the beginning, 'md' is important as a
14910 chaining variable. However, the output function chains only half
14911 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14912 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14913 in all of 'state' being rewritten, with the new values depending
14914 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14918 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14919 the handshake is continued after ssl_verify_cert_chain();
14920 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14921 can lead to 'unexplainable' connection aborts later.
14923 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14925 * Major EVP API cipher revision.
14926 Add hooks for extra EVP features. This allows various cipher
14927 parameters to be set in the EVP interface. Support added for variable
14928 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14929 setting of RC2 and RC5 parameters.
14931 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14934 Remove lots of duplicated code from the EVP library. For example *every*
14935 cipher init() function handles the 'iv' in the same way according to the
14936 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14937 for CFB and OFB modes they zero ctx->num.
14939 New functionality allows removal of S/MIME code RC2 hack.
14941 Most of the routines have the same form and so can be declared in terms
14944 By shifting this to the top level EVP_CipherInit() it can be removed from
14945 all individual ciphers. If the cipher wants to handle IVs or keys
14946 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14949 Change lots of functions like EVP_EncryptUpdate() to now return a
14950 value: although software versions of the algorithms cannot fail
14951 any installed hardware versions can.
14955 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14956 this option is set, tolerate broken clients that send the negotiated
14957 protocol version number instead of the requested protocol version
14962 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14963 i.e. non-zero for export ciphersuites, zero otherwise.
14964 Previous versions had this flag inverted, inconsistent with
14965 rsa_tmp_cb (..._TMP_RSA_CB).
14967 *Bodo Moeller; problem reported by Amit Chopra*
14969 * Add missing DSA library text string. Work around for some IIS
14970 key files with invalid SEQUENCE encoding.
14974 * Add a document (doc/standards.txt) that list all kinds of standards
14975 and so on that are implemented in OpenSSL.
14979 * Enhance c_rehash script. Old version would mishandle certificates
14980 with the same subject name hash and wouldn't handle CRLs at all.
14981 Added -fingerprint option to crl utility, to support new c_rehash
14986 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14990 * Fix for SSL server purpose checking. Server checking was
14991 rejecting certificates which had extended key usage present
14992 but no ssl client purpose.
14994 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14996 * Make PKCS#12 code work with no password. The PKCS#12 spec
14997 is a little unclear about how a blank password is handled.
14998 Since the password in encoded as a BMPString with terminating
14999 double NULL a zero length password would end up as just the
15000 double NULL. However no password at all is different and is
15001 handled differently in the PKCS#12 key generation code. NS
15002 treats a blank password as zero length. MSIE treats it as no
15003 password on export: but it will try both on import. We now do
15004 the same: PKCS12_parse() tries zero length and no password if
15005 the password is set to "" or NULL (NULL is now a valid password:
15006 it wasn't before) as does the pkcs12 application.
15010 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15011 perror when PEM_read_bio_X509_REQ fails, the error message must
15012 be obtained from the error queue.
15016 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15017 it in ERR_remove_state if appropriate, and change ERR_get_state
15018 accordingly to avoid race conditions (this is necessary because
15019 thread_hash is no longer constant once set).
15023 * Bugfix for linux-elf makefile.one.
15027 * RSA_get_default_method() will now cause a default
15028 RSA_METHOD to be chosen if one doesn't exist already.
15029 Previously this was only set during a call to RSA_new()
15030 or RSA_new_method(NULL) meaning it was possible for
15031 RSA_get_default_method() to return NULL.
15035 * Added native name translation to the existing DSO code
15036 that will convert (if the flag to do so is set) filenames
15037 that are sufficiently small and have no path information
15038 into a canonical native form. Eg. "blah" converted to
15039 "libblah.so" or "blah.dll" etc.
15043 * New function ERR_error_string_n(e, buf, len) which is like
15044 ERR_error_string(e, buf), but writes at most 'len' bytes
15045 including the 0 terminator. For ERR_error_string_n, 'buf'
15048 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15050 * CONF library reworked to become more general. A new CONF
15051 configuration file reader "class" is implemented as well as a
15052 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15053 old `CONF_*` functions are still there, but are reimplemented to
15054 work in terms of the new functions. Also, a set of functions
15055 to handle the internal storage of the configuration data is
15056 provided to make it easier to write new configuration file
15057 reader "classes" (I can definitely see something reading a
15058 configuration file in XML format, for example), called `_CONF_*`,
15059 or "the configuration storage API"...
15061 The new configuration file reading functions are:
15063 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15064 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15066 NCONF_default, NCONF_WIN32
15068 NCONF_dump_fp, NCONF_dump_bio
15070 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15071 NCONF_new creates a new CONF object. This works in the same way
15072 as other interfaces in OpenSSL, like the BIO interface.
15073 `NCONF_dump_*` dump the internal storage of the configuration file,
15074 which is useful for debugging. All other functions take the same
15075 arguments as the old `CONF_*` functions with the exception of the
15076 first that must be a `CONF *` instead of a `LHASH *`.
15078 To make it easier to use the new classes with the old `CONF_*` functions,
15079 the function CONF_set_default_method is provided.
15083 * Add '-tls1' option to 'openssl ciphers', which was already
15084 mentioned in the documentation but had not been implemented.
15085 (This option is not yet really useful because even the additional
15086 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15090 * Initial DSO code added into libcrypto for letting OpenSSL (and
15091 OpenSSL-based applications) load shared libraries and bind to
15092 them in a portable way.
15094 *Geoff Thorpe, with contributions from Richard Levitte*
15096 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15098 * Make sure _lrotl and _lrotr are only used with MSVC.
15100 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15101 (the default implementation of RAND_status).
15103 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15104 to '-clrext' (= clear extensions), as intended and documented.
15105 *Bodo Moeller; inconsistency pointed out by Michael Attili
15106 <attili@amaxo.com>*
15108 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15109 was larger than the MD block size.
15111 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15113 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15114 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15115 using the passed key: if the passed key was a private key the result
15116 of X509_print(), for example, would be to print out all the private key
15121 * des_quad_cksum() byte order bug fix.
15122 *Ulf Möller, using the problem description in krb4-0.9.7, where
15123 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15125 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15128 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15130 * For easily testing in shell scripts whether some command
15131 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15132 returns with exit code 0 iff no command of the given name is available.
15133 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15134 the output goes to stdout and nothing is printed to stderr.
15135 Additional arguments are always ignored.
15137 Since for each cipher there is a command of the same name,
15138 the 'no-cipher' compilation switches can be tested this way.
15140 ('openssl no-XXX' is not able to detect pseudo-commands such
15141 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15145 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15149 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15150 is set; it will be thrown away anyway because each handshake creates
15152 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15153 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15154 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15155 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15159 * New s_client option -ign_eof: EOF at stdin is ignored, and
15160 'Q' and 'R' lose their special meanings (quit/renegotiate).
15161 This is part of what -quiet does; unlike -quiet, -ign_eof
15162 does not suppress any output.
15166 * Add compatibility options to the purpose and trust code. The
15167 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15168 accepts a certificate or CA, this was the previous behaviour,
15169 with all the associated security issues.
15171 X509_TRUST_COMPAT is the old trust behaviour: only and
15172 automatically trust self signed roots in certificate store. A
15173 new trust setting X509_TRUST_DEFAULT is used to specify that
15174 a purpose has no associated trust setting and it should instead
15175 use the value in the default purpose.
15179 * Fix the PKCS#8 DSA private key code so it decodes keys again
15180 and fix a memory leak.
15184 * In util/mkerr.pl (which implements 'make errors'), preserve
15185 reason strings from the previous version of the .c file, as
15186 the default to have only downcase letters (and digits) in
15187 automatically generated reasons codes is not always appropriate.
15191 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15192 using strerror. Previously, ERR_reason_error_string() returned
15193 library names as reason strings for SYSerr; but SYSerr is a special
15194 case where small numbers are errno values, not library numbers.
15198 * Add '-dsaparam' option to 'openssl dhparam' application. This
15199 converts DSA parameters into DH parameters. (When creating parameters,
15200 DSA_generate_parameters is used.)
15204 * Include 'length' (recommended exponent length) in C code generated
15205 by 'openssl dhparam -C'.
15209 * The second argument to set_label in perlasm was already being used
15210 so couldn't be used as a "file scope" flag. Moved to third argument
15215 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15216 instead of RAND_bytes for encryption IVs and salts.
15220 * Include RAND_status() into RAND_METHOD instead of implementing
15221 it only for md_rand.c Otherwise replacing the PRNG by calling
15222 RAND_set_rand_method would be impossible.
15226 * Don't let DSA_generate_key() enter an infinite loop if the random
15227 number generation fails.
15231 * New 'rand' application for creating pseudo-random output.
15235 * Added configuration support for Linux/IA64
15237 *Rolf Haberrecker <rolf@suse.de>*
15239 * Assembler module support for Mingw32.
15243 * Shared library support for HPUX (in shlib/).
15245 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15247 * Shared library support for Solaris gcc.
15249 *Lutz Behnke <behnke@trustcenter.de>*
15251 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15253 * PKCS7_encrypt() was adding text MIME headers twice because they
15254 were added manually and by SMIME_crlf_copy().
15258 * In bntest.c don't call BN_rand with zero bits argument.
15260 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15262 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15263 case was implemented. This caused BN_div_recp() to fail occasionally.
15267 * Add an optional second argument to the set_label() in the perl
15268 assembly language builder. If this argument exists and is set
15269 to 1 it signals that the assembler should use a symbol whose
15270 scope is the entire file, not just the current function. This
15271 is needed with MASM which uses the format label:: for this scope.
15273 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15275 * Change the ASN1 types so they are typedefs by default. Before
15276 almost all types were #define'd to ASN1_STRING which was causing
15277 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15282 * Change names of new functions to the new get1/get0 naming
15283 convention: After 'get1', the caller owns a reference count
15284 and has to call `..._free`; 'get0' returns a pointer to some
15285 data structure without incrementing reference counters.
15286 (Some of the existing 'get' functions increment a reference
15287 counter, some don't.)
15288 Similarly, 'set1' and 'add1' functions increase reference
15289 counters or duplicate objects.
15293 * Allow for the possibility of temp RSA key generation failure:
15294 the code used to assume it always worked and crashed on failure.
15298 * Fix potential buffer overrun problem in BIO_printf().
15299 *Ulf Möller, using public domain code by Patrick Powell; problem
15300 pointed out by David Sacerdote <das33@cornell.edu>*
15302 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15303 RAND_egd() and RAND_status(). In the command line application,
15304 the EGD socket can be specified like a seed file using RANDFILE
15309 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15310 Some CAs (e.g. Verisign) distribute certificates in this form.
15314 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15315 list to exclude them. This means that no special compilation option
15316 is needed to use anonymous DH: it just needs to be included in the
15321 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15322 EVP_MD_type. The old functionality is available in a new macro called
15323 EVP_MD_md(). Change code that uses it and update docs.
15327 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15328 where the `void *` argument is replaced by a function pointer argument.
