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 * Remove the RAND_DRBG API
28 The RAND_DRBG API did not fit well into the new provider concept as
29 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
30 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
31 and some of its API calls are rather low-level. This holds in particular
32 for the callback mechanism (RAND_DRBG_set_callbacks()).
34 Adding a compatibility layer to continue supporting the RAND_DRBG API as
35 a legacy API for a regular deprecation period turned out to come at the
36 price of complicating the new provider API unnecessarily. Since the
37 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
40 *Paul Dale and Matthias St. Pierre*
42 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
43 as well as actual hostnames.
47 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
48 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
49 conversely, silently ignore DTLS protocol version bounds when configuring
50 TLS-based contexts. The commands can be repeated to set bounds of both
51 types. The same applies with the corresponding "min_protocol" and
52 "max_protocol" command-line switches, in case some application uses both TLS
55 SSL_CTX instances that are created for a fixed protocol version (e.g.
56 TLSv1_server_method()) also silently ignore version bounds. Previously
57 attempts to apply bounds to these protocol versions would result in an
58 error. Now only the "version-flexible" SSL_CTX instances are subject to
59 limits in configuration files in command-line options.
63 * Deprecated the `ENGINE` API. Engines should be replaced with providers
68 * Reworked the recorded ERR codes to make better space for system errors.
69 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
70 given code is a system error (true) or an OpenSSL error (false).
74 * Reworked the test perl framework to better allow parallel testing.
76 *Nicola Tuveri and David von Oheimb*
78 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
79 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
83 * 'Configure' has been changed to figure out the configuration target if
84 none is given on the command line. Consequently, the 'config' script is
85 now only a mere wrapper. All documentation is changed to only mention
88 *Rich Salz and Richard Levitte*
90 * Added a library context that applications as well as other
91 libraries can use to form a separate context within which libcrypto
92 operations are performed.
94 There are two ways this can be used:
96 - Directly, by passing a library context to functions that take
97 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
99 - Indirectly, by creating a new library context and then assigning
100 it as the new default, with `OPENSSL_CTX_set0_default`.
102 All public OpenSSL functions that take an `OPENSSL_CTX` pointer,
103 apart from the functions directly related to `OPENSSL_CTX`, accept
104 NULL to indicate that the default library context should be used.
106 Library code that changes the default library context using
107 `OPENSSL_CTX_set0_default` should take care to restore it with a
108 second call before returning to the caller.
112 * Handshake now fails if Extended Master Secret extension is dropped
117 * Dropped interactive mode from the 'openssl' program. From now on,
118 the `openssl` command without arguments is equivalent to `openssl
123 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
124 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
125 While the old function names have been retained for backward compatibility
126 they should not be used in new developments
127 because their return values are confusing: Unlike other `_cmp()` functions
128 they do not return 0 in case their arguments are equal.
132 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
133 EC_GROUP_get_field_type().
137 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
138 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
139 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
140 Applications should rely on the library automatically assigning a suitable
141 EC_METHOD internally upon EC_GROUP construction.
145 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
146 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
147 assigned internally without application intervention.
148 Users of EC_GROUP_new() should switch to a different suitable constructor.
152 * Add CAdES-BES signature verification support, mostly derived
153 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
155 *Filipe Raimundo da Silva*
157 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
161 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
162 functions are not widely used and now OpenSSL automatically perform this
163 conversion when needed.
167 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
168 EC_KEY_precompute_mult(). These functions are not widely used and
169 applications should instead switch to named curves which OpenSSL has
170 hardcoded lookup tables for.
174 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
175 should instead use the L<EC_POINT_mul(3)> function.
179 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
180 that are not applicable to the new provider model. Applications should
181 instead use EVP_default_properties_is_fips_enabled() and
182 EVP_default_properties_enable_fips().
186 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
187 is set, an unexpected EOF is ignored, it pretends a close notify was received
188 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
192 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
193 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
194 used and applications should instead use the
195 L<EC_POINT_set_affine_coordinates(3)> and
196 L<EC_POINT_get_affine_coordinates(3)> functions.
200 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
201 arrays to be more easily constructed via a series of utility functions.
202 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
203 the various push functions and finally convert to a passable OSSL_PARAM
204 array using OSSL_PARAM_BLD_to_param().
208 * The security strength of SHA1 and MD5 based signatures in TLS has been
209 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
210 working at the default security level of 1 and instead requires security
211 level 0. The security level can be changed either using the cipher string
212 with @SECLEVEL, or calling SSL_CTX_set_security_level().
216 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
217 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
218 internal keys, if they correspond to one of those built in types.
222 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
223 contain a provider side internal key.
227 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
228 They are old functions that we don't use, and that you could disable with
229 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
233 * Project text documents not yet having a proper file name extension
234 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
235 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
239 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
240 have been converted to Markdown with the goal to produce documents
241 which not only look pretty when viewed online in the browser, but
242 remain well readable inside a plain text editor.
244 To achieve this goal, a 'minimalistic' Markdown style has been applied
245 which avoids formatting elements that interfere too much with the
246 reading flow in the text file. For example, it
248 * avoids [ATX headings][] and uses [setext headings][] instead
249 (which works for `<h1>` and `<h2>` headings only).
250 * avoids [inline links][] and uses [reference links][] instead.
251 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
253 [ATX headings]: https://github.github.com/gfm/#atx-headings
254 [setext headings]: https://github.github.com/gfm/#setext-headings
255 [inline links]: https://github.github.com/gfm/#inline-link
256 [reference links]: https://github.github.com/gfm/#reference-link
257 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
258 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
260 *Matthias St. Pierre*
262 * The test suite is changed to preserve results of each test recipe.
263 A new directory test-runs/ with subdirectories named like the
264 test recipes are created in the build tree for this purpose.
268 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
269 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
270 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
272 *David von Oheimb, Martin Peylo*
274 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
275 The legacy OCSP-focused and only partly documented API is retained.
276 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
280 * Added `util/check-format.pl`, a tool for checking adherence to the
281 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
282 The checks performed are incomplete and yield some false positives.
283 Still the tool should be useful for detecting most typical glitches.
287 * BIO_do_connect and BIO_do_handshake have been extended:
288 If domain name resolution yields multiple IP addresses all of them are tried
289 after connect() failures.
293 * All of the low level RSA functions have been deprecated including:
295 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
296 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
297 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
298 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
299 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
300 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
301 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
302 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
303 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
304 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
305 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
306 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
307 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
308 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
309 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
310 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
311 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
312 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
313 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
314 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
315 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
316 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
317 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
318 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
319 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
320 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
321 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
322 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
324 Use of these low level functions has been informally discouraged for a long
325 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
326 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
327 L<EVP_PKEY_decrypt(3)>.
331 * X509 certificates signed using SHA1 are no longer allowed at security
333 In TLS/SSL the default security level is 1. It can be set either
334 using the cipher string with @SECLEVEL, or calling
335 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
336 a call to SSL_CTX_use_certificate() will fail if the security level is not
338 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
339 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
344 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
345 modified to use PKEY APIs. These commands are now in maintenance mode
346 and no new features will be added to them.
350 * The command line utility rsautl has been deprecated.
351 Instead use the pkeyutl program.
355 * The command line utilities genrsa and rsa have been modified to use PKEY
356 APIs These commands are now in maintenance mode and no new features will
361 * All of the low level DH functions have been deprecated including:
363 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
364 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
365 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
366 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
367 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
368 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
369 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
370 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
371 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
372 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
373 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
374 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
375 DH_meth_get_generate_params and DH_meth_set_generate_params.
377 Use of these low level functions has been informally discouraged for a long
378 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
379 and L<EVP_PKEY_derive(3)>.
383 * All of the low level DSA functions have been deprecated including:
385 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
386 DSA_get_default_method, DSA_set_method, DSA_get_method,
387 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
388 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
389 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
390 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
391 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
392 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
393 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
394 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
395 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
396 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
397 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
400 Use of these low level functions has been informally discouraged for a long
401 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
402 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
406 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
407 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
408 This means that applications don't have to look at the curve NID and
409 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
410 However, they still can, that EVP_PKEY_set_alias_type() call acts as
411 a no-op when the EVP_PKEY is already of the given type.
413 Parameter and key generation is also reworked to make it possible
414 to generate EVP_PKEY_SM2 parameters and keys without having to go
415 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
416 However, code that does the latter will still work as before.
420 * Deprecated low level ECDH and ECDSA functions. These include:
422 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
423 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
426 Use of these low level functions has been informally discouraged for a long
427 time. Instead applications should use the EVP_PKEY_derive(3),
428 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
432 * Deprecated the EC_KEY_METHOD functions. These include:
434 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
435 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
436 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
437 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
438 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
439 EC_KEY_METHOD_get_verify.
441 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
445 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
446 and EVP_PKEY_decrypt() instead.
447 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
448 and EVP_PKEY_encrypt() instead.
452 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
453 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
454 a new formulation to include all the things it can be used for,
455 as well as words of caution.
459 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
460 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
464 * All of the low level HMAC functions have been deprecated including:
466 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
467 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
470 Use of these low level functions has been informally discouraged for a long
471 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
472 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
473 and L<EVP_MAC_final(3)>.
477 * Over two thousand fixes were made to the documentation, including:
478 - Common options (such as -rand/-writerand, TLS version control, etc)
479 were refactored and point to newly-enhanced descriptions in openssl.pod.
480 - Added style conformance for all options (with help from Richard Levitte),
481 documented all reported missing options, added a CI build to check
482 that all options are documented and that no unimplemented options
484 - Documented some internals, such as all use of environment variables.
485 - Addressed all internal broken L<> references.
489 * All of the low level CMAC functions have been deprecated including:
491 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
492 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
494 Use of these low level functions has been informally discouraged for a long
495 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
496 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
497 and L<EVP_MAC_final(3)>.
501 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
502 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
505 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
506 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
507 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
508 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
509 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
510 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
511 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
512 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
513 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
514 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
516 Use of these low level functions has been informally discouraged
517 for a long time. Applications should use the EVP_DigestInit_ex(3),
518 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
522 * Corrected the documentation of the return values from the `EVP_DigestSign*`
523 set of functions. The documentation mentioned negative values for some
524 errors, but this was never the case, so the mention of negative values
527 Code that followed the documentation and thereby check with something
528 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
532 * All of the low level cipher functions have been deprecated including:
534 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
535 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
536 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
537 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
538 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
539 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
540 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
541 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
542 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
543 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
544 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
545 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
546 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
547 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
548 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
549 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
550 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
551 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
552 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
553 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
554 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
555 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
556 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
557 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
558 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
559 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
560 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
561 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
562 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
564 Use of these low level functions has been informally discouraged for
565 a long time. Applications should use the high level EVP APIs, e.g.
566 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
567 equivalently named decrypt functions instead.
569 *Matt Caswell and Paul Dale*
571 * Removed include/openssl/opensslconf.h.in and replaced it with
572 include/openssl/configuration.h.in, which differs in not including
573 <openssl/macros.h>. A short header include/openssl/opensslconf.h
574 was added to include both.
576 This allows internal hacks where one might need to modify the set
577 of configured macros, for example this if deprecated symbols are
578 still supposed to be available internally:
580 #include <openssl/configuration.h>
582 #undef OPENSSL_NO_DEPRECATED
583 #define OPENSSL_SUPPRESS_DEPRECATED
585 #include <openssl/macros.h>
587 This should not be used by applications that use the exported
588 symbols, as that will lead to linking errors.
592 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
593 used in exponentiation with 512-bit moduli. No EC algorithms are
594 affected. Analysis suggests that attacks against 2-prime RSA1024,
595 3-prime RSA1536, and DSA1024 as a result of this defect would be very
596 difficult to perform and are not believed likely. Attacks against DH512
597 are considered just feasible. However, for an attack the target would
598 have to re-use the DH512 private key, which is not recommended anyway.
599 Also applications directly using the low level API BN_mod_exp may be
600 affected if they use BN_FLG_CONSTTIME.
605 * Most memory-debug features have been deprecated, and the functionality
606 replaced with no-ops.
610 * Added documentation for the STACK API. OpenSSL only defines the STACK
611 functions where they are used.
615 * Introduced a new method type and API, OSSL_SERIALIZER, to
616 represent generic serializers. An implementation is expected to
617 be able to serialize an object associated with a given name (such
618 as an algorithm name for an asymmetric key) into forms given by
619 implementation properties.
621 Serializers are primarily used from inside libcrypto, through
622 calls to functions like EVP_PKEY_print_private(),
623 PEM_write_bio_PrivateKey() and similar.
625 Serializers are specified in such a way that they can be made to
626 directly handle the provider side portion of an object, if this
627 provider side part comes from the same provider as the serializer
628 itself, but can also be made to handle objects in parametrized
629 form (as an OSSL_PARAM array of data). This allows a provider to
630 offer generic serializers as a service for any other provider.