15329 Previously `void *` was abused to point to functions, which works on
15330 many platforms, but is not correct. As these functions are usually
15331 called by macros defined in OpenSSL header files, most source code
15332 should work without changes.
15336 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15337 sections with information on -D... compiler switches used for
15338 compiling the library so that applications can see them. To enable
15339 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15340 must be defined. E.g.,
15341 #define OPENSSL_ALGORITHM_DEFINES
15342 #include <openssl/opensslconf.h>
15343 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15345 *Richard Levitte, Ulf and Bodo Möller*
15347 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15352 * Change the 'other' type in certificate aux info to a STACK_OF
15353 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15354 the required ASN1 format: arbitrary types determined by an OID.
15358 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15359 argument to 'req'. This is not because the function is newer or
15360 better than others it just uses the work 'NEW' in the certificate
15361 request header lines. Some software needs this.
15365 * Reorganise password command line arguments: now passwords can be
15366 obtained from various sources. Delete the PEM_cb function and make
15367 it the default behaviour: i.e. if the callback is NULL and the
15368 usrdata argument is not NULL interpret it as a null terminated pass
15369 phrase. If usrdata and the callback are NULL then the pass phrase
15370 is prompted for as usual.
15374 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15375 the support is automatically enabled. The resulting binaries will
15376 autodetect the card and use it if present.
15378 *Ben Laurie and Compaq Inc.*
15380 * Work around for Netscape hang bug. This sends certificate request
15381 and server done in one record. Since this is perfectly legal in the
15382 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15383 the bugs/SSLv3 entry for more info.
15387 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15391 * Add -rand argument to smime and pkcs12 applications and read/write
15396 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15400 * Add command line password options to the remaining applications.
15404 * Bug fix for BN_div_recp() for numerators with an even number of
15409 * More tests in bntest.c, and changed test_bn output.
15413 * ./config recognizes MacOS X now.
15417 * Bug fix for BN_div() when the first words of num and divisor are
15418 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15422 * Add support for various broken PKCS#8 formats, and command line
15423 options to produce them.
15427 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15428 get temporary BIGNUMs from a BN_CTX.
15432 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15437 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15438 include a #define from the old name to the new. The original intent
15439 was that statically linked binaries could for example just call
15440 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15441 link with digests. This never worked because SSLeay_add_all_digests()
15442 and SSLeay_add_all_ciphers() were in the same source file so calling
15443 one would link with the other. They are now in separate source files.
15447 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15451 * Use a less unusual form of the Miller-Rabin primality test (it used
15452 a binary algorithm for exponentiation integrated into the Miller-Rabin
15453 loop, our standard modexp algorithms are faster).
15457 * Support for the EBCDIC character set completed.
15459 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15461 * Source code cleanups: use const where appropriate, eliminate casts,
15462 use `void *` instead of `char *` in lhash.
15466 * Bugfix: ssl3_send_server_key_exchange was not restartable
15467 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15468 this the server could overwrite ephemeral keys that the client
15473 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15474 using 50 iterations of the Rabin-Miller test.
15476 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15477 iterations of the Rabin-Miller test as required by the appendix
15478 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15479 As BN_is_prime_fasttest includes trial division, DSA parameter
15480 generation becomes much faster.
15482 This implies a change for the callback functions in DSA_is_prime
15483 and DSA_generate_parameters: The callback function is called once
15484 for each positive witness in the Rabin-Miller test, not just
15485 occasionally in the inner loop; and the parameters to the
15486 callback function now provide an iteration count for the outer
15487 loop rather than for the current invocation of the inner loop.
15488 DSA_generate_parameters additionally can call the callback
15489 function with an 'iteration count' of -1, meaning that a
15490 candidate has passed the trial division test (when q is generated
15491 from an application-provided seed, trial division is skipped).
15495 * New function BN_is_prime_fasttest that optionally does trial
15496 division before starting the Rabin-Miller test and has
15497 an additional BN_CTX * argument (whereas BN_is_prime always
15498 has to allocate at least one BN_CTX).
15499 'callback(1, -1, cb_arg)' is called when a number has passed the
15500 trial division stage.
15504 * Fix for bug in CRL encoding. The validity dates weren't being handled
15509 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15513 * New function BN_pseudo_rand().
15517 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15518 bignum version of BN_from_montgomery() with the working code from
15519 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15524 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15525 made it impossible to use the same SSL_SESSION data structure in
15526 SSL2 clients in multiple threads.
15530 * The return value of RAND_load_file() no longer counts bytes obtained
15531 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15532 to seed the PRNG (previously an explicit byte count was required).
15534 *Ulf Möller, Bodo Möller*
15536 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15537 used `char *` instead of `void *` and had casts all over the place.
15541 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15545 * Retain source code compatibility for BN_prime_checks macro:
15546 BN_is_prime(..., BN_prime_checks, ...) now uses
15547 BN_prime_checks_for_size to determine the appropriate number of
15548 Rabin-Miller iterations.
15552 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15553 DH_CHECK_P_NOT_SAFE_PRIME.
15554 (Check if this is true? OpenPGP calls them "strong".)
15558 * Merge the functionality of "dh" and "gendh" programs into a new program
15559 "dhparam". The old programs are retained for now but will handle DH keys
15560 (instead of parameters) in future.
15564 * Make the ciphers, s_server and s_client programs check the return values
15565 when a new cipher list is set.
15569 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15570 ciphers. Before when the 56bit ciphers were enabled the sorting was
15573 The syntax for the cipher sorting has been extended to support sorting by
15574 cipher-strength (using the strength_bits hard coded in the tables).
15575 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15577 Fix a bug in the cipher-command parser: when supplying a cipher command
15578 string with an "undefined" symbol (neither command nor alphanumeric
15579 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15580 an error is flagged.
15582 Due to the strength-sorting extension, the code of the
15583 ssl_create_cipher_list() function was completely rearranged. I hope that
15584 the readability was also increased :-)
15586 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15588 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15589 for the first serial number and places 2 in the serial number file. This
15590 avoids problems when the root CA is created with serial number zero and
15591 the first user certificate has the same issuer name and serial number
15596 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15597 the new code. Add documentation for this stuff.
15601 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15602 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15603 structures and behave in an analogous way to the X509v3 functions:
15604 they shouldn't be called directly but wrapper functions should be used
15607 So we also now have some wrapper functions that call the X509at functions
15608 when passed certificate requests. (TO DO: similar things can be done with
15609 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15610 things. Some of these need some d2i or i2d and print functionality
15611 because they handle more complex structures.)
15615 * Add missing #ifndefs that caused missing symbols when building libssl
15616 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15617 NO_RSA in `ssl/s2*.c`.
15619 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15621 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15622 has a return value which indicates the quality of the random data
15623 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15624 error queue. New function RAND_pseudo_bytes() generates output that is
15625 guaranteed to be unique but not unpredictable. RAND_add is like
15626 RAND_seed, but takes an extra argument for an entropy estimate
15627 (RAND_seed always assumes full entropy).
15631 * Do more iterations of Rabin-Miller probable prime test (specifically,
15632 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15633 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15634 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15635 false-positive rate of at most 2^-80 for random input.
15639 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15643 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15644 in the 0.9.5 release), this returns the chain
15645 from an X509_CTX structure with a dup of the stack and all
15646 the X509 reference counts upped: so the stack will exist
15647 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15650 Also make SSL_SESSION_print() print out the verify return
15655 * Add manpage for the pkcs12 command. Also change the default
15656 behaviour so MAC iteration counts are used unless the new
15657 -nomaciter option is used. This improves file security and
15658 only older versions of MSIE (4.0 for example) need it.
15662 * Honor the no-xxx Configure options when creating .DEF files.
15666 * Add PKCS#10 attributes to field table: challengePassword,
15667 unstructuredName and unstructuredAddress. These are taken from
15668 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15669 international characters are used.
15671 More changes to X509_ATTRIBUTE code: allow the setting of types
15672 based on strings. Remove the 'loc' parameter when adding
15673 attributes because these will be a SET OF encoding which is sorted
15678 * Initial changes to the 'req' utility to allow request generation
15679 automation. This will allow an application to just generate a template
15680 file containing all the field values and have req construct the
15683 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15684 used all over the place including certificate requests and PKCS#7
15685 structures. They are currently handled manually where necessary with
15686 some primitive wrappers for PKCS#7. The new functions behave in a
15687 manner analogous to the X509 extension functions: they allow
15688 attributes to be looked up by NID and added.
15690 Later something similar to the X509V3 code would be desirable to
15691 automatically handle the encoding, decoding and printing of the
15692 more complex types. The string types like challengePassword can
15693 be handled by the string table functions.
15695 Also modified the multi byte string table handling. Now there is
15696 a 'global mask' which masks out certain types. The table itself
15697 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15698 is useful when for example there is only one permissible type
15699 (as in countryName) and using the mask might result in no valid
15704 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15705 SSL_get_peer_finished to allow applications to obtain the latest
15706 Finished messages sent to the peer or expected from the peer,
15707 respectively. (SSL_get_peer_finished is usually the Finished message
15708 actually received from the peer, otherwise the protocol will be aborted.)
15710 As the Finished message are message digests of the complete handshake
15711 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15712 be used for external authentication procedures when the authentication
15713 provided by SSL/TLS is not desired or is not enough.
15717 * Enhanced support for Alpha Linux is added. Now ./config checks if
15718 the host supports BWX extension and if Compaq C is present on the
15719 $PATH. Just exploiting of the BWX extension results in 20-30%
15720 performance kick for some algorithms, e.g. DES and RC4 to mention
15721 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15726 * Add support for MS "fast SGC". This is arguably a violation of the
15727 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15728 weak crypto and after checking the certificate is SGC a second one
15729 with strong crypto. MS SGC stops the first handshake after receiving
15730 the server certificate message and sends a second client hello. Since
15731 a server will typically do all the time consuming operations before
15732 expecting any further messages from the client (server key exchange
15733 is the most expensive) there is little difference between the two.
15735 To get OpenSSL to support MS SGC we have to permit a second client
15736 hello message after we have sent server done. In addition we have to
15737 reset the MAC if we do get this second client hello.
15741 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15742 if a DER encoded private key is RSA or DSA traditional format. Changed
15743 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15744 format DER encoded private key. Newer code should use PKCS#8 format which
15745 has the key type encoded in the ASN1 structure. Added DER private key
15746 support to pkcs8 application.
15750 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15751 ciphersuites has been selected (as required by the SSL 3/TLS 1
15752 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15753 is set, we interpret this as a request to violate the specification
15754 (the worst that can happen is a handshake failure, and 'correct'
15755 behaviour would result in a handshake failure anyway).
15759 * In SSL_CTX_add_session, take into account that there might be multiple
15760 SSL_SESSION structures with the same session ID (e.g. when two threads
15761 concurrently obtain them from an external cache).
15762 The internal cache can handle only one SSL_SESSION with a given ID,
15763 so if there's a conflict, we now throw out the old one to achieve
15768 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15769 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15770 some routines that use cipher OIDs: some ciphers do not have OIDs
15771 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15776 * Simplify the trust setting structure and code. Now we just have
15777 two sequences of OIDs for trusted and rejected settings. These will
15778 typically have values the same as the extended key usage extension
15779 and any application specific purposes.
15781 The trust checking code now has a default behaviour: it will just
15782 check for an object with the same NID as the passed id. Functions can
15783 be provided to override either the default behaviour or the behaviour
15784 for a given id. SSL client, server and email already have functions
15785 in place for compatibility: they check the NID and also return "trusted"
15786 if the certificate is self signed.
15790 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15791 traditional format into an EVP_PKEY structure.
15795 * Add a password callback function PEM_cb() which either prompts for
15796 a password if usr_data is NULL or otherwise assumes it is a null
15797 terminated password. Allow passwords to be passed on command line
15798 environment or config files in a few more utilities.
15802 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15803 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15804 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15805 Update documentation.
15809 * Support for ASN1 "NULL" type. This could be handled before by using
15810 ASN1_TYPE but there wasn't any function that would try to read a NULL
15811 and produce an error if it couldn't. For compatibility we also have
15812 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15813 don't allocate anything because they don't need to.
15817 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15820 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15822 * Rebuild of the memory allocation routines used by OpenSSL code and
15823 possibly others as well. The purpose is to make an interface that
15824 provide hooks so anyone can build a separate set of allocation and
15825 deallocation routines to be used by OpenSSL, for example memory
15826 pool implementations, or something else, which was previously hard
15827 since Malloc(), Realloc() and Free() were defined as macros having
15828 the values malloc, realloc and free, respectively (except for Win32
15829 compilations). The same is provided for memory debugging code.
15830 OpenSSL already comes with functionality to find memory leaks, but
15831 this gives people a chance to debug other memory problems.
15833 With these changes, a new set of functions and macros have appeared:
15835 CRYPTO_set_mem_debug_functions() [F]
15836 CRYPTO_get_mem_debug_functions() [F]
15837 CRYPTO_dbg_set_options() [F]
15838 CRYPTO_dbg_get_options() [F]
15839 CRYPTO_malloc_debug_init() [M]
15841 The memory debug functions are NULL by default, unless the library
15842 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15843 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15844 gives the standard debugging functions that come with OpenSSL) or
15845 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15846 provided by the library user) must be used. When the standard
15847 debugging functions are used, CRYPTO_dbg_set_options can be used to
15848 request additional information:
15849 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15850 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15852 Also, things like CRYPTO_set_mem_functions will always give the
15853 expected result (the new set of functions is used for allocation
15854 and deallocation) at all times, regardless of platform and compiler
15857 To finish it up, some functions that were never use in any other
15858 way than through macros have a new API and new semantic:
15860 CRYPTO_dbg_malloc()
15861 CRYPTO_dbg_realloc()
15864 All macros of value have retained their old syntax.
15866 *Richard Levitte and Bodo Moeller*
15868 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15869 ordering of SMIMECapabilities wasn't in "strength order" and there
15870 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15875 * Some ASN1 types with illegal zero length encoding (INTEGER,
15876 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15878 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15880 * Merge in my S/MIME library for OpenSSL. This provides a simple
15881 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15882 functionality to handle multipart/signed properly) and a utility
15883 called 'smime' to call all this stuff. This is based on code I
15884 originally wrote for Celo who have kindly allowed it to be
15885 included in OpenSSL.
15889 * Add variants des_set_key_checked and des_set_key_unchecked of
15890 des_set_key (aka des_key_sched). Global variable des_check_key
15891 decides which of these is called by des_set_key; this way
15892 des_check_key behaves as it always did, but applications and
15893 the library itself, which was buggy for des_check_key == 1,
15894 have a cleaner way to pick the version they need.
15898 * New function PKCS12_newpass() which changes the password of a
15903 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15904 dynamic mix. In both cases the ids can be used as an index into the
15905 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15906 functions so they accept a list of the field values and the
15907 application doesn't need to directly manipulate the X509_TRUST
15912 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15917 * Modify the way the V3 extension code looks up extensions. This now
15918 works in a similar way to the object code: we have some "standard"
15919 extensions in a static table which is searched with OBJ_bsearch()
15920 and the application can add dynamic ones if needed. The file
15921 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15922 updated whenever a new extension is added to the core code and kept
15923 in ext_nid order. There is a simple program 'tabtest.c' which checks
15924 this. New extensions are not added too often so this file can readily
15925 be maintained manually.
15927 There are two big advantages in doing things this way. The extensions
15928 can be looked up immediately and no longer need to be "added" using
15929 X509V3_add_standard_extensions(): this function now does nothing.
15930 Side note: I get *lots* of email saying the extension code doesn't
15931 work because people forget to call this function.
15932 Also no dynamic allocation is done unless new extensions are added:
15933 so if we don't add custom extensions there is no need to call
15934 X509V3_EXT_cleanup().
15938 * Modify enc utility's salting as follows: make salting the default. Add a
15939 magic header, so unsalted files fail gracefully instead of just decrypting
15940 to garbage. This is because not salting is a big security hole, so people
15941 should be discouraged from doing it.
15945 * Fixes and enhancements to the 'x509' utility. It allowed a message
15946 digest to be passed on the command line but it only used this
15947 parameter when signing a certificate. Modified so all relevant
15948 operations are affected by the digest parameter including the
15949 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15950 DSA key was used because it didn't fix the digest.
15954 * Initial certificate chain verify code. Currently tests the untrusted
15955 certificates for consistency with the verify purpose (which is set
15956 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15958 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15959 this is because it will reject chains with invalid extensions whereas
15960 every previous version of OpenSSL and SSLeay made no checks at all.
15962 Trust code: checks the root CA for the relevant trust settings. Trust
15963 settings have an initial value consistent with the verify purpose: e.g.
15964 if the verify purpose is for SSL client use it expects the CA to be
15965 trusted for SSL client use. However the default value can be changed to
15966 permit custom trust settings: one example of this would be to only trust
15967 certificates from a specific "secure" set of CAs.
15969 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15970 which should be used for version portability: especially since the
15971 verify structure is likely to change more often now.
15973 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15974 to set them. If not set then assume SSL clients will verify SSL servers
15977 Two new options to the verify program: -untrusted allows a set of
15978 untrusted certificates to be passed in and -purpose which sets the
15979 intended purpose of the certificate. If a purpose is set then the
15980 new chain verify code is used to check extension consistency.
15984 * Support for the authority information access extension.
15988 * Modify RSA and DSA PEM read routines to transparently handle
15989 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15990 public keys in a format compatible with certificate
15991 SubjectPublicKeyInfo structures. Unfortunately there were already
15992 functions called *_PublicKey_* which used various odd formats so
15993 these are retained for compatibility: however the DSA variants were
15994 never in a public release so they have been deleted. Changed dsa/rsa
15995 utilities to handle the new format: note no releases ever handled public
15996 keys so we should be OK.
15998 The primary motivation for this change is to avoid the same fiasco
15999 that dogs private keys: there are several incompatible private key
16000 formats some of which are standard and some OpenSSL specific and
16001 require various evil hacks to allow partial transparent handling and
16002 even then it doesn't work with DER formats. Given the option anything
16003 other than PKCS#8 should be dumped: but the other formats have to
16004 stay in the name of compatibility.
16006 With public keys and the benefit of hindsight one standard format
16007 is used which works with EVP_PKEY, RSA or DSA structures: though
16008 it clearly returns an error if you try to read the wrong kind of key.
16010 Added a -pubkey option to the 'x509' utility to output the public key.
16011 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16012 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16013 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16014 that do the same as the `EVP_PKEY_assign_*()` except they up the
16015 reference count of the added key (they don't "swallow" the
16020 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16021 CRLs would fail if the file contained no certificates or no CRLs:
16022 added a new function to read in both types and return the number
16023 read: this means that if none are read it will be an error. The
16024 DER versions of the certificate and CRL reader would always fail
16025 because it isn't possible to mix certificates and CRLs in DER format
16026 without choking one or the other routine. Changed this to just read
16027 a certificate: this is the best we can do. Also modified the code
16028 in apps/verify.c to take notice of return codes: it was previously
16029 attempting to read in certificates from NULL pointers and ignoring
16030 any errors: this is one reason why the cert and CRL reader seemed
16031 to work. It doesn't check return codes from the default certificate
16032 routines: these may well fail if the certificates aren't installed.
16036 * Code to support otherName option in GeneralName.
16040 * First update to verify code. Change the verify utility
16041 so it warns if it is passed a self signed certificate:
16042 for consistency with the normal behaviour. X509_verify
16043 has been modified to it will now verify a self signed
16044 certificate if *exactly* the same certificate appears
16045 in the store: it was previously impossible to trust a
16046 single self signed certificate. This means that:
16047 openssl verify ss.pem
16048 now gives a warning about a self signed certificate but
16049 openssl verify -CAfile ss.pem ss.pem
16054 * For servers, store verify_result in SSL_SESSION data structure
16055 (and add it to external session representation).
16056 This is needed when client certificate verifications fails,
16057 but an application-provided verification callback (set by
16058 SSL_CTX_set_cert_verify_callback) allows accepting the session
16059 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16060 but returns 1): When the session is reused, we have to set
16061 ssl->verify_result to the appropriate error code to avoid
16064 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16066 * Fix a bug in the new PKCS#7 code: it didn't consider the
16067 case in PKCS7_dataInit() where the signed PKCS7 structure
16068 didn't contain any existing data because it was being created.
16070 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16072 * Add a salt to the key derivation routines in enc.c. This
16073 forms the first 8 bytes of the encrypted file. Also add a
16074 -S option to allow a salt to be input on the command line.
16078 * New function X509_cmp(). Oddly enough there wasn't a function
16079 to compare two certificates. We do this by working out the SHA1
16080 hash and comparing that. X509_cmp() will be needed by the trust
16085 * SSL_get1_session() is like SSL_get_session(), but increments
16086 the reference count in the SSL_SESSION returned.
16088 *Geoff Thorpe <geoff@eu.c2.net>*
16090 * Fix for 'req': it was adding a null to request attributes.
16091 Also change the X509_LOOKUP and X509_INFO code to handle
16092 certificate auxiliary information.
16096 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16101 * Add the possibility to add extra information to the memory leak
16102 detecting output, to form tracebacks, showing from where each
16103 allocation was originated: CRYPTO_push_info("constant string") adds
16104 the string plus current file name and line number to a per-thread
16105 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16106 is like calling CYRPTO_pop_info() until the stack is empty.
16107 Also updated memory leak detection code to be multi-thread-safe.
16111 * Add options -text and -noout to pkcs7 utility and delete the
16112 encryption options which never did anything. Update docs.