634 * Added a .pragma directive to the syntax of configuration files, to
635 allow varying behavior in a supported and predictable manner.
636 Currently added pragma:
640 This allows dollar signs to be a keyword character unless it's
641 followed by a opening brace or parenthesis. This is useful for
642 platforms where dollar signs are commonly used in names, such as
643 volume names and system directory names on VMS.
647 * Added functionality to create an EVP_PKEY from user data. This
648 is effectively the same as creating a RSA, DH or DSA object and
649 then assigning them to an EVP_PKEY, but directly using algorithm
650 agnostic EVP functions. A benefit is that this should be future
651 proof for public key algorithms to come.
655 * Change the interpretation of the '--api' configuration option to
656 mean that this is a desired API compatibility level with no
657 further meaning. The previous interpretation, that this would
658 also mean to remove all deprecated symbols up to and including
659 the given version, no requires that 'no-deprecated' is also used
660 in the configuration.
662 When building applications, the desired API compatibility level
663 can be set with the OPENSSL_API_COMPAT macro like before. For
664 API compatibility version below 3.0, the old style numerical
665 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
666 For version 3.0 and on, the value is expected to be the decimal
667 value calculated from the major and minor version like this:
669 MAJOR * 10000 + MINOR * 100
673 -DOPENSSL_API_COMPAT=30000 For 3.0
674 -DOPENSSL_API_COMPAT=30200 For 3.2
676 To hide declarations that are deprecated up to and including the
677 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
678 given when building the application as well.
682 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
683 access to certificate and CRL stores via URIs and OSSL_STORE
686 This adds the following functions:
688 - X509_LOOKUP_store()
689 - X509_STORE_load_file()
690 - X509_STORE_load_path()
691 - X509_STORE_load_store()
692 - SSL_add_store_cert_subjects_to_stack()
693 - SSL_CTX_set_default_verify_store()
694 - SSL_CTX_load_verify_file()
695 - SSL_CTX_load_verify_dir()
696 - SSL_CTX_load_verify_store()
700 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
701 The presence of this system service is determined at run-time.
705 * Added functionality to create an EVP_PKEY context based on data
706 for methods from providers. This takes an algorithm name and a
707 property query string and simply stores them, with the intent
708 that any operation that uses this context will use those strings
709 to fetch the needed methods implicitly, thereby making the port
710 of application written for pre-3.0 OpenSSL easier.
714 * The undocumented function NCONF_WIN32() has been deprecated; for
715 conversion details see the HISTORY section of doc/man5/config.pod
719 * Introduced the new functions EVP_DigestSignInit_ex() and
720 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
721 EVP_DigestVerifyUpdate() have been converted to functions. See the man
722 pages for further details.
726 * Over two thousand fixes were made to the documentation, including:
727 adding missing command flags, better style conformance, documentation
730 *Rich Salz, Richard Levitte*
732 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
733 X25519, X448, Ed25519 and Ed448.
737 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
742 * Deprecated the public definition of ERR_STATE as well as the function
743 ERR_get_state(). This is done in preparation of making ERR_STATE an
748 * Added ERR functionality to give callers access to the stored function
749 names that have replaced the older function code based functions.
751 New functions are ERR_get_error_func(), ERR_peek_error_func(),
752 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
753 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
754 and ERR_peek_last_error_all().
756 These functions have become deprecated: ERR_get_error_line_data(),
757 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
758 ERR_func_error_string().
762 * Extended testing to be verbose for failing tests only. The make variables
763 VERBOSE_FAILURE or VF can be used to enable this:
765 $ make VF=1 test # Unix
766 $ mms /macro=(VF=1) test ! OpenVMS
767 $ nmake VF=1 test # Windows
771 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
772 used even when parsing explicit parameters, when loading a serialized key
773 or calling `EC_GROUP_new_from_ecpkparameters()`/
774 `EC_GROUP_new_from_ecparameters()`.
775 This prevents bypass of security hardening and performance gains,
776 especially for curves with specialized EC_METHODs.
777 By default, if a key encoded with explicit parameters is loaded and later
778 serialized, the output is still encoded with explicit parameters, even if
779 internally a "named" EC_GROUP is used for computation.
783 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
784 this change, EC_GROUP_set_generator would accept order and/or cofactor as
785 NULL. After this change, only the cofactor parameter can be NULL. It also
786 does some minimal sanity checks on the passed order.
791 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
792 An attack is simple, if the first CMS_recipientInfo is valid but the
793 second CMS_recipientInfo is chosen ciphertext. If the second
794 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
795 encryption key will be replaced by garbage, and the message cannot be
796 decoded, but if the RSA decryption fails, the correct encryption key is
797 used and the recipient will not notice the attack.
798 As a work around for this potential attack the length of the decrypted
799 key must be equal to the cipher default key length, in case the
800 certifiate is not given and all recipientInfo are tried out.
801 The old behaviour can be re-enabled in the CMS code by setting the
802 CMS_DEBUG_DECRYPT flag.
806 * Early start up entropy quality from the DEVRANDOM seed source has been
807 improved for older Linux systems. The RAND subsystem will wait for
808 /dev/random to be producing output before seeding from /dev/urandom.
809 The seeded state is stored for future library initialisations using
810 a system global shared memory segment. The shared memory identifier
811 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
812 the desired value. The default identifier is 114.
816 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
817 when primes for RSA keys are computed.
818 Since we previously always generated primes == 2 (mod 3) for RSA keys,
819 the 2-prime and 3-prime RSA modules were easy to distinguish, since
820 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
821 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
822 This avoids possible fingerprinting of newly generated RSA modules.
826 * Correct the extended master secret constant on EBCDIC systems. Without this
827 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
828 negotiate EMS will fail. Unfortunately this also means that TLS connections
829 between EBCDIC systems with this fix, and EBCDIC systems without this
830 fix will fail if they negotiate EMS.
834 * Changed the library initialisation so that the config file is now loaded
835 by default. This was already the case for libssl. It now occurs for both
836 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
837 OPENSSL_init_crypto() to suppress automatic loading of a config file.
841 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
842 where the former acts as a replacement for ERR_put_error(), and the
843 latter replaces the combination ERR_put_error()+ERR_add_error_data().
844 ERR_raise_data() adds more flexibility by taking a format string and
845 an arbitrary number of arguments following it, to be processed with
850 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
851 to check if a named provider is loaded and available. When called, it
852 will also activate all fallback providers if such are still present.
856 * Enforce a minimum DH modulus size of 512 bits.
860 * Changed DH parameters to generate the order q subgroup instead of 2q.
861 Previously generated DH parameters are still accepted by DH_check
862 but DH_generate_key works around that by clearing bit 0 of the
863 private key for those. This avoids leaking bit 0 of the private key.
867 * Significantly reduce secure memory usage by the randomness pools.
871 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
876 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
877 algorithms. An implementation of a key exchange algorithm can be obtained
878 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
879 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
880 the older EVP_PKEY_derive_init() function. See the man pages for the new
881 functions for further details.
885 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
889 * Removed the function names from error messages and deprecated the
892 * Removed NextStep support and the macro OPENSSL_UNISTD
896 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
897 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
898 Also removed "export var as function" capability; we do not export
899 variables, only functions.
903 * RC5_32_set_key has been changed to return an int type, with 0 indicating
904 an error and 1 indicating success. In previous versions of OpenSSL this
905 was a void type. If a key was set longer than the maximum possible this
910 * Support SM2 signing and verification schemes with X509 certificate.
914 * Use SHA256 as the default digest for TS query in the ts app.
918 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
919 This checks that the salt length is at least 128 bits, the derived key
920 length is at least 112 bits, and that the iteration count is at least 1000.
921 For backwards compatibility these checks are disabled by default in the
922 default provider, but are enabled by default in the fips provider.
923 To enable or disable these checks use the control
924 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
928 * Default cipher lists/suites are now available via a function, the
929 #defines are deprecated.
933 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
934 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
935 for Windows Store apps easier. Also, the "no-uplink" option has been added.
939 * Join the directories crypto/x509 and crypto/x509v3
943 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
944 This changes the size when using the genpkey app when no size is given. It
945 fixes an omission in earlier changes that changed all RSA, DSA and DH
946 generation apps to use 2048 bits by default.
950 * Added command 'openssl kdf' that uses the EVP_KDF API.
954 * Added command 'openssl mac' that uses the EVP_MAC API.
958 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
959 as default directories. Also added the command 'openssl info'
960 for scripting purposes.
964 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
965 deprecated. These undocumented functions were never integrated into the EVP
966 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
967 Bi-directional IGE mode. These modes were never formally standardised and
968 usage of these functions is believed to be very small. In particular
969 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
970 is ever used. The security implications are believed to be minimal, but
971 this issue was never fixed for backwards compatibility reasons. New code
972 should not use these modes.
976 * Add prediction resistance to the DRBG reseeding process.
980 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
981 mandated by IEEE Std 1619-2018.
985 * Added newline escaping functionality to a filename when using openssl dgst.
986 This output format is to replicate the output format found in the `*sum`
987 checksum programs. This aims to preserve backward compatibility.
989 *Matt Eaton, Richard Levitte, and Paul Dale*
991 * Removed the heartbeat message in DTLS feature, as it has very
992 little usage and doesn't seem to fulfill a valuable purpose.
993 The configuration option is now deprecated.
997 * Changed the output of 'openssl {digestname} < file' to display the
998 digest name in its output.
1002 * Added a new generic trace API which provides support for enabling
1003 instrumentation through trace output. This feature is mainly intended
1004 as an aid for developers and is disabled by default. To utilize it,
1005 OpenSSL needs to be configured with the `enable-trace` option.
1007 If the tracing API is enabled, the application can activate trace output
1008 by registering BIOs as trace channels for a number of tracing and debugging
1011 The 'openssl' application has been expanded to enable any of the types
1012 available via environment variables defined by the user, and serves as
1013 one possible example on how to use this functionality.
1015 *Richard Levitte & Matthias St. Pierre*
1017 * Added build tests for C++. These are generated files that only do one
1018 thing, to include one public OpenSSL head file each. This tests that
1019 the public header files can be usefully included in a C++ application.
1021 This test isn't enabled by default. It can be enabled with the option
1022 'enable-buildtest-c++'.
1026 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1030 * Add KMAC to EVP_MAC.
1034 * Added property based algorithm implementation selection framework to
1039 * Added SCA hardening for modular field inversion in EC_GROUP through
1040 a new dedicated field_inv() pointer in EC_METHOD.
1041 This also addresses a leakage affecting conversions from projective
1042 to affine coordinates.
1044 *Billy Bob Brumley, Nicola Tuveri*
1046 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1047 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1048 those algorithms that were already supported through the EVP_PKEY API
1049 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1050 and scrypt are now wrappers that call EVP_KDF.
1054 * Build devcrypto engine as a dynamic engine.
1056 *Eneas U de Queiroz*
1058 * Add keyed BLAKE2 to EVP_MAC.
1062 * Fix a bug in the computation of the endpoint-pair shared secret used
1063 by DTLS over SCTP. This breaks interoperability with older versions
1064 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1065 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1066 interoperability with such broken implementations. However, enabling
1067 this switch breaks interoperability with correct implementations.
1069 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1070 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1074 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1078 * Change the license to the Apache License v2.0.
1082 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1084 - Major releases (indicated by incrementing the MAJOR release number)
1085 may introduce incompatible API/ABI changes.
1086 - Minor releases (indicated by incrementing the MINOR release number)
1087 may introduce new features but retain API/ABI compatibility.
1088 - Patch releases (indicated by incrementing the PATCH number)
1089 are intended for bug fixes and other improvements of existing
1090 features only (like improving performance or adding documentation)
1091 and retain API/ABI compatibility.
1095 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1099 * Remove the 'dist' target and add a tarball building script. The
1100 'dist' target has fallen out of use, and it shouldn't be
1101 necessary to configure just to create a source distribution.
1105 * Recreate the OS390-Unix config target. It no longer relies on a
1106 special script like it did for OpenSSL pre-1.1.0.
1110 * Instead of having the source directories listed in Configure, add
1111 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1116 * Add GMAC to EVP_MAC.
1120 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1124 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1125 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1126 to facilitate the continued use of MACs through raw private keys in
1127 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1131 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1132 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1136 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1137 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1138 are retained for backwards compatibility.
1142 * AES-XTS mode now enforces that its two keys are different to mitigate
1143 the attacked described in "Efficient Instantiations of Tweakable
1144 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1145 Details of this attack can be obtained from:
1146 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1150 * Rename the object files, i.e. give them other names than in previous
1151 versions. Their names now include the name of the final product, as
1152 well as its type mnemonic (bin, lib, shlib).
1156 * Added new option for 'openssl list', '-objects', which will display the
1157 list of built in objects, i.e. OIDs with names.
1161 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1162 improves application performance by removing data copies and providing
1163 applications with zero-copy system calls such as sendfile and splice.
1167 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1168 option is set, openssl cleanses (zeroize) plaintext bytes from
1169 internal buffers after delivering them to the application. Note,
1170 the application is still responsible for cleansing other copies
1171 (e.g.: data received by SSL_read(3)).
1178 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1180 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1182 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1183 while reading in libssl then we would report an error back to the
1184 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1185 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1186 therefore give a hint as to what went wrong.
1190 * Check that ed25519 and ed448 are allowed by the security level. Previously
1191 signature algorithms not using an MD were not being checked that they were
1192 allowed by the security level.
1196 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1197 was not quite right. The behaviour was not consistent between resumption
1198 and normal handshakes, and also not quite consistent with historical
1199 behaviour. The behaviour in various scenarios has been clarified and
1200 it has been updated to make it match historical behaviour as closely as
1205 * *[VMS only]* The header files that the VMS compilers include automatically,
1206 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1207 that the C++ compiler doesn't understand. This is a shortcoming in the
1208 compiler, but can be worked around with `__cplusplus` guards.
1210 C++ applications that use OpenSSL libraries must be compiled using the
1211 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1212 functions. Otherwise, only functions with symbols of less than 31
1213 characters can be used, as the linker will not be able to successfully
1214 resolve symbols with longer names.
1218 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1219 The presence of this system service is determined at run-time.
1223 * Added newline escaping functionality to a filename when using openssl dgst.
1224 This output format is to replicate the output format found in the `*sum`
1225 checksum programs. This aims to preserve backward compatibility.
1227 *Matt Eaton, Richard Levitte, and Paul Dale*
1229 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1234 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1236 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1237 number generator (RNG). This was intended to include protection in the
1238 event of a fork() system call in order to ensure that the parent and child
1239 processes did not share the same RNG state. However this protection was not
1240 being used in the default case.
1242 A partial mitigation for this issue is that the output from a high
1243 precision timer is mixed into the RNG state so the likelihood of a parent
1244 and child process sharing state is significantly reduced.
1246 If an application already calls OPENSSL_init_crypto() explicitly using
1247 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1250 *Matthias St. Pierre*
1252 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1253 used even when parsing explicit parameters, when loading a serialized key
1254 or calling `EC_GROUP_new_from_ecpkparameters()`/
1255 `EC_GROUP_new_from_ecparameters()`.
1256 This prevents bypass of security hardening and performance gains,
1257 especially for curves with specialized EC_METHODs.
1258 By default, if a key encoded with explicit parameters is loaded and later
1259 serialized, the output is still encoded with explicit parameters, even if
1260 internally a "named" EC_GROUP is used for computation.
1264 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1265 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1266 NULL. After this change, only the cofactor parameter can be NULL. It also
1267 does some minimal sanity checks on the passed order.
1272 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1273 An attack is simple, if the first CMS_recipientInfo is valid but the
1274 second CMS_recipientInfo is chosen ciphertext. If the second
1275 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1276 encryption key will be replaced by garbage, and the message cannot be
1277 decoded, but if the RSA decryption fails, the correct encryption key is
1278 used and the recipient will not notice the attack.
1279 As a work around for this potential attack the length of the decrypted
1280 key must be equal to the cipher default key length, in case the
1281 certifiate is not given and all recipientInfo are tried out.
1282 The old behaviour can be re-enabled in the CMS code by setting the
1283 CMS_DEBUG_DECRYPT flag.
1288 * Early start up entropy quality from the DEVRANDOM seed source has been
1289 improved for older Linux systems. The RAND subsystem will wait for
1290 /dev/random to be producing output before seeding from /dev/urandom.
1291 The seeded state is stored for future library initialisations using
1292 a system global shared memory segment. The shared memory identifier
1293 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1294 the desired value. The default identifier is 114.
1298 * Correct the extended master secret constant on EBCDIC systems. Without this
1299 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1300 negotiate EMS will fail. Unfortunately this also means that TLS connections
1301 between EBCDIC systems with this fix, and EBCDIC systems without this
1302 fix will fail if they negotiate EMS.
1306 * Use Windows installation paths in the mingw builds
1308 Mingw isn't a POSIX environment per se, which means that Windows
1309 paths should be used for installation.
1314 * Changed DH_check to accept parameters with order q and 2q subgroups.
1315 With order 2q subgroups the bit 0 of the private key is not secret
1316 but DH_generate_key works around that by clearing bit 0 of the
1317 private key for those. This avoids leaking bit 0 of the private key.
1321 * Significantly reduce secure memory usage by the randomness pools.
1325 * Revert the DEVRANDOM_WAIT feature for Linux systems
1327 The DEVRANDOM_WAIT feature added a select() call to wait for the
1328 /dev/random device to become readable before reading from the
1329 /dev/urandom device.
1331 It turned out that this change had negative side effects on
1332 performance which were not acceptable. After some discussion it
1333 was decided to revert this feature and leave it up to the OS
1334 resp. the platform maintainer to ensure a proper initialization
1335 during early boot time.
1337 *Matthias St. Pierre*
1339 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1341 * Add build tests for C++. These are generated files that only do one
1342 thing, to include one public OpenSSL head file each. This tests that
1343 the public header files can be usefully included in a C++ application.
1345 This test isn't enabled by default. It can be enabled with the option
1346 'enable-buildtest-c++'.
1350 * Enable SHA3 pre-hashing for ECDSA and DSA.
1354 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1355 This changes the size when using the genpkey app when no size is given. It
1356 fixes an omission in earlier changes that changed all RSA, DSA and DH
1357 generation apps to use 2048 bits by default.
1361 * Reorganize the manual pages to consistently have RETURN VALUES,
1362 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1363 util/fix-doc-nits accordingly.
1365 *Paul Yang, Joshua Lock*
1367 * Add the missing accessor EVP_PKEY_get0_engine()
1371 * Have apps like 's_client' and 's_server' output the signature scheme
1372 along with other cipher suite parameters when debugging.
1376 * Make OPENSSL_config() error agnostic again.
1380 * Do the error handling in RSA decryption constant time.
1384 * Prevent over long nonces in ChaCha20-Poly1305.
1386 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1387 for every encryption operation. RFC 7539 specifies that the nonce value
1388 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1389 and front pads the nonce with 0 bytes if it is less than 12
1390 bytes. However it also incorrectly allows a nonce to be set of up to 16
1391 bytes. In this case only the last 12 bytes are significant and any
1392 additional leading bytes are ignored.
1394 It is a requirement of using this cipher that nonce values are
1395 unique. Messages encrypted using a reused nonce value are susceptible to
1396 serious confidentiality and integrity attacks. If an application changes
1397 the default nonce length to be longer than 12 bytes and then makes a
1398 change to the leading bytes of the nonce expecting the new value to be a
1399 new unique nonce then such an application could inadvertently encrypt
1400 messages with a reused nonce.
1402 Additionally the ignored bytes in a long nonce are not covered by the
1403 integrity guarantee of this cipher. Any application that relies on the
1404 integrity of these ignored leading bytes of a long nonce may be further
1405 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1406 is safe because no such use sets such a long nonce value. However user
1407 applications that use this cipher directly and set a non-default nonce
1408 length to be longer than 12 bytes may be vulnerable.
1410 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1416 * Add DEVRANDOM_WAIT feature for Linux systems
1418 On older Linux systems where the getrandom() system call is not available,
1419 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1420 Contrary to getrandom(), the /dev/urandom device will not block during
1421 early boot when the kernel CSPRNG has not been seeded yet.
1423 To mitigate this known weakness, use select() to wait for /dev/random to
1424 become readable before reading from /dev/urandom.
1426 * Ensure that SM2 only uses SM3 as digest algorithm
1430 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1432 * Change the info callback signals for the start and end of a post-handshake
1433 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1434 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1435 confused by this and assume that a TLSv1.2 renegotiation has started. This
1436 can break KeyUpdate handling. Instead we no longer signal the start and end
1437 of a post handshake message exchange (although the messages themselves are
1438 still signalled). This could break some applications that were expecting
1439 the old signals. However without this KeyUpdate is not usable for many
1444 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1446 * Timing vulnerability in DSA signature generation
1448 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1449 timing side channel attack. An attacker could use variations in the signing
1450 algorithm to recover the private key.
1452 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1457 * Timing vulnerability in ECDSA signature generation
1459 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1460 timing side channel attack. An attacker could use variations in the signing
1461 algorithm to recover the private key.
1463 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1468 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1469 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1470 of two gigabytes and the error handling improved.
1472 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1473 categorized as a normal bug, not a security issue, because the DRBG reseeds
1474 automatically and is fully functional even without additional randomness
1475 provided by the application.
1477 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1479 * Add a new ClientHello callback. Provides a callback interface that gives
1480 the application the ability to adjust the nascent SSL object at the
1481 earliest stage of ClientHello processing, immediately after extensions have
1482 been collected but before they have been processed. In particular, this
1483 callback can adjust the supported TLS versions in response to the contents
1488 * Add SM2 base algorithm support.
1492 * s390x assembly pack: add (improved) hardware-support for the following
1493 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1494 aes-cfb/cfb8, aes-ecb.
1498 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1499 parameter is no longer accepted, as it leads to a corrupt table. NULL
1500 pem_str is reserved for alias entries only.
1504 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1505 step for prime curves. The new implementation is based on formulae from
1506 differential addition-and-doubling in homogeneous projective coordinates
1507 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1508 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1509 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1510 to work in projective coordinates.
1512 *Billy Bob Brumley, Nicola Tuveri*
1514 * Change generating and checking of primes so that the error rate of not
1515 being prime depends on the intended use based on the size of the input.
1516 For larger primes this will result in more rounds of Miller-Rabin.
1517 The maximal error rate for primes with more than 1080 bits is lowered
1520 *Kurt Roeckx, Annie Yousar*
1522 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1526 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1527 moving between systems, and to avoid confusion when a Windows build is
1528 done with mingw vs with MSVC. For POSIX installs, there's still a
1529 symlink or copy named 'tsget' to avoid that confusion as well.
1533 * Revert blinding in ECDSA sign and instead make problematic addition
1534 length-invariant. Switch even to fixed-length Montgomery multiplication.
1538 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1539 step for binary curves. The new implementation is based on formulae from
1540 differential addition-and-doubling in mixed Lopez-Dahab projective
1541 coordinates, modified to independently blind the operands.
1543 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1545 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1546 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1547 EC_METHODs to implement their own specialized "ladder step", to take
1548 advantage of more favorable coordinate systems or more efficient
1549 differential addition-and-doubling algorithms.
1551 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1553 * Modified the random device based seed sources to keep the relevant
1554 file descriptors open rather than reopening them on each access.
1555 This allows such sources to operate in a chroot() jail without
1556 the associated device nodes being available. This behaviour can be
1557 controlled using RAND_keep_random_devices_open().
1561 * Numerous side-channel attack mitigations have been applied. This may have
1562 performance impacts for some algorithms for the benefit of improved
1563 security. Specific changes are noted in this change log by their respective
1568 * AIX shared library support overhaul. Switch to AIX "natural" way of
1569 handling shared libraries, which means collecting shared objects of
1570 different versions and bitnesses in one common archive. This allows to
1571 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1572 doesn't affect the way 3rd party applications are linked, only how
1573 multi-version installation is managed.
1577 * Make ec_group_do_inverse_ord() more robust and available to other
1578 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1579 mitigations are applied to the fallback BN_mod_inverse().
1580 When using this function rather than BN_mod_inverse() directly, new
1581 EC cryptosystem implementations are then safer-by-default.
1585 * Add coordinate blinding for EC_POINT and implement projective
1586 coordinate blinding for generic prime curves as a countermeasure to
1587 chosen point SCA attacks.
1589 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1591 * Add blinding to ECDSA and DSA signatures to protect against side channel
1592 attacks discovered by Keegan Ryan (NCC Group).
1596 * Enforce checking in the pkeyutl command line app to ensure that the input
1597 length does not exceed the maximum supported digest length when performing
1598 a sign, verify or verifyrecover operation.
1602 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1603 I/O in combination with something like select() or poll() will hang. This
1604 can be turned off again using SSL_CTX_clear_mode().
1605 Many applications do not properly handle non-application data records, and
1606 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1607 around the problems in those applications, but can also break some.
1608 It's recommended to read the manpages about SSL_read(), SSL_write(),
1609 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1610 SSL_CTX_set_read_ahead() again.