16116 * Add options to some of the utilities to allow the pass phrase
16117 to be included on either the command line (not recommended on
16118 OSes like Unix) or read from the environment. Update the
16119 manpages and fix a few bugs.
16123 * Add a few manpages for some of the openssl commands.
16127 * Fix the -revoke option in ca. It was freeing up memory twice,
16128 leaking and not finding already revoked certificates.
16132 * Extensive changes to support certificate auxiliary information.
16133 This involves the use of X509_CERT_AUX structure and X509_AUX
16134 functions. An X509_AUX function such as PEM_read_X509_AUX()
16135 can still read in a certificate file in the usual way but it
16136 will also read in any additional "auxiliary information". By
16137 doing things this way a fair degree of compatibility can be
16138 retained: existing certificates can have this information added
16139 using the new 'x509' options.
16141 Current auxiliary information includes an "alias" and some trust
16142 settings. The trust settings will ultimately be used in enhanced
16143 certificate chain verification routines: currently a certificate
16144 can only be trusted if it is self signed and then it is trusted
16149 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16150 The problem was that one of the replacement routines had not been working
16151 since SSLeay releases. For now the offending routine has been replaced
16152 with non-optimised assembler. Even so, this now gives around 95%
16153 performance improvement for 1024 bit RSA signs.
16157 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16158 handling. Most clients have the effective key size in bits equal to
16159 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16160 A few however don't do this and instead use the size of the decrypted key
16161 to determine the RC2 key length and the AlgorithmIdentifier to determine
16162 the effective key length. In this case the effective key length can still
16163 be 40 bits but the key length can be 168 bits for example. This is fixed
16164 by manually forcing an RC2 key into the EVP_PKEY structure because the
16165 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16166 the key length and effective key length are equal.
16170 * Add a bunch of functions that should simplify the creation of
16171 X509_NAME structures. Now you should be able to do:
16172 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16173 and have it automatically work out the correct field type and fill in
16174 the structures. The more adventurous can try:
16175 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16176 and it will (hopefully) work out the correct multibyte encoding.
16180 * Change the 'req' utility to use the new field handling and multibyte
16181 copy routines. Before the DN field creation was handled in an ad hoc
16182 way in req, ca, and x509 which was rather broken and didn't support
16183 BMPStrings or UTF8Strings. Since some software doesn't implement
16184 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16185 using the dirstring_type option. See the new comment in the default
16186 openssl.cnf for more info.
16190 * Make crypto/rand/md_rand.c more robust:
16191 - Assure unique random numbers after fork().
16192 - Make sure that concurrent threads access the global counter and
16193 md serializably so that we never lose entropy in them
16194 or use exactly the same state in multiple threads.
16195 Access to the large state is not always serializable because
16196 the additional locking could be a performance killer, and
16197 md should be large enough anyway.
16201 * New file apps/app_rand.c with commonly needed functionality
16202 for handling the random seed file.
16204 Use the random seed file in some applications that previously did not:
16206 dsaparam -genkey (which also ignored its '-rand' option),
16209 x509 (when signing).
16210 Except on systems with /dev/urandom, it is crucial to have a random
16211 seed file at least for key creation, DSA signing, and for DH exchanges;
16212 for RSA signatures we could do without one.
16214 gendh and gendsa (unlike genrsa) used to read only the first byte
16215 of each file listed in the '-rand' option. The function as previously
16216 found in genrsa is now in app_rand.c and is used by all programs
16217 that support '-rand'.
16221 * In RAND_write_file, use mode 0600 for creating files;
16222 don't just chmod when it may be too late.
16226 * Report an error from X509_STORE_load_locations
16227 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16231 * New function ASN1_mbstring_copy() this copies a string in either
16232 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16233 into an ASN1_STRING type. A mask of permissible types is passed
16234 and it chooses the "minimal" type to use or an error if not type
16239 * Add function equivalents to the various macros in asn1.h. The old
16240 macros are retained with an `M_` prefix. Code inside the library can
16241 use the `M_` macros. External code (including the openssl utility)
16242 should *NOT* in order to be "shared library friendly".
16246 * Add various functions that can check a certificate's extensions
16247 to see if it usable for various purposes such as SSL client,
16248 server or S/MIME and CAs of these types. This is currently
16249 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16250 verification. Also added a -purpose flag to x509 utility to
16251 print out all the purposes.
16255 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16260 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16261 for, obtain and decode and extension and obtain its critical flag.
16262 This allows all the necessary extension code to be handled in a
16263 single function call.
16267 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16268 platforms. See crypto/rc4/rc4_enc.c for further details.
16272 * New -noout option to asn1parse. This causes no output to be produced
16273 its main use is when combined with -strparse and -out to extract data
16274 from a file (which may not be in ASN.1 format).
16278 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16279 when producing the local key id.
16281 *Richard Levitte <levitte@stacken.kth.se>*
16283 * New option -dhparam in s_server. This allows a DH parameter file to be
16284 stated explicitly. If it is not stated then it tries the first server
16285 certificate file. The previous behaviour hard coded the filename
16290 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16291 a public key to be input or output. For example:
16292 openssl rsa -in key.pem -pubout -out pubkey.pem
16293 Also added necessary DSA public key functions to handle this.
16297 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16298 in the message. This was handled by allowing
16299 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16301 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16303 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16304 to the end of the strings whereas this didn't. This would cause problems
16305 if strings read with d2i_ASN1_bytes() were later modified.
16307 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16309 * Fix for base64 decode bug. When a base64 bio reads only one line of
16310 data and it contains EOF it will end up returning an error. This is
16311 caused by input 46 bytes long. The cause is due to the way base64
16312 BIOs find the start of base64 encoded data. They do this by trying a
16313 trial decode on each line until they find one that works. When they
16314 do a flag is set and it starts again knowing it can pass all the
16315 data directly through the decoder. Unfortunately it doesn't reset
16316 the context it uses. This means that if EOF is reached an attempt
16317 is made to pass two EOFs through the context and this causes the
16318 resulting error. This can also cause other problems as well. As is
16319 usual with these problems it takes *ages* to find and the fix is
16320 trivial: move one line.
16322 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16324 * Ugly workaround to get s_client and s_server working under Windows. The
16325 old code wouldn't work because it needed to select() on sockets and the
16326 tty (for keypresses and to see if data could be written). Win32 only
16327 supports select() on sockets so we select() with a 1s timeout on the
16328 sockets and then see if any characters are waiting to be read, if none
16329 are present then we retry, we also assume we can always write data to
16330 the tty. This isn't nice because the code then blocks until we've
16331 received a complete line of data and it is effectively polling the
16332 keyboard at 1s intervals: however it's quite a bit better than not
16333 working at all :-) A dedicated Windows application might handle this
16334 with an event loop for example.
16338 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16339 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16340 will be called when RSA_sign() and RSA_verify() are used. This is useful
16341 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16342 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16343 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16344 This necessitated the support of an extra signature type NID_md5_sha1
16345 for SSL signatures and modifications to the SSL library to use it instead
16346 of calling RSA_public_decrypt() and RSA_private_encrypt().
16350 * Add new -verify -CAfile and -CApath options to the crl program, these
16351 will lookup a CRL issuers certificate and verify the signature in a
16352 similar way to the verify program. Tidy up the crl program so it
16353 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16354 less strict. It will now permit CRL extensions even if it is not
16355 a V2 CRL: this will allow it to tolerate some broken CRLs.
16359 * Initialize all non-automatic variables each time one of the openssl
16360 sub-programs is started (this is necessary as they may be started
16361 multiple times from the "OpenSSL>" prompt).
16363 *Lennart Bang, Bodo Moeller*
16365 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16366 removing all other RSA functionality (this is what NO_RSA does). This
16367 is so (for example) those in the US can disable those operations covered
16368 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16373 * Non-copying interface to BIO pairs.
16374 (still largely untested)
16378 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16379 ASCII string. This was handled independently in various places before.
16383 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16384 UTF8 strings a character at a time.
16388 * Use client_version from client hello to select the protocol
16389 (s23_srvr.c) and for RSA client key exchange verification
16390 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16394 * Add various utility functions to handle SPKACs, these were previously
16395 handled by poking round in the structure internals. Added new function
16396 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16397 print, verify and generate SPKACs. Based on an original idea from
16398 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16402 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16406 * Allow the config file extension section to be overwritten on the
16407 command line. Based on an original idea from Massimiliano Pala
16408 <madwolf@comune.modena.it>. The new option is called -extensions
16409 and can be applied to ca, req and x509. Also -reqexts to override
16410 the request extensions in req and -crlexts to override the crl extensions
16415 * Add new feature to the SPKAC handling in ca. Now you can include
16416 the same field multiple times by preceding it by "XXXX." for example:
16419 this is the same syntax as used in the req config file.
16423 * Allow certificate extensions to be added to certificate requests. These
16424 are specified in a 'req_extensions' option of the req section of the
16425 config file. They can be printed out with the -text option to req but
16426 are otherwise ignored at present.
16430 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16431 data read consists of only the final block it would not decrypted because
16432 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16433 A misplaced 'break' also meant the decrypted final block might not be
16434 copied until the next read.
16438 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16439 a few extra parameters to the DH structure: these will be useful if
16440 for example we want the value of 'q' or implement X9.42 DH.
16444 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16445 provides hooks that allow the default DSA functions or functions on a
16446 "per key" basis to be replaced. This allows hardware acceleration and
16447 hardware key storage to be handled without major modification to the
16448 library. Also added low level modexp hooks and CRYPTO_EX structure and
16449 associated functions.
16453 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16454 as "read only": it can't be written to and the buffer it points to will
16455 not be freed. Reading from a read only BIO is much more efficient than
16456 a normal memory BIO. This was added because there are several times when
16457 an area of memory needs to be read from a BIO. The previous method was
16458 to create a memory BIO and write the data to it, this results in two
16459 copies of the data and an O(n^2) reading algorithm. There is a new
16460 function BIO_new_mem_buf() which creates a read only memory BIO from
16461 an area of memory. Also modified the PKCS#7 routines to use read only
16466 * Bugfix: ssl23_get_client_hello did not work properly when called in
16467 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16468 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16469 but a retry condition occurred while trying to read the rest.
16473 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16474 NID_pkcs7_encrypted by default: this was wrong since this should almost
16475 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16476 the encrypted data type: this is a more sensible place to put it and it
16477 allows the PKCS#12 code to be tidied up that duplicated this
16482 * Changed obj_dat.pl script so it takes its input and output files on
16483 the command line. This should avoid shell escape redirection problems
16488 * Initial support for certificate extension requests, these are included
16489 in things like Xenroll certificate requests. Included functions to allow
16490 extensions to be obtained and added.
16494 * -crlf option to s_client and s_server for sending newlines as
16495 CRLF (as required by many protocols).
16499 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16501 * Install libRSAglue.a when OpenSSL is built with RSAref.
16503 *Ralf S. Engelschall*
16505 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16507 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16509 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16514 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16515 DH parameters/keys (q is lost during that conversion, but the resulting
16516 DH parameters contain its length).
16518 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16519 much faster than DH_generate_parameters (which creates parameters
16520 where `p = 2*q + 1`), and also the smaller q makes DH computations
16521 much more efficient (160-bit exponentiation instead of 1024-bit
16522 exponentiation); so this provides a convenient way to support DHE
16523 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16524 utter importance to use
16525 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16527 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16528 when such DH parameters are used, because otherwise small subgroup
16529 attacks may become possible!
16533 * Avoid memory leak in i2d_DHparams.
16537 * Allow the -k option to be used more than once in the enc program:
16538 this allows the same encrypted message to be read by multiple recipients.
16542 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16543 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16544 it will always use the numerical form of the OID, even if it has a short
16549 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16550 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16551 otherwise bn_mod_exp was called. In the case of hardware keys for example
16552 no private key components need be present and it might store extra data
16553 in the RSA structure, which cannot be accessed from bn_mod_exp.
16554 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16555 private key operations.
16559 * Added support for SPARC Linux.
16563 * pem_password_cb function type incompatibly changed from
16564 typedef int pem_password_cb(char *buf, int size, int rwflag);
16566 ....(char *buf, int size, int rwflag, void *userdata);
16567 so that applications can pass data to their callbacks:
16568 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16569 additional void * argument, which is just handed through whenever
16570 the password callback is called.
16572 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16574 New function SSL_CTX_set_default_passwd_cb_userdata.
16576 Compatibility note: As many C implementations push function arguments
16577 onto the stack in reverse order, the new library version is likely to
16578 interoperate with programs that have been compiled with the old
16579 pem_password_cb definition (PEM_whatever takes some data that
16580 happens to be on the stack as its last argument, and the callback
16581 just ignores this garbage); but there is no guarantee whatsoever that
16584 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16585 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16586 problems not only on Windows, but also on some Unix platforms.
16587 To avoid problematic command lines, these definitions are now in an
16588 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16589 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16593 * MIPS III/IV assembler module is reimplemented.
16597 * More DES library cleanups: remove references to srand/rand and
16598 delete an unused file.
16602 * Add support for the free Netwide assembler (NASM) under Win32,
16603 since not many people have MASM (ml) and it can be hard to obtain.
16604 This is currently experimental but it seems to work OK and pass all
16605 the tests. Check out INSTALL.W32 for info.
16609 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16610 without temporary keys kept an extra copy of the server key,
16611 and connections with temporary keys did not free everything in case
16616 * New function RSA_check_key and new openssl rsa option -check
16617 for verifying the consistency of RSA keys.
16619 *Ulf Moeller, Bodo Moeller*
16621 * Various changes to make Win32 compile work:
16622 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16623 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16624 comparison" warnings.
16625 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16629 * Add a debugging option to PKCS#5 v2 key generation function: when
16630 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16631 derived keys are printed to stderr.
16635 * Copy the flags in ASN1_STRING_dup().
16637 *Roman E. Pavlov <pre@mo.msk.ru>*
16639 * The x509 application mishandled signing requests containing DSA
16640 keys when the signing key was also DSA and the parameters didn't match.
16642 It was supposed to omit the parameters when they matched the signing key:
16643 the verifying software was then supposed to automatically use the CA's
16644 parameters if they were absent from the end user certificate.
16646 Omitting parameters is no longer recommended. The test was also
16647 the wrong way round! This was probably due to unusual behaviour in
16648 EVP_cmp_parameters() which returns 1 if the parameters match.
16649 This meant that parameters were omitted when they *didn't* match and
16650 the certificate was useless. Certificates signed with 'ca' didn't have
16653 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16655 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16656 The interface is as follows:
16657 Applications can use
16658 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16659 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16660 "off" is now the default.
16661 The library internally uses
16662 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16663 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16664 to disable memory-checking temporarily.
16666 Some inconsistent states that previously were possible (and were
16667 even the default) are now avoided.
16669 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16670 with each memory chunk allocated; this is occasionally more helpful
16671 than just having a counter.
16673 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16675 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16680 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16681 which largely parallels "options", but is for changing API behaviour,
16682 whereas "options" are about protocol behaviour.
16683 Initial "mode" flags are:
16685 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16686 a single record has been written.
16687 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16688 retries use the same buffer location.
16689 (But all of the contents must be
16694 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16697 * Fix problems with no-hmac etc.
16699 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16701 * New functions RSA_get_default_method(), RSA_set_method() and
16702 RSA_get_method(). These allows replacement of RSA_METHODs without having
16703 to mess around with the internals of an RSA structure.
16707 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16708 Also really enable memory leak checks in openssl.c and in some
16711 *Chad C. Mulligan, Bodo Moeller*
16713 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16714 up the length of negative integers. This has now been simplified to just
16715 store the length when it is first determined and use it later, rather
16716 than trying to keep track of where data is copied and updating it to
16718 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16720 * Add a new function PKCS7_signatureVerify. This allows the verification
16721 of a PKCS#7 signature but with the signing certificate passed to the
16722 function itself. This contrasts with PKCS7_dataVerify which assumes the
16723 certificate is present in the PKCS#7 structure. This isn't always the
16724 case: certificates can be omitted from a PKCS#7 structure and be
16725 distributed by "out of band" means (such as a certificate database).
16729 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16730 function prototypes in pem.h, also change util/mkdef.pl to add the
16731 necessary function names.
16735 * mk1mf.pl (used by Windows builds) did not properly read the
16736 options set by Configure in the top level Makefile, and Configure
16737 was not even able to write more than one option correctly.
16738 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16742 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16743 file to be loaded from a BIO or FILE pointer. The BIO version will
16744 for example allow memory BIOs to contain config info.
16748 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16749 Whoever hopes to achieve shared-library compatibility across versions
16750 must use this, not the compile-time macro.
16751 (Exercise 0.9.4: Which is the minimum library version required by
16753 Note: All this applies only to multi-threaded programs, others don't
16758 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16759 through a BIO pair triggered the default case, i.e.
16760 SSLerr(...,SSL_R_UNKNOWN_STATE).
16764 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16765 can use the SSL library even if none of the specific BIOs is
16770 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16771 for the encoded length.
16773 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16775 * Add initial documentation of the X509V3 functions.
16779 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16780 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16781 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16782 secure PKCS#8 private key format with a high iteration count.
16786 * Fix determination of Perl interpreter: A perl or perl5
16787 *directory* in $PATH was also accepted as the interpreter.
16789 *Ralf S. Engelschall*
16791 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16792 wrong with it but it was very old and did things like calling
16793 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16794 unusual formatting.
16798 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16799 to use the new extension code.
16803 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16804 with macros. This should make it easier to change their form, add extra
16805 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16810 * Add to configuration table a new entry that can specify an alternative
16811 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16812 according to Mark Crispin <MRC@Panda.COM>.
16816 * DES CBC did not update the IV. Weird.
16820 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16821 Changing the behaviour of the former might break existing programs --
16822 where IV updating is needed, des_ncbc_encrypt can be used.
16825 * When bntest is run from "make test" it drives bc to check its
16826 calculations, as well as internally checking them. If an internal check
16827 fails, it needs to cause bc to give a non-zero result or make test carries
16828 on without noticing the failure. Fixed.
16832 * DES library cleanups.
16836 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16837 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16838 ciphers. NOTE: although the key derivation function has been verified
16839 against some published test vectors it has not been extensively tested
16840 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16845 * Instead of "mkdir -p", which is not fully portable, use new
16846 Perl script "util/mkdir-p.pl".
16850 * Rewrite the way password based encryption (PBE) is handled. It used to
16851 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16852 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16853 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16854 the 'parameter' field of the AlgorithmIdentifier is passed to the
16855 underlying key generation function so it must do its own ASN1 parsing.
16856 This has also changed the EVP_PBE_CipherInit() function which now has a
16857 'parameter' argument instead of literal salt and iteration count values
16858 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16862 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16863 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16864 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16865 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16866 value was just used as a "magic string" and not used directly its
16867 value doesn't matter.
16871 * Introduce some semblance of const correctness to BN. Shame C doesn't
16876 * "linux-sparc64" configuration (ultrapenguin).
16878 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16879 "linux-sparc" configuration.
16881 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16883 * config now generates no-xxx options for missing ciphers.
16887 * Support the EBCDIC character set (work in progress).
16888 File ebcdic.c not yet included because it has a different license.
16890 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16892 * Support BS2000/OSD-POSIX.
16894 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16896 * Make callbacks for key generation use `void *` instead of `char *`.
16900 * Make S/MIME samples compile (not yet tested).
16904 * Additional typesafe stacks.
16908 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16912 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16914 * New configuration variant "sco5-gcc".
16916 * Updated some demos.
16918 *Sean O Riordain, Wade Scholine*
16920 * Add missing BIO_free at exit of pkcs12 application.
16924 * Fix memory leak in conf.c.
16928 * Updates for Win32 to assembler version of MD5.
16932 * Set #! path to perl in apps/der_chop to where we found it
16933 instead of using a fixed path.
16937 * SHA library changes for irix64-mips4-cc.
16941 * Improvements for VMS support.
16945 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16947 * Bignum library bug fix. IRIX 6 passes "make test" now!
16948 This also avoids the problems with SC4.2 and unpatched SC5.
16950 *Andy Polyakov <appro@fy.chalmers.se>*
16952 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16953 These are required because of the typesafe stack would otherwise break
16954 existing code. If old code used a structure member which used to be STACK
16955 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16956 sk_num or sk_value it would produce an error because the num, data members
16957 are not present in STACK_OF. Now it just produces a warning. sk_set
16958 replaces the old method of assigning a value to sk_value
16959 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16960 that does this will no longer work (and should use sk_set instead) but
16961 this could be regarded as a "questionable" behaviour anyway.
16965 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16966 correctly handle encrypted S/MIME data.
16970 * Change type of various DES function arguments from des_cblock
16971 (which means, in function argument declarations, pointer to char)
16972 to des_cblock * (meaning pointer to array with 8 char elements),
16973 which allows the compiler to do more typechecking; it was like
16974 that back in SSLeay, but with lots of ugly casts.
16976 Introduce new type const_des_cblock.
16980 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16981 problems: find RecipientInfo structure that matches recipient certificate
16982 and initialise the ASN1 structures properly based on passed cipher.
16986 * Belatedly make the BN tests actually check the results.
16990 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16991 to and from BNs: it was completely broken. New compilation option
16992 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16993 key elements as negative integers.
16997 * Reorganize and speed up MD5.
16999 *Andy Polyakov <appro@fy.chalmers.se>*
17003 *Richard Levitte <richard@levitte.org>*
17005 * New option -out to asn1parse to allow the parsed structure to be
17006 output to a file. This is most useful when combined with the -strparse
17007 option to examine the output of things like OCTET STRINGS.