1614 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1615 now allow empty (zero character) pass phrases.
1619 * Apply blinding to binary field modular inversion and remove patent
1620 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1624 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1625 binary and prime elliptic curves.
1629 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1630 constant time fixed point multiplication.
1634 * Revise elliptic curve scalar multiplication with timing attack
1635 defenses: ec_wNAF_mul redirects to a constant time implementation
1636 when computing fixed point and variable point multiplication (which
1637 in OpenSSL are mostly used with secret scalars in keygen, sign,
1638 ECDH derive operations).
1639 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1642 * Updated CONTRIBUTING
1646 * Updated DRBG / RAND to request nonce and additional low entropy
1647 randomness from the system.
1649 *Matthias St. Pierre*
1651 * Updated 'openssl rehash' to use OpenSSL consistent default.
1655 * Moved the load of the ssl_conf module to libcrypto, which helps
1656 loading engines that libssl uses before libssl is initialised.
1660 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1664 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1666 *Ingo Schwarze, Rich Salz*
1668 * Added output of accepting IP address and port for 'openssl s_server'
1672 * Added a new API for TLSv1.3 ciphersuites:
1673 SSL_CTX_set_ciphersuites()
1674 SSL_set_ciphersuites()
1678 * Memory allocation failures consistently add an error to the error
1683 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1684 in libcrypto when run as setuid/setgid.
1688 * Load any config file by default when libssl is used.
1692 * Added new public header file <openssl/rand_drbg.h> and documentation
1693 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1695 *Matthias St. Pierre*
1697 * QNX support removed (cannot find contributors to get their approval
1698 for the license change).
1702 * TLSv1.3 replay protection for early data has been implemented. See the
1703 SSL_read_early_data() man page for further details.
1707 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1708 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1709 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1710 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1711 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1712 configuration has been separated out. See the ciphers man page or the
1713 SSL_CTX_set_ciphersuites() man page for more information.
1717 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1718 in responder mode now supports the new "-multi" option, which
1719 spawns the specified number of child processes to handle OCSP
1720 requests. The "-timeout" option now also limits the OCSP
1721 responder's patience to wait to receive the full client request
1722 on a newly accepted connection. Child processes are respawned
1723 as needed, and the CA index file is automatically reloaded
1724 when changed. This makes it possible to run the "ocsp" responder
1725 as a long-running service, making the OpenSSL CA somewhat more
1726 feature-complete. In this mode, most diagnostic messages logged
1727 after entering the event loop are logged via syslog(3) rather than
1732 * Added support for X448 and Ed448. Heavily based on original work by
1737 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1738 objects loaded. This adds the functions OSSL_STORE_expect() and
1739 OSSL_STORE_find() as well as needed tools to construct searches and
1740 get the search data out of them.
1744 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1745 version of OpenSSL should review their configuration settings to ensure
1746 that they are still appropriate for TLSv1.3. For further information see:
1747 <https://wiki.openssl.org/index.php/TLS1.3>
1751 * Grand redesign of the OpenSSL random generator
1753 The default RAND method now utilizes an AES-CTR DRBG according to
1754 NIST standard SP 800-90Ar1. The new random generator is essentially
1755 a port of the default random generator from the OpenSSL FIPS 2.0
1756 object module. It is a hybrid deterministic random bit generator
1757 using an AES-CTR bit stream and which seeds and reseeds itself
1758 automatically using trusted system entropy sources.
1760 Some of its new features are:
1761 - Support for multiple DRBG instances with seed chaining.
1762 - The default RAND method makes use of a DRBG.
1763 - There is a public and private DRBG instance.
1764 - The DRBG instances are fork-safe.
1765 - Keep all global DRBG instances on the secure heap if it is enabled.
1766 - The public and private DRBG instance are per thread for lock free
1769 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1771 * Changed Configure so it only says what it does and doesn't dump
1772 so much data. Instead, ./configdata.pm should be used as a script
1773 to display all sorts of configuration data.
1777 * Added processing of "make variables" to Configure.
1781 * Added SHA512/224 and SHA512/256 algorithm support.
1785 * The last traces of Netware support, first removed in 1.1.0, have
1790 * Get rid of Makefile.shared, and in the process, make the processing
1791 of certain files (rc.obj, or the .def/.map/.opt files produced from
1792 the ordinal files) more visible and hopefully easier to trace and
1793 debug (or make silent).
1797 * Make it possible to have environment variable assignments as
1798 arguments to config / Configure.
1802 * Add multi-prime RSA (RFC 8017) support.
1806 * Add SM3 implemented according to GB/T 32905-2016
1807 *Jack Lloyd <jack.lloyd@ribose.com>,*
1808 *Ronald Tse <ronald.tse@ribose.com>,*
1809 *Erick Borsboom <erick.borsboom@ribose.com>*
1811 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1812 as documented in RFC6066.
1813 Based on a patch from Tomasz Moń
1815 *Filipe Raimundo da Silva*
1817 * Add SM4 implemented according to GB/T 32907-2016.
1818 *Jack Lloyd <jack.lloyd@ribose.com>,*
1819 *Ronald Tse <ronald.tse@ribose.com>,*
1820 *Erick Borsboom <erick.borsboom@ribose.com>*
1822 * Reimplement -newreq-nodes and ERR_error_string_n; the
1823 original author does not agree with the license change.
1827 * Add ARIA AEAD TLS support.
1831 * Some macro definitions to support VS6 have been removed. Visual
1832 Studio 6 has not worked since 1.1.0
1836 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1837 without clearing the errors.
1841 * Add "atfork" functions. If building on a system that without
1842 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1843 requirements. The RAND facility now uses/requires this.
1851 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1852 not possible to disable entirely. However, it's still possible to
1853 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1856 To disable, configure with 'no-ui-console'. 'no-ui' is still
1857 possible to use as an alias. Check at compile time with the
1858 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1859 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1863 * Add a STORE module, which implements a uniform and URI based reader of
1864 stores that can contain keys, certificates, CRLs and numerous other
1865 objects. The main API is loosely based on a few stdio functions,
1866 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1867 OSSL_STORE_error and OSSL_STORE_close.
1868 The implementation uses backends called "loaders" to implement arbitrary
1869 URI schemes. There is one built in "loader" for the 'file' scheme.
1873 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1874 then adjusted to work on FreeBSD 8.4 as well.
1875 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1876 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1880 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1881 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1882 error code calls like this:
1884 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1886 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1887 that can be encoded in C. For the foreseeable future, this will only
1890 *Richard Levitte and Tim Hudson*
1892 * Removed BSD cryptodev engine.
1896 * Add a build target 'build_all_generated', to build all generated files
1897 and only that. This can be used to prepare everything that requires
1898 things like perl for a system that lacks perl and then move everything
1899 to that system and do the rest of the build there.
1903 * In the UI interface, make it possible to duplicate the user data. This
1904 can be used by engines that need to retain the data for a longer time
1905 than just the call where this user data is passed.
1909 * Ignore the '-named_curve auto' value for compatibility of applications
1912 *Tomas Mraz <tmraz@fedoraproject.org>*
1914 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1915 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1916 alerts across multiple records (some of which could be empty). In practice
1917 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1918 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1919 support this at all. Supporting it adds significant complexity to the
1920 record layer, and its removal is unlikely to cause interoperability
1925 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1926 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1927 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1932 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1933 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1935 *Richard Levitte, Andy Polyakov*
1937 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1942 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1943 platform rather than 'mingw'.
1947 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1948 success if they are asked to add an object which already exists
1949 in the store. This change cascades to other functions which load
1950 certificates and CRLs.
1954 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1955 facilitate stack unwinding even from assembly subroutines.
1959 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1960 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1964 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1965 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1966 which is the minimum version we support.
1970 * Certificate time validation (X509_cmp_time) enforces stricter
1971 compliance with RFC 5280. Fractional seconds and timezone offsets
1972 are no longer allowed.
1976 * Add support for ARIA
1980 * s_client will now send the Server Name Indication (SNI) extension by
1981 default unless the new "-noservername" option is used. The server name is
1982 based on the host provided to the "-connect" option unless overridden by
1983 using "-servername".
1987 * Add support for SipHash
1991 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1992 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1993 prevent issues where no progress is being made and the peer continually
1994 sends unrecognised record types, using up resources processing them.
1998 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1999 using the algorithm defined in
2000 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2004 * Heartbeat support has been removed; the ABI is changed for now.
2006 *Richard Levitte, Rich Salz*
2008 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2012 * The RSA "null" method, which was partially supported to avoid patent
2013 issues, has been replaced to always returns NULL.
2020 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2022 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2023 used even when parsing explicit parameters, when loading a serialized key
2024 or calling `EC_GROUP_new_from_ecpkparameters()`/
2025 `EC_GROUP_new_from_ecparameters()`.
2026 This prevents bypass of security hardening and performance gains,
2027 especially for curves with specialized EC_METHODs.
2028 By default, if a key encoded with explicit parameters is loaded and later
2029 serialized, the output is still encoded with explicit parameters, even if
2030 internally a "named" EC_GROUP is used for computation.
2034 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2035 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2036 NULL. After this change, only the cofactor parameter can be NULL. It also
2037 does some minimal sanity checks on the passed order.
2042 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2043 An attack is simple, if the first CMS_recipientInfo is valid but the
2044 second CMS_recipientInfo is chosen ciphertext. If the second
2045 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2046 encryption key will be replaced by garbage, and the message cannot be
2047 decoded, but if the RSA decryption fails, the correct encryption key is
2048 used and the recipient will not notice the attack.
2049 As a work around for this potential attack the length of the decrypted
2050 key must be equal to the cipher default key length, in case the
2051 certifiate is not given and all recipientInfo are tried out.
2052 The old behaviour can be re-enabled in the CMS code by setting the
2053 CMS_DEBUG_DECRYPT flag.
2058 * Use Windows installation paths in the mingw builds
2060 Mingw isn't a POSIX environment per se, which means that Windows
2061 paths should be used for installation.
2066 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2068 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2069 This changes the size when using the genpkey app when no size is given. It
2070 fixes an omission in earlier changes that changed all RSA, DSA and DH
2071 generation apps to use 2048 bits by default.
2075 * Prevent over long nonces in ChaCha20-Poly1305.
2077 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2078 for every encryption operation. RFC 7539 specifies that the nonce value
2079 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2080 and front pads the nonce with 0 bytes if it is less than 12
2081 bytes. However it also incorrectly allows a nonce to be set of up to 16
2082 bytes. In this case only the last 12 bytes are significant and any
2083 additional leading bytes are ignored.
2085 It is a requirement of using this cipher that nonce values are
2086 unique. Messages encrypted using a reused nonce value are susceptible to
2087 serious confidentiality and integrity attacks. If an application changes
2088 the default nonce length to be longer than 12 bytes and then makes a
2089 change to the leading bytes of the nonce expecting the new value to be a
2090 new unique nonce then such an application could inadvertently encrypt
2091 messages with a reused nonce.
2093 Additionally the ignored bytes in a long nonce are not covered by the
2094 integrity guarantee of this cipher. Any application that relies on the
2095 integrity of these ignored leading bytes of a long nonce may be further
2096 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2097 is safe because no such use sets such a long nonce value. However user
2098 applications that use this cipher directly and set a non-default nonce
2099 length to be longer than 12 bytes may be vulnerable.
2101 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2107 * Added SCA hardening for modular field inversion in EC_GROUP through
2108 a new dedicated field_inv() pointer in EC_METHOD.
2109 This also addresses a leakage affecting conversions from projective
2110 to affine coordinates.
2112 *Billy Bob Brumley, Nicola Tuveri*
2114 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2115 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2119 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2123 * Remove the 'dist' target and add a tarball building script. The
2124 'dist' target has fallen out of use, and it shouldn't be
2125 necessary to configure just to create a source distribution.
2129 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2131 * Timing vulnerability in DSA signature generation
2133 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2134 timing side channel attack. An attacker could use variations in the signing
2135 algorithm to recover the private key.
2137 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2142 * Timing vulnerability in ECDSA signature generation
2144 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2145 timing side channel attack. An attacker could use variations in the signing
2146 algorithm to recover the private key.
2148 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2153 * Add coordinate blinding for EC_POINT and implement projective
2154 coordinate blinding for generic prime curves as a countermeasure to
2155 chosen point SCA attacks.
2157 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2159 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2161 * Client DoS due to large DH parameter
2163 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2164 malicious server can send a very large prime value to the client. This will
2165 cause the client to spend an unreasonably long period of time generating a
2166 key for this prime resulting in a hang until the client has finished. This
2167 could be exploited in a Denial Of Service attack.