17011 * Make SSL library a little more fool-proof by not requiring any longer
17012 that `SSL_set_{accept,connect}_state` be called before
17013 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17014 in many applications because usually everything *appeared* to work as
17015 intended anyway -- now it really works as intended).
17019 * Move openssl.cnf out of lib/.
17023 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17024 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17025 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17027 *Ralf S. Engelschall*
17029 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17030 handle PKCS#7 enveloped data properly.
17032 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17034 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17035 copying pointers. The cert_st handling is changed by this in
17036 various ways (and thus what used to be known as ctx->default_cert
17037 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17038 any longer when s->cert does not give us what we need).
17039 ssl_cert_instantiate becomes obsolete by this change.
17040 As soon as we've got the new code right (possibly it already is?),
17041 we have solved a couple of bugs of the earlier code where s->cert
17042 was used as if it could not have been shared with other SSL structures.
17044 Note that using the SSL API in certain dirty ways now will result
17045 in different behaviour than observed with earlier library versions:
17046 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17047 does not influence s as it used to.
17049 In order to clean up things more thoroughly, inside SSL_SESSION
17050 we don't use CERT any longer, but a new structure SESS_CERT
17051 that holds per-session data (if available); currently, this is
17052 the peer's certificate chain and, for clients, the server's certificate
17053 and temporary key. CERT holds only those values that can have
17054 meaningful defaults in an SSL_CTX.
17058 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17059 from the internal representation. Various PKCS#7 fixes: remove some
17060 evil casts and set the enc_dig_alg field properly based on the signing
17065 * Allow PKCS#12 password to be set from the command line or the
17066 environment. Let 'ca' get its config file name from the environment
17067 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17072 * Allow certificate policies extension to use an IA5STRING for the
17073 organization field. This is contrary to the PKIX definition but
17074 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17079 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17080 without disallowing inline assembler and the like for non-pedantic builds.
17084 * Support Borland C++ builder.
17086 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17092 * SHA-1 cleanups and performance enhancements.
17094 *Andy Polyakov <appro@fy.chalmers.se>*
17096 * Sparc v8plus assembler for the bignum library.
17098 *Andy Polyakov <appro@fy.chalmers.se>*
17100 * Accept any -xxx and +xxx compiler options in Configure.
17104 * Update HPUX configuration.
17108 * Add missing `sk_<type>_unshift()` function to safestack.h
17110 *Ralf S. Engelschall*
17112 * New function SSL_CTX_use_certificate_chain_file that sets the
17113 "extra_cert"s in addition to the certificate. (This makes sense
17114 only for "PEM" format files, as chains as a whole are not
17119 * Support verify_depth from the SSL API.
17120 x509_vfy.c had what can be considered an off-by-one-error:
17121 Its depth (which was not part of the external interface)
17122 was actually counting the number of certificates in a chain;
17123 now it really counts the depth.
17127 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17128 instead of X509err, which often resulted in confusing error
17129 messages since the error codes are not globally unique
17130 (e.g. an alleged error in ssl3_accept when a certificate
17131 didn't match the private key).
17133 * New function SSL_CTX_set_session_id_context that allows to set a default
17134 value (so that you don't need SSL_set_session_id_context for each
17135 connection using the SSL_CTX).
17139 * OAEP decoding bug fix.
17143 * Support INSTALL_PREFIX for package builders, as proposed by
17148 * New Configure options "threads" and "no-threads". For systems
17149 where the proper compiler options are known (currently Solaris
17150 and Linux), "threads" is the default.
17154 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17158 * Install various scripts to $(OPENSSLDIR)/misc, not to
17159 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17160 such as /usr/local/bin.
17164 * "make linux-shared" to build shared libraries.
17166 *Niels Poppe <niels@netbox.org>*
17168 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17172 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17173 extension adding in x509 utility.
17177 * Remove NOPROTO sections and error code comments.
17181 * Partial rewrite of the DEF file generator to now parse the ANSI
17186 * New Configure options --prefix=DIR and --openssldir=DIR.
17190 * Complete rewrite of the error code script(s). It is all now handled
17191 by one script at the top level which handles error code gathering,
17192 header rewriting and C source file generation. It should be much better
17193 than the old method: it now uses a modified version of Ulf's parser to
17194 read the ANSI prototypes in all header files (thus the old K&R definitions
17195 aren't needed for error creation any more) and do a better job of
17196 translating function codes into names. The old 'ASN1 error code embedded
17197 in a comment' is no longer necessary and it doesn't use .err files which
17198 have now been deleted. Also the error code call doesn't have to appear all
17199 on one line (which resulted in some large lines...).
17203 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17207 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17208 0 (which usually indicates a closed connection), but continue reading.
17212 * Fix some race conditions.
17216 * Add support for CRL distribution points extension. Add Certificate
17217 Policies and CRL distribution points documentation.
17221 * Move the autogenerated header file parts to crypto/opensslconf.h.
17225 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17226 8 of keying material. Merlin has also confirmed interop with this fix
17227 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17229 *Merlin Hughes <merlin@baltimore.ie>*
17231 * Fix lots of warnings.
17233 *Richard Levitte <levitte@stacken.kth.se>*
17235 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17236 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17238 *Richard Levitte <levitte@stacken.kth.se>*
17240 * Fix problems with sizeof(long) == 8.
17242 *Andy Polyakov <appro@fy.chalmers.se>*
17244 * Change functions to ANSI C.
17248 * Fix typos in error codes.
17250 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17252 * Remove defunct assembler files from Configure.
17256 * SPARC v8 assembler BIGNUM implementation.
17258 *Andy Polyakov <appro@fy.chalmers.se>*
17260 * Support for Certificate Policies extension: both print and set.
17261 Various additions to support the r2i method this uses.
17265 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17266 return a const string when you are expecting an allocated buffer.
17270 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17271 types DirectoryString and DisplayText.
17275 * Add code to allow r2i extensions to access the configuration database,
17276 add an LHASH database driver and add several ctx helper functions.
17280 * Fix an evil bug in bn_expand2() which caused various BN functions to
17281 fail when they extended the size of a BIGNUM.
17285 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17286 support typesafe stack.
17290 * Fix typo in SSL_[gs]et_options().
17292 *Nils Frostberg <nils@medcom.se>*
17294 * Delete various functions and files that belonged to the (now obsolete)
17295 old X509V3 handling code.
17299 * New Configure option "rsaref".
17303 * Don't auto-generate pem.h.
17307 * Introduce type-safe ASN.1 SETs.
17311 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17313 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17315 * Introduce type-safe STACKs. This will almost certainly break lots of code
17316 that links with OpenSSL (well at least cause lots of warnings), but fear
17317 not: the conversion is trivial, and it eliminates loads of evil casts. A
17318 few STACKed things have been converted already. Feel free to convert more.
17319 In the fullness of time, I'll do away with the STACK type altogether.
17323 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17324 specified in `<certfile>` by updating the entry in the index.txt file.
17325 This way one no longer has to edit the index.txt file manually for
17326 revoking a certificate. The -revoke option does the gory details now.
17328 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17330 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17331 `-text` option at all and this way the `-noout -text` combination was
17332 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17334 *Ralf S. Engelschall*
17336 * Make sure a corresponding plain text error message exists for the
17337 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17338 verify callback function determined that a certificate was revoked.
17340 *Ralf S. Engelschall*
17342 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17343 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17344 all available ciphers including rc5, which was forgotten until now.
17345 In order to let the testing shell script know which algorithms
17346 are available, a new (up to now undocumented) command
17347 `openssl list-cipher-commands` is used.
17351 * Bugfix: s_client occasionally would sleep in select() when
17352 it should have checked SSL_pending() first.
17356 * New functions DSA_do_sign and DSA_do_verify to provide access to
17357 the raw DSA values prior to ASN.1 encoding.
17361 * Tweaks to Configure
17363 *Niels Poppe <niels@netbox.org>*
17365 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17370 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17374 * New config option to avoid instructions that are illegal on the 80386.
17375 The default code is faster, but requires at least a 486.
17379 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17380 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17381 same as SSL2_VERSION anyway.
17385 * New "-showcerts" option for s_client.
17389 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17390 application. Various cleanups and fixes.
17394 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17395 modify error routines to work internally. Add error codes and PBE init
17396 to library startup routines.
17400 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17401 packing functions to asn1 and evp. Changed function names and error
17402 codes along the way.
17406 * PKCS12 integration: and so it begins... First of several patches to
17407 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17408 objects to objects.h
17412 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17413 and display support for Thawte strong extranet extension.
17417 * Add LinuxPPC support.
17419 *Jeff Dubrule <igor@pobox.org>*
17421 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17422 bn_div_words in alpha.s.
17424 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17426 * Make sure the RSA OAEP test is skipped under -DRSAref because
17427 OAEP isn't supported when OpenSSL is built with RSAref.
17429 *Ulf Moeller <ulf@fitug.de>*
17431 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17432 so they no longer are missing under -DNOPROTO.
17434 *Soren S. Jorvang <soren@t.dk>*
17436 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17438 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17439 doesn't work when the session is reused. Coming soon!
17443 * Fix a security hole, that allows sessions to be reused in the wrong
17444 context thus bypassing client cert protection! All software that uses
17445 client certs and session caches in multiple contexts NEEDS PATCHING to
17446 allow session reuse! A fuller solution is in the works.
17448 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17450 * Some more source tree cleanups (removed obsolete files
17451 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17452 permission on "config" script to be executable) and a fix for the INSTALL
17455 *Ulf Moeller <ulf@fitug.de>*
17457 * Remove some legacy and erroneous uses of malloc, free instead of
17460 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17462 * Make rsa_oaep_test return non-zero on error.
17464 *Ulf Moeller <ulf@fitug.de>*
17466 * Add support for native Solaris shared libraries. Configure
17467 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17468 if someone would make that last step automatic.
17470 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17472 * ctx_size was not built with the right compiler during "make links". Fixed.
17476 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17477 except NULL ciphers". This means the default cipher list will no longer
17478 enable NULL ciphers. They need to be specifically enabled e.g. with
17479 the string "DEFAULT:eNULL".
17483 * Fix to RSA private encryption routines: if p < q then it would
17484 occasionally produce an invalid result. This will only happen with
17485 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17489 * Be less restrictive and allow also `perl util/perlpath.pl
17490 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17491 because this way one can also use an interpreter named `perl5` (which is
17492 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17493 installed as `perl`).
17495 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17497 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17499 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17501 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17502 advapi32.lib to Win32 build and change the pem test comparison
17503 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17504 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17505 and crypto/des/ede_cbcm_enc.c.
17509 * DES quad checksum was broken on big-endian architectures. Fixed.
17513 * Comment out two functions in bio.h that aren't implemented. Fix up the
17514 Win32 test batch file so it (might) work again. The Win32 test batch file
17515 is horrible: I feel ill....
17519 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17520 in e_os.h. Audit of header files to check ANSI and non ANSI
17521 sections: 10 functions were absent from non ANSI section and not exported
17522 from Windows DLLs. Fixed up libeay.num for new functions.
17526 * Make `openssl version` output lines consistent.
17528 *Ralf S. Engelschall*
17530 * Fix Win32 symbol export lists for BIO functions: Added
17531 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17532 to ms/libeay{16,32}.def.
17534 *Ralf S. Engelschall*
17536 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17537 fine under Unix and passes some trivial tests I've now added. But the
17538 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17539 added to make sure no one expects that this stuff really works in the
17540 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17541 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17544 *Ralf S. Engelschall*
17546 * Fix the generation of two part addresses in perl.
17548 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17550 * Add config entry for Linux on MIPS.
17552 *John Tobey <jtobey@channel1.com>*
17554 * Make links whenever Configure is run, unless we are on Windoze.
17558 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17559 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17564 * Add a useful kludge to allow package maintainers to specify compiler and
17565 other platforms details on the command line without having to patch the
17566 Configure script every time: One now can use
17567 `perl Configure <id>:<details>`,
17568 i.e. platform ids are allowed to have details appended
17569 to them (separated by colons). This is treated as there would be a static
17570 pre-configured entry in Configure's %table under key `<id>` with value
17571 `<details>` and `perl Configure <id>` is called. So, when you want to
17572 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17573 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17574 now, which overrides the FreeBSD-elf entry on-the-fly.
17576 *Ralf S. Engelschall*
17578 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17582 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17583 on the `perl Configure ...` command line. This way one can compile
17584 OpenSSL libraries with Position Independent Code (PIC) which is needed
17585 for linking it into DSOs.
17587 *Ralf S. Engelschall*
17589 * Remarkably, export ciphers were totally broken and no-one had noticed!
17594 * Cleaned up the LICENSE document: The official contact for any license
17595 questions now is the OpenSSL core team under openssl-core@openssl.org.
17596 And add a paragraph about the dual-license situation to make sure people
17597 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17598 to the OpenSSL toolkit.
17600 *Ralf S. Engelschall*
17602 * General source tree makefile cleanups: Made `making xxx in yyy...`
17603 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17604 Additionally cleaned up the `make links` target: Remove unnecessary
17605 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17606 to speed processing and no longer clutter the display with confusing
17607 stuff. Instead only the actually done links are displayed.
17609 *Ralf S. Engelschall*
17611 * Permit null encryption ciphersuites, used for authentication only. It used
17612 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17613 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17618 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17619 signed attributes when verifying signatures (this would break them),
17620 the detached data encoding was wrong and public keys obtained using
17621 X509_get_pubkey() weren't freed.
17625 * Add text documentation for the BUFFER functions. Also added a work around
17626 to a Win95 console bug. This was triggered by the password read stuff: the
17627 last character typed gets carried over to the next fread(). If you were
17628 generating a new cert request using 'req' for example then the last
17629 character of the passphrase would be CR which would then enter the first
17634 * Added the new 'Includes OpenSSL Cryptography Software' button as
17635 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17636 button and can be used by applications based on OpenSSL to show the
17637 relationship to the OpenSSL project.
17639 *Ralf S. Engelschall*
17641 * Remove confusing variables in function signatures in files
17642 ssl/ssl_lib.c and ssl/ssl.h.
17644 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17646 * Don't install bss_file.c under PREFIX/include/
17648 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17650 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17651 functions that return function pointers and has support for NT specific
17652 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17653 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17654 unsigned to signed types: this was killing the Win32 compile.
17658 * Add new certificate file to stack functions,
17659 SSL_add_dir_cert_subjects_to_stack() and
17660 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17661 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17662 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17663 This means that Apache-SSL and similar packages don't have to mess around
17664 to add as many CAs as they want to the preferred list.
17668 * Experiment with doxygen documentation. Currently only partially applied to
17670 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17671 openssl.doxy as the configuration file.
17675 * Get rid of remaining C++-style comments which strict C compilers hate.
17677 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17679 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17680 compiled in by default: it has problems with large keys.
17684 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17685 DH private keys and/or callback functions which directly correspond to
17686 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17687 is needed for applications which have to configure certificates on a
17688 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17690 For the RSA certificate situation is makes no difference, but
17691 for the DSA certificate situation this fixes the "no shared cipher"
17692 problem where the OpenSSL cipher selection procedure failed because the
17693 temporary keys were not overtaken from the context and the API provided
17694 no way to reconfigure them.
17695 The new functions now let applications reconfigure the stuff and they
17696 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17697 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17698 non-public-API function ssl_cert_instantiate() is used as a helper
17699 function and also to reduce code redundancy inside ssl_rsa.c.
17701 *Ralf S. Engelschall*
17703 * Move s_server -dcert and -dkey options out of the undocumented feature
17704 area because they are useful for the DSA situation and should be
17705 recognized by the users.
17707 *Ralf S. Engelschall*
17709 * Fix the cipher decision scheme for export ciphers: the export bits are
17710 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17711 SSL_EXP_MASK. So, the original variable has to be used instead of the
17712 already masked variable.
17714 *Richard Levitte <levitte@stacken.kth.se>*
17716 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17718 *Richard Levitte <levitte@stacken.kth.se>*
17720 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17721 from `int` to `unsigned int` because it is a length and initialized by
17722 EVP_DigestFinal() which expects an `unsigned int *`.
17724 *Richard Levitte <levitte@stacken.kth.se>*
17726 * Don't hard-code path to Perl interpreter on shebang line of Configure
17727 script. Instead use the usual Shell->Perl transition trick.
17729 *Ralf S. Engelschall*
17731 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17732 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17733 -noout -modulus` as it's already the case for `openssl rsa -noout
17734 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17735 currently the public key is printed (a decision which was already done by
17736 `openssl dsa -modulus` in the past) which serves a similar purpose.
17737 Additionally the NO_RSA no longer completely removes the whole -modulus
17738 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17741 *Ralf S. Engelschall*
17743 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17744 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17746 *Arne Ansper <arne@ats.cyber.ee>*
17748 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17749 to be added. Now both 'req' and 'ca' can use new objects defined in the
17754 * Add cool BIO that does syslog (or event log on NT).
17756 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17758 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17759 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17760 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17761 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17765 * Add preliminary config info for new extension code.
17769 * Make RSA_NO_PADDING really use no padding.
17771 *Ulf Moeller <ulf@fitug.de>*
17773 * Generate errors when private/public key check is done.
17777 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17778 for some CRL extensions and new objects added.
17782 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17783 key usage extension and fuller support for authority key id.
17787 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17788 padding method for RSA, which is recommended for new applications in PKCS
17789 #1 v2.0 (RFC 2437, October 1998).
17790 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17791 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17792 against Bleichbacher's attack on RSA.
17793 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17796 * Updates to the new SSL compression code
17798 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17800 * Fix so that the version number in the master secret, when passed
17801 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17802 (because the server will not accept higher), that the version number
17803 is 0x03,0x01, not 0x03,0x00
17805 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17807 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17808 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17809 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17813 * Support for RAW extensions where an arbitrary extension can be
17814 created by including its DER encoding. See apps/openssl.cnf for
17819 * Make sure latest Perl versions don't interpret some generated C array
17820 code as Perl array code in the crypto/err/err_genc.pl script.
17822 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17824 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17825 not many people have the assembler. Various Win32 compilation fixes and
17826 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17827 build instructions.
17831 * Modify configure script 'Configure' to automatically create crypto/date.h
17832 file under Win32 and also build pem.h from pem.org. New script
17833 util/mkfiles.pl to create the MINFO file on environments that can't do a
17834 'make files': perl util/mkfiles.pl >MINFO should work.
17838 * Major rework of DES function declarations, in the pursuit of correctness
17839 and purity. As a result, many evil casts evaporated, and some weirdness,
17840 too. You may find this causes warnings in your code. Zapping your evil
17841 casts will probably fix them. Mostly.
17845 * Fix for a typo in asn1.h. Bug fix to object creation script
17846 obj_dat.pl. It considered a zero in an object definition to mean
17847 "end of object": none of the objects in objects.h have any zeros
17848 so it wasn't spotted.
17850 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17852 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17853 Masking (CBCM). In the absence of test vectors, the best I have been able
17854 to do is check that the decrypt undoes the encrypt, so far. Send me test
17855 vectors if you have them.
17859 * Correct calculation of key length for export ciphers (too much space was
17860 allocated for null ciphers). This has not been tested!
17864 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17865 message is now correct (it understands "crypto" and "ssl" on its
17866 command line). There is also now an "update" option. This will update
17867 the util/ssleay.num and util/libeay.num files with any new functions.
17869 perl util/mkdef.pl crypto ssl update
17870 it will update them.
17874 * Overhauled the Perl interface:
17875 - ported BN stuff to OpenSSL's different BN library
17876 - made the perl/ source tree CVS-aware
17877 - renamed the package from SSLeay to OpenSSL (the files still contain
17878 their history because I've copied them in the repository)
17879 - removed obsolete files (the test scripts will be replaced
17880 by better Test::Harness variants in the future)
17882 *Ralf S. Engelschall*
17884 * First cut for a very conservative source tree cleanup:
17885 1. merge various obsolete readme texts into doc/ssleay.txt
17886 where we collect the old documents and readme texts.
17887 2. remove the first part of files where I'm already sure that we no
17888 longer need them because of three reasons: either they are just temporary
17889 files which were left by Eric or they are preserved original files where
17890 I've verified that the diff is also available in the CVS via "cvs diff
17891 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17892 the crypto/md/ stuff).
17894 *Ralf S. Engelschall*
17896 * More extension code. Incomplete support for subject and issuer alt
17897 name, issuer and authority key id. Change the i2v function parameters
17898 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17899 what that's for :-) Fix to ASN1 macro which messed up
17900 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17904 * Preliminary support for ENUMERATED type. This is largely copied from the
17909 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17911 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17913 * Make sure `make rehash` target really finds the `openssl` program.
17915 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17917 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17918 like to hear about it if this slows down other processors.
17922 * Add CygWin32 platform information to Configure script.
17924 *Alan Batie <batie@aahz.jf.intel.com>*
17926 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17928 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17930 * New program nseq to manipulate netscape certificate sequences
17934 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17939 * Fixes to BN code. Previously the default was to define BN_RECURSION
17940 but the BN code had some problems that would cause failures when
17941 doing certificate verification and some other functions.
17943 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17945 * Add ASN1 and PEM code to support netscape certificate sequences.
17949 * Add ASN1 and PEM code to support netscape certificate sequences.
17953 * Add several PKIX and private extended key usage OIDs.
17957 * Modify the 'ca' program to handle the new extension code. Modify
17958 openssl.cnf for new extension format, add comments.