2169 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2174 * Cache timing vulnerability in RSA Key Generation
2176 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2177 a cache timing side channel attack. An attacker with sufficient access to
2178 mount cache timing attacks during the RSA key generation process could
2179 recover the private key.
2181 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2182 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2187 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2188 parameter is no longer accepted, as it leads to a corrupt table. NULL
2189 pem_str is reserved for alias entries only.
2193 * Revert blinding in ECDSA sign and instead make problematic addition
2194 length-invariant. Switch even to fixed-length Montgomery multiplication.
2198 * Change generating and checking of primes so that the error rate of not
2199 being prime depends on the intended use based on the size of the input.
2200 For larger primes this will result in more rounds of Miller-Rabin.
2201 The maximal error rate for primes with more than 1080 bits is lowered
2204 *Kurt Roeckx, Annie Yousar*
2206 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2210 * Add blinding to ECDSA and DSA signatures to protect against side channel
2211 attacks discovered by Keegan Ryan (NCC Group).
2215 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2216 now allow empty (zero character) pass phrases.
2220 * Certificate time validation (X509_cmp_time) enforces stricter
2221 compliance with RFC 5280. Fractional seconds and timezone offsets
2222 are no longer allowed.
2226 * Fixed a text canonicalisation bug in CMS
2228 Where a CMS detached signature is used with text content the text goes
2229 through a canonicalisation process first prior to signing or verifying a
2230 signature. This process strips trailing space at the end of lines, converts
2231 line terminators to CRLF and removes additional trailing line terminators
2232 at the end of a file. A bug in the canonicalisation process meant that
2233 some characters, such as form-feed, were incorrectly treated as whitespace
2234 and removed. This is contrary to the specification (RFC5485). This fix
2235 could mean that detached text data signed with an earlier version of
2236 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2237 signed with a fixed OpenSSL may fail to verify with an earlier version of
2238 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2239 and use the "-binary" flag (for the "cms" command line application) or set
2240 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2244 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2246 * Constructed ASN.1 types with a recursive definition could exceed the stack
2248 Constructed ASN.1 types with a recursive definition (such as can be found
2249 in PKCS7) could eventually exceed the stack given malicious input with
2250 excessive recursion. This could result in a Denial Of Service attack. There
2251 are no such structures used within SSL/TLS that come from untrusted sources
2252 so this is considered safe.
2254 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2260 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2262 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2263 effectively reduced to only comparing the least significant bit of each
2264 byte. This allows an attacker to forge messages that would be considered as
2265 authenticated in an amount of tries lower than that guaranteed by the
2266 security claims of the scheme. The module can only be compiled by the
2267 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2269 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2275 * Add a build target 'build_all_generated', to build all generated files
2276 and only that. This can be used to prepare everything that requires
2277 things like perl for a system that lacks perl and then move everything
2278 to that system and do the rest of the build there.
2282 * Backport SSL_OP_NO_RENGOTIATION
2284 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2285 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2286 changes this is no longer possible in 1.1.0. Therefore the new
2287 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2288 1.1.0 to provide equivalent functionality.
2290 Note that if an application built against 1.1.0h headers (or above) is run
2291 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2292 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2296 * Removed the OS390-Unix config target. It relied on a script that doesn't
2301 * rsaz_1024_mul_avx2 overflow bug on x86_64
2303 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2304 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2305 Analysis suggests that attacks against RSA and DSA as a result of this
2306 defect would be very difficult to perform and are not believed likely.
2307 Attacks against DH1024 are considered just feasible, because most of the
2308 work necessary to deduce information about a private key may be performed
2309 offline. The amount of resources required for such an attack would be
2310 significant. However, for an attack on TLS to be meaningful, the server
2311 would have to share the DH1024 private key among multiple clients, which is
2312 no longer an option since CVE-2016-0701.
2314 This only affects processors that support the AVX2 but not ADX extensions
2315 like Intel Haswell (4th generation).
2317 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2318 was originally found via the OSS-Fuzz project.
2323 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2325 * bn_sqrx8x_internal carry bug on x86_64
2327 There is a carry propagating bug in the x86_64 Montgomery squaring
2328 procedure. No EC algorithms are affected. Analysis suggests that attacks
2329 against RSA and DSA as a result of this defect would be very difficult to
2330 perform and are not believed likely. Attacks against DH are considered just
2331 feasible (although very difficult) because most of the work necessary to
2332 deduce information about a private key may be performed offline. The amount
2333 of resources required for such an attack would be very significant and
2334 likely only accessible to a limited number of attackers. An attacker would
2335 additionally need online access to an unpatched system using the target
2336 private key in a scenario with persistent DH parameters and a private
2337 key that is shared between multiple clients.
2339 This only affects processors that support the BMI1, BMI2 and ADX extensions
2340 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2342 This issue was reported to OpenSSL by the OSS-Fuzz project.
2347 * Malformed X.509 IPAddressFamily could cause OOB read
2349 If an X.509 certificate has a malformed IPAddressFamily extension,
2350 OpenSSL could do a one-byte buffer overread. The most likely result
2351 would be an erroneous display of the certificate in text format.
2353 This issue was reported to OpenSSL by the OSS-Fuzz project.
2358 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2360 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2361 platform rather than 'mingw'.
2365 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2366 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2367 which is the minimum version we support.
2371 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2373 * Encrypt-Then-Mac renegotiation crash
2375 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2376 negotiated where it was not in the original handshake (or vice-versa) then
2377 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2378 and servers are affected.
2380 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2385 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2387 * Truncated packet could crash via OOB read
2389 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2390 cipher is being used, then a truncated packet can cause that host to
2391 perform an out-of-bounds read, usually resulting in a crash.
2393 This issue was reported to OpenSSL by Robert Święcki of Google.
2398 * Bad (EC)DHE parameters cause a client crash
2400 If a malicious server supplies bad parameters for a DHE or ECDHE key
2401 exchange then this can result in the client attempting to dereference a
2402 NULL pointer leading to a client crash. This could be exploited in a Denial
2405 This issue was reported to OpenSSL by Guido Vranken.
2410 * BN_mod_exp may produce incorrect results on x86_64
2412 There is a carry propagating bug in the x86_64 Montgomery squaring
2413 procedure. No EC algorithms are affected. Analysis suggests that attacks
2414 against RSA and DSA as a result of this defect would be very difficult to
2415 perform and are not believed likely. Attacks against DH are considered just
2416 feasible (although very difficult) because most of the work necessary to
2417 deduce information about a private key may be performed offline. The amount
2418 of resources required for such an attack would be very significant and
2419 likely only accessible to a limited number of attackers. An attacker would
2420 additionally need online access to an unpatched system using the target
2421 private key in a scenario with persistent DH parameters and a private
2422 key that is shared between multiple clients. For example this can occur by
2423 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2424 similar to CVE-2015-3193 but must be treated as a separate problem.
2426 This issue was reported to OpenSSL by the OSS-Fuzz project.
2431 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2433 * ChaCha20/Poly1305 heap-buffer-overflow
2435 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2436 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2437 crash. This issue is not considered to be exploitable beyond a DoS.
2439 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2444 * CMS Null dereference
2446 Applications parsing invalid CMS structures can crash with a NULL pointer
2447 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2448 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2449 structure callback if an attempt is made to free certain invalid encodings.
2450 Only CHOICE structures using a callback which do not handle NULL value are
2453 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2458 * Montgomery multiplication may produce incorrect results
2460 There is a carry propagating bug in the Broadwell-specific Montgomery
2461 multiplication procedure that handles input lengths divisible by, but
2462 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2463 and DH private keys are impossible. This is because the subroutine in
2464 question is not used in operations with the private key itself and an input
2465 of the attacker's direct choice. Otherwise the bug can manifest itself as
2466 transient authentication and key negotiation failures or reproducible
2467 erroneous outcome of public-key operations with specially crafted input.
2468 Among EC algorithms only Brainpool P-512 curves are affected and one
2469 presumably can attack ECDH key negotiation. Impact was not analyzed in
2470 detail, because pre-requisites for attack are considered unlikely. Namely
2471 multiple clients have to choose the curve in question and the server has to
2472 share the private key among them, neither of which is default behaviour.
2473 Even then only clients that chose the curve will be affected.
2475 This issue was publicly reported as transient failures and was not
2476 initially recognized as a security issue. Thanks to Richard Morgan for
2477 providing reproducible case.
2482 * Removed automatic addition of RPATH in shared libraries and executables,
2483 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2487 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2489 * Fix Use After Free for large message sizes
2491 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2492 message larger than approx 16k is received then the underlying buffer to
2493 store the incoming message is reallocated and moved. Unfortunately a
2494 dangling pointer to the old location is left which results in an attempt to
2495 write to the previously freed location. This is likely to result in a
2496 crash, however it could potentially lead to execution of arbitrary code.
2498 This issue only affects OpenSSL 1.1.0a.
2500 This issue was reported to OpenSSL by Robert Święcki.
2505 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2507 * OCSP Status Request extension unbounded memory growth
2509 A malicious client can send an excessively large OCSP Status Request
2510 extension. If that client continually requests renegotiation, sending a
2511 large OCSP Status Request extension each time, then there will be unbounded
2512 memory growth on the server. This will eventually lead to a Denial Of
2513 Service attack through memory exhaustion. Servers with a default
2514 configuration are vulnerable even if they do not support OCSP. Builds using
2515 the "no-ocsp" build time option are not affected.
2517 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2522 * SSL_peek() hang on empty record
2524 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2525 sends an empty record. This could be exploited by a malicious peer in a
2526 Denial Of Service attack.
2528 This issue was reported to OpenSSL by Alex Gaynor.
2533 * Excessive allocation of memory in tls_get_message_header() and
2534 dtls1_preprocess_fragment()
2536 A (D)TLS message includes 3 bytes for its length in the header for the
2537 message. This would allow for messages up to 16Mb in length. Messages of
2538 this length are excessive and OpenSSL includes a check to ensure that a
2539 peer is sending reasonably sized messages in order to avoid too much memory
2540 being consumed to service a connection. A flaw in the logic of version
2541 1.1.0 means that memory for the message is allocated too early, prior to
2542 the excessive message length check. Due to way memory is allocated in
2543 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2544 to service a connection. This could lead to a Denial of Service through
2545 memory exhaustion. However, the excessive message length check still takes
2546 place, and this would cause the connection to immediately fail. Assuming
2547 that the application calls SSL_free() on the failed connection in a timely
2548 manner then the 21Mb of allocated memory will then be immediately freed
2549 again. Therefore the excessive memory allocation will be transitory in
2550 nature. This then means that there is only a security impact if:
2552 1) The application does not call SSL_free() in a timely manner in the event
2553 that the connection fails
2555 2) The application is working in a constrained environment where there is
2556 very little free memory
2558 3) The attacker initiates multiple connection attempts such that there are
2559 multiple connections in a state where memory has been allocated for the
2560 connection; SSL_free() has not yet been called; and there is insufficient
2561 memory to service the multiple requests.
2563 Except in the instance of (1) above any Denial Of Service is likely to be
2564 transitory because as soon as the connection fails the memory is
2565 subsequently freed again in the SSL_free() call. However there is an
2566 increased risk during this period of application crashes due to the lack of
2567 memory - which would then mean a more serious Denial of Service.
2569 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2570 (CVE-2016-6307 and CVE-2016-6308)
2574 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2575 had to be removed. Primary reason is that vendor assembler can't
2576 assemble our modules with -KPIC flag. As result it, assembly
2577 support, was not even available as option. But its lack means
2578 lack of side-channel resistant code, which is incompatible with
2579 security by todays standards. Fortunately gcc is readily available
2580 prepackaged option, which we firmly point at...
2584 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2586 * Windows command-line tool supports UTF-8 opt-in option for arguments
2587 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2588 (to any value) allows Windows user to access PKCS#12 file generated
2589 with Windows CryptoAPI and protected with non-ASCII password, as well
2590 as files generated under UTF-8 locale on Linux also protected with
2595 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2596 have been disabled by default and removed from DEFAULT, just like RC4.
2597 See the RC4 item below to re-enable both.
2601 * The method for finding the storage location for the Windows RAND seed file
2602 has changed. First we check %RANDFILE%. If that is not set then we check
2603 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2604 all else fails we fall back to C:\.
2608 * The EVP_EncryptUpdate() function has had its return type changed from void
2609 to int. A return of 0 indicates and error while a return of 1 indicates
2614 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2615 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2616 off the constant time implementation for RSA, DSA and DH have been made
2617 no-ops and deprecated.
2621 * Windows RAND implementation was simplified to only get entropy by
2622 calling CryptGenRandom(). Various other RAND-related tickets
2625 *Joseph Wylie Yandle, Rich Salz*
2627 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2628 and `OPENSSL_LH_`, respectively. The old names are available
2629 with API compatibility. They new names are now completely documented.