17962 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17963 and add a sample to openssl.cnf so req -x509 now adds appropriate
17968 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17969 error code, add initial support to X509_print() and x509 application.
17973 * Takes a deep breath and start adding X509 V3 extension support code. Add
17974 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17975 stuff is currently isolated and isn't even compiled yet.
17979 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17980 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17981 Removed the versions check from X509 routines when loading extensions:
17982 this allows certain broken certificates that don't set the version
17983 properly to be processed.
17987 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17988 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17989 can still be regenerated with "make depend".
17993 * Spelling mistake in C version of CAST-128.
17995 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17997 * Changes to the error generation code. The perl script err-code.pl
17998 now reads in the old error codes and retains the old numbers, only
17999 adding new ones if necessary. It also only changes the .err files if new
18000 codes are added. The makefiles have been modified to only insert errors
18001 when needed (to avoid needlessly modifying header files). This is done
18002 by only inserting errors if the .err file is newer than the auto generated
18003 C file. To rebuild all the error codes from scratch (the old behaviour)
18004 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18005 or delete all the .err files.
18009 * CAST-128 was incorrectly implemented for short keys. The C version has
18010 been fixed, but is untested. The assembler versions are also fixed, but
18011 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18012 to regenerate it if needed.
18013 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18014 Hagino <itojun@kame.net>*
18016 * File was opened incorrectly in randfile.c.
18018 *Ulf Möller <ulf@fitug.de>*
18020 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18021 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18022 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18023 al: it's just almost always a UTCTime. Note this patch adds new error
18024 codes so do a "make errors" if there are problems.
18028 * Correct Linux 1 recognition in config.
18030 *Ulf Möller <ulf@fitug.de>*
18032 * Remove pointless MD5 hash when using DSA keys in ca.
18034 *Anonymous <nobody@replay.com>*
18036 * Generate an error if given an empty string as a cert directory. Also
18037 generate an error if handed NULL (previously returned 0 to indicate an
18038 error, but didn't set one).
18040 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18042 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18046 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18047 parameters. This was causing a warning which killed off the Win32 compile.
18051 * Remove C++ style comments from crypto/bn/bn_local.h.
18053 *Neil Costigan <neil.costigan@celocom.com>*
18055 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18056 based on a text string, looking up short and long names and finally
18057 "dot" format. The "dot" format stuff didn't work. Added new function
18058 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18059 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18060 OID is not part of the table.
18064 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18065 X509_LOOKUP_by_alias().
18069 * Sort openssl functions by name.
18073 * Get the gendsa program working (hopefully) and add it to app list. Remove
18074 encryption from sample DSA keys (in case anyone is interested the password
18079 * Make *all* `*_free` functions accept a NULL pointer.
18081 *Frans Heymans <fheymans@isaserver.be>*
18083 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18086 *Anonymous <nobody@replay.com>*
18088 * s_server should send the CAfile as acceptable CAs, not its own cert.
18090 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18092 * Don't blow it for numeric -newkey arguments to apps/req.
18094 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18096 * Temp key "for export" tests were wrong in s3_srvr.c.
18098 *Anonymous <nobody@replay.com>*
18100 * Add prototype for temp key callback functions
18101 SSL_CTX_set_tmp_{rsa,dh}_callback().
18105 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18106 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18110 * X509_name_add_entry() freed the wrong thing after an error.
18112 *Arne Ansper <arne@ats.cyber.ee>*
18114 * rsa_eay.c would attempt to free a NULL context.
18116 *Arne Ansper <arne@ats.cyber.ee>*
18118 * BIO_s_socket() had a broken should_retry() on Windoze.
18120 *Arne Ansper <arne@ats.cyber.ee>*
18122 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18124 *Arne Ansper <arne@ats.cyber.ee>*
18126 * Make sure the already existing X509_STORE->depth variable is initialized
18127 in X509_STORE_new(), but document the fact that this variable is still
18128 unused in the certificate verification process.
18130 *Ralf S. Engelschall*
18132 * Fix the various library and apps files to free up pkeys obtained from
18133 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18137 * Fix reference counting in X509_PUBKEY_get(). This makes
18138 demos/maurice/example2.c work, amongst others, probably.
18140 *Steve Henson and Ben Laurie*
18142 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18143 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18144 are no longer created. This way we have a single and consistent command
18145 line interface `openssl <command>`, similar to `cvs <command>`.
18147 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18149 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18150 BIT STRING wrapper always have zero unused bits.
18154 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18158 * Make the top-level INSTALL documentation easier to understand.
18162 * Makefiles updated to exit if an error occurs in a sub-directory
18163 make (including if user presses ^C) [Paul Sutton]
18165 * Make Montgomery context stuff explicit in RSA data structure.
18169 * Fix build order of pem and err to allow for generated pem.h.
18173 * Fix renumbering bug in X509_NAME_delete_entry().
18177 * Enhanced the err-ins.pl script so it makes the error library number
18178 global and can add a library name. This is needed for external ASN1 and
18179 other error libraries.
18183 * Fixed sk_insert which never worked properly.
18187 * Fix ASN1 macros so they can handle indefinite length constructed
18188 EXPLICIT tags. Some non standard certificates use these: they can now
18193 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18194 into a single doc/ssleay.txt bundle. This way the information is still
18195 preserved but no longer messes up this directory. Now it's new room for
18196 the new set of documentation files.
18198 *Ralf S. Engelschall*
18200 * SETs were incorrectly DER encoded. This was a major pain, because they
18201 shared code with SEQUENCEs, which aren't coded the same. This means that
18202 almost everything to do with SETs or SEQUENCEs has either changed name or
18203 number of arguments.
18205 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18207 * Fix test data to work with the above.
18211 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18212 was already fixed by Eric for 0.9.1 it seems.
18214 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18216 * Autodetect FreeBSD3.
18220 * Fix various bugs in Configure. This affects the following platforms:
18224 unixware-2.0-pentium
18229 * Eliminate generated files from CVS. Reorder tests to regenerate files
18230 before they are needed.
18234 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18238 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18240 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18241 changed SSLeay to OpenSSL in version strings.
18243 *Ralf S. Engelschall*
18245 * Some fixups to the top-level documents.
18249 * Fixed the nasty bug where rsaref.h was not found under compile-time
18250 because the symlink to include/ was missing.
18252 *Ralf S. Engelschall*
18254 * Incorporated the popular no-RSA/DSA-only patches
18255 which allow to compile a RSA-free SSLeay.
18257 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18259 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18260 when "ssleay" is still not found.
18262 *Ralf S. Engelschall*
18264 * Added more platforms to Configure: Cray T3E, HPUX 11,
18266 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18268 * Updated the README file.
18270 *Ralf S. Engelschall*
18272 * Added various .cvsignore files in the CVS repository subdirs
18273 to make a "cvs update" really silent.
18275 *Ralf S. Engelschall*
18277 * Recompiled the error-definition header files and added
18278 missing symbols to the Win32 linker tables.
18280 *Ralf S. Engelschall*
18282 * Cleaned up the top-level documents;
18283 o new files: CHANGES and LICENSE
18284 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18285 o merged COPYRIGHT into LICENSE
18286 o removed obsolete TODO file
18287 o renamed MICROSOFT to INSTALL.W32
18289 *Ralf S. Engelschall*
18291 * Removed dummy files from the 0.9.1b source tree:
18292 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18293 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18294 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18295 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18296 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18298 *Ralf S. Engelschall*
18300 * Added various platform portability fixes.
18304 * The Genesis of the OpenSSL rpject:
18305 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18306 Young and Tim J. Hudson created while they were working for C2Net until
18309 *The OpenSSL Project*
18311 ### Changes between 0.9.0b and 0.9.1b [not released]
18313 * Updated a few CA certificates under certs/
18317 * Changed some BIGNUM api stuff.
18321 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18322 DGUX x86, Linux Alpha, etc.
18326 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18327 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18332 * Add -strparse option to asn1pars program which parses nested
18335 *Dr Stephen Henson <shenson@bigfoot.com>*
18337 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18341 * DSA fix for "ca" program.
18345 * Added "-genkey" option to "dsaparam" program.
18349 * Added RIPE MD160 (rmd160) message digest.
18353 * Added -a (all) option to "ssleay version" command.
18357 * Added PLATFORM define which is the id given to Configure.
18361 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18365 * Extended the ASN.1 parser routines.
18369 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18373 * Added a BN_CTX to the BN library.
18377 * Fixed the weak key values in DES library
18381 * Changed API in EVP library for cipher aliases.
18385 * Added support for RC2/64bit cipher.
18389 * Converted the lhash library to the crypto/mem.c functions.
18393 * Added more recognized ASN.1 object ids.
18397 * Added more RSA padding checks for SSL/TLS.
18401 * Added BIO proxy/filter functionality.
18405 * Added extra_certs to SSL_CTX which can be used
18406 send extra CA certificates to the client in the CA cert chain sending
18407 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18411 * Now Fortezza is denied in the authentication phase because
18412 this is key exchange mechanism is not supported by SSLeay at all.
18416 * Additional PKCS1 checks.
18420 * Support the string "TLSv1" for all TLS v1 ciphers.
18424 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18425 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18429 * Fixed a few memory leaks.
18433 * Fixed various code and comment typos.
18437 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18438 bytes sent in the client random.
18440 *Edward Bishop <ebishop@spyglass.com>*
18444 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18445 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18446 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18447 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18448 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18449 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18450 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18451 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18452 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18453 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18454 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18455 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18456 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18457 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18458 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18459 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18460 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18461 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18462 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18463 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18464 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18465 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18466 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18467 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18468 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18469 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18470 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18471 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18472 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18473 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18474 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18475 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18476 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18477 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18478 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18479 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18480 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18481 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18482 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18483 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18484 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18485 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18486 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18487 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18488 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18489 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18490 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18491 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18492 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18493 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18494 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18495 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18496 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18497 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18498 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18499 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18500 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18501 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18502 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18503 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18504 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18505 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18506 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18507 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18508 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18509 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18510 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18511 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18512 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18513 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18514 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18515 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18516 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18517 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18518 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18519 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18520 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18521 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18522 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18523 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18524 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18525 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18526 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18527 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18528 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18529 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18530 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18531 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18532 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18533 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18534 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18535 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18536 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18537 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18538 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18539 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18540 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18541 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18542 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18543 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18544 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18545 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18546 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18547 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18548 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18549 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18550 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18551 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18552 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18553 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18554 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18555 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18556 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18557 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18558 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18559 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18560 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18561 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18562 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18563 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18564 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18565 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18566 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18567 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18568 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18569 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18570 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18571 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18572 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18573 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18574 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18575 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18576 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18577 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18578 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18579 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18580 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18581 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18582 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18583 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18584 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18585 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18586 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18587 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18588 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18589 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18590 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18591 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18592 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18593 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18594 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18595 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18596 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18597 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18598 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18599 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18600 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18601 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18602 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18603 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655