2633 * Unify TYPE_up_ref(obj) methods signature.
2634 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2635 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2636 int (instead of void) like all others TYPE_up_ref() methods.
2637 So now these methods also check the return value of CRYPTO_atomic_add(),
2638 and the validity of object reference counter.
2640 *fdasilvayy@gmail.com*
2642 * With Windows Visual Studio builds, the .pdb files are installed
2643 alongside the installed libraries and executables. For a static
2644 library installation, ossl_static.pdb is the associate compiler
2645 generated .pdb file to be used when linking programs.
2649 * Remove openssl.spec. Packaging files belong with the packagers.
2653 * Automatic Darwin/OSX configuration has had a refresh, it will now
2654 recognise x86_64 architectures automatically. You can still decide
2655 to build for a different bitness with the environment variable
2656 KERNEL_BITS (can be 32 or 64), for example:
2658 KERNEL_BITS=32 ./config
2662 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2663 256 bit AES and HMAC with SHA256.
2667 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2671 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2675 * To enable users to have their own config files and build file templates,
2676 Configure looks in the directory indicated by the environment variable
2677 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2678 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2679 name and is used as is.
2683 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2684 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2685 X509_CERT_FILE_CTX was removed.
2689 * "shared" builds are now the default. To create only static libraries use
2690 the "no-shared" Configure option.
2694 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2695 All of these option have not worked for some while and are fundamental
2700 * Make various cleanup routines no-ops and mark them as deprecated. Most
2701 global cleanup functions are no longer required because they are handled
2702 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2703 Explicitly de-initing can cause problems (e.g. where a library that uses
2704 OpenSSL de-inits, but an application is still using it). The affected
2705 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2706 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2707 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2708 COMP_zlib_cleanup().
2712 * --strict-warnings no longer enables runtime debugging options
2713 such as REF_DEBUG. Instead, debug options are automatically
2714 enabled with '--debug' builds.
2716 *Andy Polyakov, Emilia Käsper*
2718 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2719 have been moved out of the public header files. New functions for managing
2720 these have been added.
2724 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2725 objects have been moved out of the public header files. New
2726 functions for managing these have been added.
2730 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2731 have been moved out of the public header files. New functions for managing
2732 these have been added.
2736 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2737 moved out of the public header files. New functions for managing these
2742 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2746 * Removed the mk1mf build scripts.
2750 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2751 it is always safe to #include a header now.
2755 * Removed the aged BC-32 config and all its supporting scripts
2759 * Removed support for Ultrix, Netware, and OS/2.
2763 * Add support for HKDF.
2765 *Alessandro Ghedini*
2767 * Add support for blake2b and blake2s
2771 * Added support for "pipelining". Ciphers that have the
2772 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2773 encryptions/decryptions simultaneously. There are currently no built-in
2774 ciphers with this property but the expectation is that engines will be able
2775 to offer it to significantly improve throughput. Support has been extended
2776 into libssl so that multiple records for a single connection can be
2777 processed in one go (for >=TLS 1.1).
2781 * Added the AFALG engine. This is an async capable engine which is able to
2782 offload work to the Linux kernel. In this initial version it only supports
2783 AES128-CBC. The kernel must be version 4.1.0 or greater.
2787 * OpenSSL now uses a new threading API. It is no longer necessary to
2788 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2789 are two supported threading models: pthreads and windows threads. It is
2790 also possible to configure OpenSSL at compile time for "no-threads". The
2791 old threading API should no longer be used. The functions have been
2792 replaced with "no-op" compatibility macros.
2794 *Alessandro Ghedini, Matt Caswell*
2796 * Modify behavior of ALPN to invoke callback after SNI/servername
2797 callback, such that updates to the SSL_CTX affect ALPN.
2801 * Add SSL_CIPHER queries for authentication and key-exchange.
2805 * Changes to the DEFAULT cipherlist:
2806 - Prefer (EC)DHE handshakes over plain RSA.
2807 - Prefer AEAD ciphers over legacy ciphers.
2808 - Prefer ECDSA over RSA when both certificates are available.
2809 - Prefer TLSv1.2 ciphers/PRF.
2810 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2815 * Change the ECC default curve list to be this, in order: x25519,
2816 secp256r1, secp521r1, secp384r1.
2820 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2821 disabled by default. They can be re-enabled using the
2822 enable-weak-ssl-ciphers option to Configure.
2826 * If the server has ALPN configured, but supports no protocols that the
2827 client advertises, send a fatal "no_application_protocol" alert.
2828 This behaviour is SHALL in RFC 7301, though it isn't universally
2829 implemented by other servers.
2833 * Add X25519 support.
2834 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2835 for public and private key encoding using the format documented in
2836 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2837 key generation and key derivation.
2839 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2844 * Deprecate SRP_VBASE_get_by_user.
2845 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2846 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2847 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2848 seed, even if the seed is configured.
2850 Users should use SRP_VBASE_get1_by_user instead. Note that in
2851 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2852 also that even though configuring the SRP seed attempts to hide
2853 invalid usernames by continuing the handshake with fake
2854 credentials, this behaviour is not constant time and no strong
2855 guarantees are made that the handshake is indistinguishable from
2856 that of a valid user.
2860 * Configuration change; it's now possible to build dynamic engines
2861 without having to build shared libraries and vice versa. This
2862 only applies to the engines in engines/, those in crypto/engine/
2863 will always be built into libcrypto (i.e. "static").
2865 Building dynamic engines is enabled by default; to disable, use
2866 the configuration option "disable-dynamic-engine".
2868 The only requirements for building dynamic engines are the
2869 presence of the DSO module and building with position independent
2870 code, so they will also automatically be disabled if configuring
2871 with "disable-dso" or "disable-pic".
2873 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2874 are also taken away from openssl/opensslconf.h, as they are
2879 * Configuration change; if there is a known flag to compile
2880 position independent code, it will always be applied on the
2881 libcrypto and libssl object files, and never on the application
2882 object files. This means other libraries that use routines from
2883 libcrypto / libssl can be made into shared libraries regardless
2884 of how OpenSSL was configured.
2886 If this isn't desirable, the configuration options "disable-pic"
2887 or "no-pic" can be used to disable the use of PIC. This will
2888 also disable building shared libraries and dynamic engines.
2892 * Removed JPAKE code. It was experimental and has no wide use.
2896 * The INSTALL_PREFIX Makefile variable has been renamed to
2897 DESTDIR. That makes for less confusion on what this variable
2898 is for. Also, the configuration option --install_prefix is
2903 * Heartbeat for TLS has been removed and is disabled by default
2904 for DTLS; configure with enable-heartbeats. Code that uses the
2905 old #define's might need to be updated.
2907 *Emilia Käsper, Rich Salz*
2909 * Rename REF_CHECK to REF_DEBUG.
2913 * New "unified" build system
2915 The "unified" build system is aimed to be a common system for all
2916 platforms we support. With it comes new support for VMS.
2918 This system builds supports building in a different directory tree
2919 than the source tree. It produces one Makefile (for unix family
2920 or lookalikes), or one descrip.mms (for VMS).
2922 The source of information to make the Makefile / descrip.mms is
2923 small files called 'build.info', holding the necessary
2924 information for each directory with source to compile, and a
2925 template in Configurations, like unix-Makefile.tmpl or
2928 With this change, the library names were also renamed on Windows
2929 and on VMS. They now have names that are closer to the standard
2930 on Unix, and include the major version number, and in certain
2931 cases, the architecture they are built for. See "Notes on shared
2932 libraries" in INSTALL.
2934 We rely heavily on the perl module Text::Template.
2938 * Added support for auto-initialisation and de-initialisation of the library.
2939 OpenSSL no longer requires explicit init or deinit routines to be called,
2940 except in certain circumstances. See the OPENSSL_init_crypto() and
2941 OPENSSL_init_ssl() man pages for further information.
2945 * The arguments to the DTLSv1_listen function have changed. Specifically the
2946 "peer" argument is now expected to be a BIO_ADDR object.
2948 * Rewrite of BIO networking library. The BIO library lacked consistent
2949 support of IPv6, and adding it required some more extensive
2950 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2951 which hold all types of addresses and chains of address information.
2952 It also introduces a new API, with functions like BIO_socket,
2953 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2954 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2955 have been adapted accordingly.
2959 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2964 * CRIME protection: disable compression by default, even if OpenSSL is
2965 compiled with zlib enabled. Applications can still enable compression
2966 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2967 using the SSL_CONF library to configure compression.
2971 * The signature of the session callback configured with
2972 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2973 was explicitly marked as `const unsigned char*` instead of
2978 * Always DPURIFY. Remove the use of uninitialized memory in the
2979 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2983 * Removed many obsolete configuration items, including
2984 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2985 MD2_CHAR, MD2_INT, MD2_LONG
2987 IDEA_SHORT, IDEA_LONG
2988 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2990 *Rich Salz, with advice from Andy Polyakov*
2992 * Many BN internals have been moved to an internal header file.
2994 *Rich Salz with help from Andy Polyakov*
2996 * Configuration and writing out the results from it has changed.
2997 Files such as Makefile include/openssl/opensslconf.h and are now
2998 produced through general templates, such as Makefile.in and
2999 crypto/opensslconf.h.in and some help from the perl module
3002 Also, the center of configuration information is no longer
3003 Makefile. Instead, Configure produces a perl module in
3004 configdata.pm which holds most of the config data (in the hash
3005 table %config), the target data that comes from the target
3006 configuration in one of the `Configurations/*.conf` files (in
3011 * To clarify their intended purposes, the Configure options
3012 --prefix and --openssldir change their semantics, and become more
3013 straightforward and less interdependent.
3015 --prefix shall be used exclusively to give the location INSTALLTOP
3016 where programs, scripts, libraries, include files and manuals are
3017 going to be installed. The default is now /usr/local.
3019 --openssldir shall be used exclusively to give the default
3020 location OPENSSLDIR where certificates, private keys, CRLs are
3021 managed. This is also where the default openssl.cnf gets
3023 If the directory given with this option is a relative path, the
3024 values of both the --prefix value and the --openssldir value will
3025 be combined to become OPENSSLDIR.
3026 The default for --openssldir is INSTALLTOP/ssl.
3028 Anyone who uses --openssldir to specify where OpenSSL is to be
3029 installed MUST change to use --prefix instead.
3033 * The GOST engine was out of date and therefore it has been removed. An up
3034 to date GOST engine is now being maintained in an external repository.
3035 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3036 support for GOST ciphersuites (these are only activated if a GOST engine
3041 * EGD is no longer supported by default; use enable-egd when
3044 *Ben Kaduk and Rich Salz*
3046 * The distribution now has Makefile.in files, which are used to
3047 create Makefile's when Configure is run. *Configure must be run
3048 before trying to build now.*
3052 * The return value for SSL_CIPHER_description() for error conditions
3057 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3059 Obtaining and performing DNSSEC validation of TLSA records is
3060 the application's responsibility. The application provides
3061 the TLSA records of its choice to OpenSSL, and these are then
3062 used to authenticate the peer.
3064 The TLSA records need not even come from DNS. They can, for
3065 example, be used to implement local end-entity certificate or
3066 trust-anchor "pinning", where the "pin" data takes the form
3067 of TLSA records, which can augment or replace verification
3068 based on the usual WebPKI public certification authorities.
3072 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3073 continues to support deprecated interfaces in default builds.
3074 However, applications are strongly advised to compile their
3075 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3076 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3077 or the 1.1.0 releases.
3079 In environments in which all applications have been ported to
3080 not use any deprecated interfaces OpenSSL's Configure script
3081 should be used with the --api=1.1.0 option to entirely remove
3082 support for the deprecated features from the library and
3083 unconditionally disable them in the installed headers.
3084 Essentially the same effect can be achieved with the "no-deprecated"
3085 argument to Configure, except that this will always restrict
3086 the build to just the latest API, rather than a fixed API
3089 As applications are ported to future revisions of the API,
3090 they should update their compile-time OPENSSL_API_COMPAT define
3091 accordingly, but in most cases should be able to continue to
3092 compile with later releases.
3094 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3095 0x10000000L and 0x00908000L, respectively. However those
3096 versions did not support the OPENSSL_API_COMPAT feature, and
3097 so applications are not typically tested for explicit support
3098 of just the undeprecated features of either release.
3102 * Add support for setting the minimum and maximum supported protocol.
3103 It can bet set via the SSL_set_min_proto_version() and
3104 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3105 MaxProtocol. It's recommended to use the new APIs to disable
3106 protocols instead of disabling individual protocols using
3107 SSL_set_options() or SSL_CONF's Protocol. This change also
3108 removes support for disabling TLS 1.2 in the OpenSSL TLS
3109 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3113 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3117 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3118 and integrates ECDSA and ECDH functionality into EC. Implementations can
3119 now redirect key generation and no longer need to convert to or from
3122 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3123 include the ec.h header file instead.
3127 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3128 ciphers who are no longer supported and drops support the ephemeral RSA key
3129 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3133 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3134 opaque. For HMAC_CTX, the following constructors and destructors
3137 HMAC_CTX *HMAC_CTX_new(void);
3138 void HMAC_CTX_free(HMAC_CTX *ctx);
3140 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3141 destroy such methods has been added. See EVP_MD_meth_new(3) and
3142 EVP_CIPHER_meth_new(3) for documentation.
3145 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3146 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3147 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3148 an already created structure.
3149 2) For consistency with the majority of our object creators and
3150 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3151 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3152 for deprecated builds.
3156 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3157 cryptographic operations to be performed asynchronously as long as an
3158 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3159 further details. Libssl has also had this capability integrated with the
3160 introduction of the new mode SSL_MODE_ASYNC and associated error
3161 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3162 pages. This work was developed in partnership with Intel Corp.
3166 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3167 always enabled now. If you want to disable the support you should
3168 exclude it using the list of supported ciphers. This also means that the
3169 "-no_ecdhe" option has been removed from s_server.
3173 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3174 SSL_{CTX_}set1_curves() which can set a list.
3178 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3179 curve you want to support using SSL_{CTX_}set1_curves().
3183 * State machine rewrite. The state machine code has been significantly
3184 refactored in order to remove much duplication of code and solve issues
3185 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3186 further details). This change does have some associated API changes.
3187 Notably the SSL_state() function has been removed and replaced by
3188 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3189 SSL_set_state() has been removed altogether. The previous handshake states
3190 defined in ssl.h and ssl3.h have also been removed.
3194 * All instances of the string "ssleay" in the public API were replaced
3195 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3196 Some error codes related to internal RSA_eay API's were renamed.
3200 * The demo files in crypto/threads were moved to demo/threads.
3204 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3207 *Matt Caswell, Rich Salz*
3209 * New ASN.1 embed macro.
3211 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3212 structure is not allocated: it is part of the parent. That is instead of
3220 This reduces memory fragmentation and make it impossible to accidentally
3221 set a mandatory field to NULL.
3223 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3224 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3225 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3230 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3234 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3235 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3236 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3237 DES and RC4 ciphersuites.
3241 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3242 This changes the decoding behaviour for some invalid messages,
3243 though the change is mostly in the more lenient direction, and
3244 legacy behaviour is preserved as much as possible.
3248 * Fix no-stdio build.
3249 *David Woodhouse <David.Woodhouse@intel.com> and also*
3250 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3252 * New testing framework
3253 The testing framework has been largely rewritten and is now using
3254 perl and the perl modules Test::Harness and an extended variant of
3255 Test::More called OpenSSL::Test to do its work. All test scripts in
3256 test/ have been rewritten into test recipes, and all direct calls to
3257 executables in test/Makefile have become individual recipes using the
3258 simplified testing OpenSSL::Test::Simple.
3260 For documentation on our testing modules, do:
3262 perldoc test/testlib/OpenSSL/Test/Simple.pm
3263 perldoc test/testlib/OpenSSL/Test.pm
3267 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3268 are used; the latter aborts on memory leaks (usually checked on exit).
3269 Some undocumented "set malloc, etc., hooks" functions were removed
3270 and others were changed. All are now documented.
3274 * In DSA_generate_parameters_ex, if the provided seed is too short,
3277 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3279 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3280 from RFC4279, RFC4785, RFC5487, RFC5489.
3282 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3283 original RSA_PSK patch.
3287 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3288 era flag was never set throughout the codebase (only read). Also removed
3289 SSL3_FLAGS_POP_BUFFER which was only used if
3290 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3294 * Changed the default name options in the "ca", "crl", "req" and "x509"
3295 to be "oneline" instead of "compat".
3299 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3300 not aware of clients that still exhibit this bug, and the workaround
3301 hasn't been working properly for a while.
3305 * The return type of BIO_number_read() and BIO_number_written() as well as
3306 the corresponding num_read and num_write members in the BIO structure has
3307 changed from unsigned long to uint64_t. On platforms where an unsigned
3308 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3313 * Given the pervasive nature of TLS extensions it is inadvisable to run
3314 OpenSSL without support for them. It also means that maintaining
3315 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3316 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3320 * Removed support for the two export grade static DH ciphersuites
3321 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3322 were newly added (along with a number of other static DH ciphersuites) to
3323 1.0.2. However the two export ones have *never* worked since they were
3324 introduced. It seems strange in any case to be adding new export
3325 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3329 * Version negotiation has been rewritten. In particular SSLv23_method(),
3330 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3331 and turned into macros which simply call the new preferred function names
3332 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3333 should use the new names instead. Also as part of this change the ssl23.h
3334 header file has been removed.
3338 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3339 code and the associated standard is no longer considered fit-for-purpose.
3343 * RT2547 was closed. When generating a private key, try to make the
3344 output file readable only by the owner. This behavior change might
3345 be noticeable when interacting with other software.
3347 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3352 * Added HTTP GET support to the ocsp command.
3356 * Changed default digest for the dgst and enc commands from MD5 to
3361 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3365 * Added support for TLS extended master secret from
3366 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3367 initial patch which was a great help during development.
3371 * All libssl internal structures have been removed from the public header
3372 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3373 now redundant). Users should not attempt to access internal structures
3374 directly. Instead they should use the provided API functions.
3378 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3379 Access to deprecated functions can be re-enabled by running config with
3380 "enable-deprecated". In addition applications wishing to use deprecated
3381 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3382 will, by default, disable some transitive includes that previously existed
3383 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3387 * Added support for OCB mode. OpenSSL has been granted a patent license
3388 compatible with the OpenSSL license for use of OCB. Details are available
3389 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3390 for OCB can be removed by calling config with no-ocb.
3394 * SSLv2 support has been removed. It still supports receiving a SSLv2
3395 compatible client hello.
3399 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3400 done while fixing the error code for the key-too-small case.
3402 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3404 * CA.sh has been removed; use CA.pl instead.
3408 * Removed old DES API.
3412 * Remove various unsupported platforms:
3418 Sinix/ReliantUNIX RM400
3423 16-bit platforms such as WIN16
3427 * Clean up OPENSSL_NO_xxx #define's
3428 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3429 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3430 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3431 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3432 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3433 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3434 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3435 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3436 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3437 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3441 * Cleaned up dead code
3442 Remove all but one '#ifdef undef' which is to be looked at.
3446 * Clean up calling of xxx_free routines.
3447 Just like free(), fix most of the xxx_free routines to accept
3448 NULL. Remove the non-null checks from callers. Save much code.
3452 * Add secure heap for storage of private keys (when possible).
3453 Add BIO_s_secmem(), CBIGNUM, etc.
3454 Contributed by Akamai Technologies under our Corporate CLA.
3458 * Experimental support for a new, fast, unbiased prime candidate generator,
3459 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3461 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3463 * New output format NSS in the sess_id command line tool. This allows
3464 exporting the session id and the master key in NSS keylog format.
3466 *Martin Kaiser <martin@kaiser.cx>*
3468 * Harmonize version and its documentation. -f flag is used to display
3471 *mancha <mancha1@zoho.com>*
3473 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3474 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3476 *mancha <mancha1@zoho.com>*
3478 * Fix some double frees. These are not thought to be exploitable.
3480 *mancha <mancha1@zoho.com>*
3482 * A missing bounds check in the handling of the TLS heartbeat extension
3483 can be used to reveal up to 64k of memory to a connected client or
3486 Thanks for Neel Mehta of Google Security for discovering this bug and to
3487 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3488 preparing the fix [CVE-2014-0160][]
3490 *Adam Langley, Bodo Moeller*
3492 * Fix for the attack described in the paper "Recovering OpenSSL
3493 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3494 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3495 <http://eprint.iacr.org/2014/140>
3497 Thanks to Yuval Yarom and Naomi Benger for discovering this
3498 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3500 *Yuval Yarom and Naomi Benger*
3502 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3503 this fixes a limitation in previous versions of OpenSSL.
3507 * Experimental encrypt-then-mac support.
3509 Experimental support for encrypt then mac from
3510 draft-gutmann-tls-encrypt-then-mac-02.txt
3512 To enable it set the appropriate extension number (0x42 for the test
3513 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3515 For non-compliant peers (i.e. just about everything) this should have no
3518 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3522 * Add EVP support for key wrapping algorithms, to avoid problems with
3523 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3524 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3525 algorithms and include tests cases.
3529 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3534 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3535 MGF1 digest and OAEP label.
3539 * Make openssl verify return errors.
3541 *Chris Palmer <palmer@google.com> and Ben Laurie*
3543 * New function ASN1_TIME_diff to calculate the difference between two
3544 ASN1_TIME structures or one structure and the current time.
3548 * Update fips_test_suite to support multiple command line options. New
3549 test to induce all self test errors in sequence and check expected
3554 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3555 sign or verify all in one operation.
3559 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3560 test programs and fips_test_suite. Includes functionality to parse
3561 the minimal script output of fipsalgest.pl directly.
3565 * Add authorisation parameter to FIPS_module_mode_set().
3569 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3573 * Use separate DRBG fields for internal and external flags. New function
3574 FIPS_drbg_health_check() to perform on demand health checking. Add
3575 generation tests to fips_test_suite with reduced health check interval to
3576 demonstrate periodic health checking. Add "nodh" option to
3577 fips_test_suite to skip very slow DH test.
3581 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3586 * More extensive health check for DRBG checking many more failure modes.
3587 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3588 combination: call this in fips_test_suite.
3592 * Add support for canonical generation of DSA parameter 'g'. See
3595 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3596 POST to handle HMAC cases.
3600 * Add functions FIPS_module_version() and FIPS_module_version_text()
3601 to return numerical and string versions of the FIPS module number.
3605 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3606 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3607 outside the validated module in the FIPS capable OpenSSL.
3611 * Minor change to DRBG entropy callback semantics. In some cases
3612 there is no multiple of the block length between min_len and
3613 max_len. Allow the callback to return more than max_len bytes
3614 of entropy but discard any extra: it is the callback's responsibility
3615 to ensure that the extra data discarded does not impact the
3616 requested amount of entropy.
3620 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3621 information in FIPS186-3, SP800-57 and SP800-131A.
3625 * CCM support via EVP. Interface is very similar to GCM case except we
3626 must supply all data in one chunk (i.e. no update, final) and the
3627 message length must be supplied if AAD is used. Add algorithm test
3632 * Initial version of POST overhaul. Add POST callback to allow the status
3633 of POST to be monitored and/or failures induced. Modify fips_test_suite
3634 to use callback. Always run all selftests even if one fails.
3638 * XTS support including algorithm test driver in the fips_gcmtest program.
3639 Note: this does increase the maximum key length from 32 to 64 bytes but
3640 there should be no binary compatibility issues as existing applications
3641 will never use XTS mode.
3645 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3646 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3647 performs algorithm blocking for unapproved PRNG types. Also do not
3648 set PRNG type in FIPS_mode_set(): leave this to the application.
3649 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3650 the standard OpenSSL PRNG: set additional data to a date time vector.
3654 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3655 This shouldn't present any incompatibility problems because applications
3656 shouldn't be using these directly and any that are will need to rethink
3657 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3661 * Extensive self tests and health checking required by SP800-90 DRBG.
3662 Remove strength parameter from FIPS_drbg_instantiate and always
3663 instantiate at maximum supported strength.
3667 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3671 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3675 * New function DH_compute_key_padded() to compute a DH key and pad with
3676 leading zeroes if needed: this complies with SP800-56A et al.
3680 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3681 anything, incomplete, subject to change and largely untested at present.
3685 * Modify fipscanisteronly build option to only build the necessary object
3686 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3690 * Add experimental option FIPSSYMS to give all symbols in
3691 fipscanister.o and FIPS or fips prefix. This will avoid
3692 conflicts with future versions of OpenSSL. Add perl script
3693 util/fipsas.pl to preprocess assembly language source files
3694 and rename any affected symbols.
3698 * Add selftest checks and algorithm block of non-fips algorithms in
3699 FIPS mode. Remove DES2 from selftests.
3703 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3704 return internal method without any ENGINE dependencies. Add new
3705 tiny fips sign and verify functions.
3709 * New build option no-ec2m to disable characteristic 2 code.
3713 * New build option "fipscanisteronly". This only builds fipscanister.o
3714 and (currently) associated fips utilities. Uses the file Makefile.fips
3715 instead of Makefile.org as the prototype.
3719 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3720 Update fips_gcmtest to use IV generator.
3724 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3725 setting output buffer to NULL. The `*Final` function must be
3726 called although it will not retrieve any additional data. The tag
3727 can be set or retrieved with a ctrl. The IV length is by default 12
3728 bytes (96 bits) but can be set to an alternative value. If the IV
3729 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3734 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3735 underlying do_cipher function handles all cipher semantics itself
3736 including padding and finalisation. This is useful if (for example)
3737 an ENGINE cipher handles block padding itself. The behaviour of
3738 do_cipher is subtly changed if this flag is set: the return value
3739 is the number of characters written to the output buffer (zero is
3740 no longer an error code) or a negative error code. Also if the
3741 input buffer is NULL and length 0 finalisation should be performed.
3745 * If a candidate issuer certificate is already part of the constructed
3746 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3750 * Improve forward-security support: add functions
3752 void SSL_CTX_set_not_resumable_session_callback(
3753 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3754 void SSL_set_not_resumable_session_callback(
3755 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3757 for use by SSL/TLS servers; the callback function will be called whenever a
3758 new session is created, and gets to decide whether the session may be
3759 cached to make it resumable (return 0) or not (return 1). (As by the
3760 SSL/TLS protocol specifications, the session_id sent by the server will be
3761 empty to indicate that the session is not resumable; also, the server will
3762 not generate RFC 4507 (RFC 5077) session tickets.)
3764 A simple reasonable callback implementation is to return is_forward_secure.
3765 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3766 by the SSL/TLS server library, indicating whether it can provide forward
3769 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3771 * New -verify_name option in command line utilities to set verification
3776 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3777 Add CMAC pkey methods.
3781 * Experimental renegotiation in s_server -www mode. If the client
3782 browses /reneg connection is renegotiated. If /renegcert it is
3783 renegotiated requesting a certificate.
3787 * Add an "external" session cache for debugging purposes to s_server. This
3788 should help trace issues which normally are only apparent in deployed
3789 multi-process servers.
3793 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3794 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3795 BIO_set_cipher() and some obscure PEM functions were changed so they
3796 can now return an error. The RAND changes required a change to the
3797 RAND_METHOD structure.
3801 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3802 a gcc attribute to warn if the result of a function is ignored. This
3803 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3804 whose return value is often ignored.
3808 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3809 These allow SCTs (signed certificate timestamps) to be requested and
3810 validated when establishing a connection.
3812 *Rob Percival <robpercival@google.com>*
3817 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3819 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3820 used even when parsing explicit parameters, when loading a serialized key
3821 or calling `EC_GROUP_new_from_ecpkparameters()`/
3822 `EC_GROUP_new_from_ecparameters()`.
3823 This prevents bypass of security hardening and performance gains,
3824 especially for curves with specialized EC_METHODs.
3825 By default, if a key encoded with explicit parameters is loaded and later
3826 serialized, the output is still encoded with explicit parameters, even if
3827 internally a "named" EC_GROUP is used for computation.
3831 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3832 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3833 NULL. After this change, only the cofactor parameter can be NULL. It also
3834 does some minimal sanity checks on the passed order.
3839 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3840 An attack is simple, if the first CMS_recipientInfo is valid but the
3841 second CMS_recipientInfo is chosen ciphertext. If the second
3842 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3843 encryption key will be replaced by garbage, and the message cannot be
3844 decoded, but if the RSA decryption fails, the correct encryption key is
3845 used and the recipient will not notice the attack.
3846 As a work around for this potential attack the length of the decrypted
3847 key must be equal to the cipher default key length, in case the
3848 certifiate is not given and all recipientInfo are tried out.
3849 The old behaviour can be re-enabled in the CMS code by setting the
3850 CMS_DEBUG_DECRYPT flag.
3855 * Document issue with installation paths in diverse Windows builds
3857 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3858 binaries and run-time config file.
3863 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3865 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3866 This changes the size when using the genpkey app when no size is given. It
3867 fixes an omission in earlier changes that changed all RSA, DSA and DH
3868 generation apps to use 2048 bits by default.
3872 * Add FIPS support for Android Arm 64-bit
3874 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3875 Module in Version 2.0.10. For some reason, the corresponding target
3876 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3877 built with FIPS support on Android Arm 64-bit. This omission has been
3880 *Matthias St. Pierre*
3882 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3884 * 0-byte record padding oracle
3886 If an application encounters a fatal protocol error and then calls
3887 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3888 then OpenSSL can respond differently to the calling application if a 0 byte
3889 record is received with invalid padding compared to if a 0 byte record is
3890 received with an invalid MAC. If the application then behaves differently
3891 based on that in a way that is detectable to the remote peer, then this
3892 amounts to a padding oracle that could be used to decrypt data.
3894 In order for this to be exploitable "non-stitched" ciphersuites must be in
3895 use. Stitched ciphersuites are optimised implementations of certain
3896 commonly used ciphersuites. Also the application must call SSL_shutdown()
3897 twice even if a protocol error has occurred (applications should not do
3898 this but some do anyway).
3900 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3901 Aviram, with additional investigation by Steven Collison and Andrew
3902 Hourselt. It was reported to OpenSSL on 10th December 2018.
3907 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3911 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3913 * Microarchitecture timing vulnerability in ECC scalar multiplication
3915 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3916 shown to be vulnerable to a microarchitecture timing side channel attack.
3917 An attacker with sufficient access to mount local timing attacks during
3918 ECDSA signature generation could recover the private key.
3920 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3921 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3927 * Timing vulnerability in DSA signature generation
3929 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3930 timing side channel attack. An attacker could use variations in the signing
3931 algorithm to recover the private key.
3933 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3938 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3939 Module, accidentally introduced while backporting security fixes from the
3940 development branch and hindering the use of ECC in FIPS mode.
3944 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3946 * Client DoS due to large DH parameter
3948 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3949 malicious server can send a very large prime value to the client. This will
3950 cause the client to spend an unreasonably long period of time generating a
3951 key for this prime resulting in a hang until the client has finished. This
3952 could be exploited in a Denial Of Service attack.
3954 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3959 * Cache timing vulnerability in RSA Key Generation
3961 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3962 a cache timing side channel attack. An attacker with sufficient access to
3963 mount cache timing attacks during the RSA key generation process could
3964 recover the private key.
3966 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3967 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3972 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3973 parameter is no longer accepted, as it leads to a corrupt table. NULL
3974 pem_str is reserved for alias entries only.
3978 * Revert blinding in ECDSA sign and instead make problematic addition
3979 length-invariant. Switch even to fixed-length Montgomery multiplication.
3983 * Change generating and checking of primes so that the error rate of not
3984 being prime depends on the intended use based on the size of the input.
3985 For larger primes this will result in more rounds of Miller-Rabin.
3986 The maximal error rate for primes with more than 1080 bits is lowered
3989 *Kurt Roeckx, Annie Yousar*
3991 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3995 * Add blinding to ECDSA and DSA signatures to protect against side channel
3996 attacks discovered by Keegan Ryan (NCC Group).
4000 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4001 now allow empty (zero character) pass phrases.
4005 * Certificate time validation (X509_cmp_time) enforces stricter
4006 compliance with RFC 5280. Fractional seconds and timezone offsets
4007 are no longer allowed.
4011 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4013 * Constructed ASN.1 types with a recursive definition could exceed the stack
4015 Constructed ASN.1 types with a recursive definition (such as can be found
4016 in PKCS7) could eventually exceed the stack given malicious input with
4017 excessive recursion. This could result in a Denial Of Service attack. There
4018 are no such structures used within SSL/TLS that come from untrusted sources
4019 so this is considered safe.
4021 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4027 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4029 * Read/write after SSL object in error state
4031 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4032 mechanism. The intent was that if a fatal error occurred during a handshake
4033 then OpenSSL would move into the error state and would immediately fail if
4034 you attempted to continue the handshake. This works as designed for the
4035 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4036 SSL_connect()), however due to a bug it does not work correctly if
4037 SSL_read() or SSL_write() is called directly. In that scenario, if the
4038 handshake fails then a fatal error will be returned in the initial function
4039 call. If SSL_read()/SSL_write() is subsequently called by the application
4040 for the same SSL object then it will succeed and the data is passed without
4041 being decrypted/encrypted directly from the SSL/TLS record layer.
4043 In order to exploit this issue an application bug would have to be present
4044 that resulted in a call to SSL_read()/SSL_write() being issued after having
4045 already received a fatal error.
4047 This issue was reported to OpenSSL by David Benjamin (Google).
4052 * rsaz_1024_mul_avx2 overflow bug on x86_64
4054 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4055 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4056 Analysis suggests that attacks against RSA and DSA as a result of this
4057 defect would be very difficult to perform and are not believed likely.
4058 Attacks against DH1024 are considered just feasible, because most of the
4059 work necessary to deduce information about a private key may be performed
4060 offline. The amount of resources required for such an attack would be
4061 significant. However, for an attack on TLS to be meaningful, the server
4062 would have to share the DH1024 private key among multiple clients, which is
4063 no longer an option since CVE-2016-0701.
4065 This only affects processors that support the AVX2 but not ADX extensions
4066 like Intel Haswell (4th generation).
4068 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4069 was originally found via the OSS-Fuzz project.
4074 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4076 * bn_sqrx8x_internal carry bug on x86_64
4078 There is a carry propagating bug in the x86_64 Montgomery squaring
4079 procedure. No EC algorithms are affected. Analysis suggests that attacks
4080 against RSA and DSA as a result of this defect would be very difficult to
4081 perform and are not believed likely. Attacks against DH are considered just
4082 feasible (although very difficult) because most of the work necessary to
4083 deduce information about a private key may be performed offline. The amount
4084 of resources required for such an attack would be very significant and
4085 likely only accessible to a limited number of attackers. An attacker would
4086 additionally need online access to an unpatched system using the target
4087 private key in a scenario with persistent DH parameters and a private
4088 key that is shared between multiple clients.
4090 This only affects processors that support the BMI1, BMI2 and ADX extensions
4091 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4093 This issue was reported to OpenSSL by the OSS-Fuzz project.
4098 * Malformed X.509 IPAddressFamily could cause OOB read
4100 If an X.509 certificate has a malformed IPAddressFamily extension,
4101 OpenSSL could do a one-byte buffer overread. The most likely result
4102 would be an erroneous display of the certificate in text format.
4104 This issue was reported to OpenSSL by the OSS-Fuzz project.
4109 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4111 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4112 platform rather than 'mingw'.
4116 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4118 * Truncated packet could crash via OOB read
4120 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4121 cipher is being used, then a truncated packet can cause that host to
4122 perform an out-of-bounds read, usually resulting in a crash.
4124 This issue was reported to OpenSSL by Robert Święcki of Google.
4129 * BN_mod_exp may produce incorrect results on x86_64
4131 There is a carry propagating bug in the x86_64 Montgomery squaring
4132 procedure. No EC algorithms are affected. Analysis suggests that attacks
4133 against RSA and DSA as a result of this defect would be very difficult to
4134 perform and are not believed likely. Attacks against DH are considered just
4135 feasible (although very difficult) because most of the work necessary to
4136 deduce information about a private key may be performed offline. The amount
4137 of resources required for such an attack would be very significant and
4138 likely only accessible to a limited number of attackers. An attacker would
4139 additionally need online access to an unpatched system using the target
4140 private key in a scenario with persistent DH parameters and a private
4141 key that is shared between multiple clients. For example this can occur by
4142 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4143 similar to CVE-2015-3193 but must be treated as a separate problem.
4145 This issue was reported to OpenSSL by the OSS-Fuzz project.
4150 * Montgomery multiplication may produce incorrect results
4152 There is a carry propagating bug in the Broadwell-specific Montgomery
4153 multiplication procedure that handles input lengths divisible by, but
4154 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4155 and DH private keys are impossible. This is because the subroutine in
4156 question is not used in operations with the private key itself and an input
4157 of the attacker's direct choice. Otherwise the bug can manifest itself as
4158 transient authentication and key negotiation failures or reproducible
4159 erroneous outcome of public-key operations with specially crafted input.
4160 Among EC algorithms only Brainpool P-512 curves are affected and one
4161 presumably can attack ECDH key negotiation. Impact was not analyzed in
4162 detail, because pre-requisites for attack are considered unlikely. Namely
4163 multiple clients have to choose the curve in question and the server has to
4164 share the private key among them, neither of which is default behaviour.
4165 Even then only clients that chose the curve will be affected.
4167 This issue was publicly reported as transient failures and was not
4168 initially recognized as a security issue. Thanks to Richard Morgan for
4169 providing reproducible case.
4174 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4175 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4176 prevent issues where no progress is being made and the peer continually
4177 sends unrecognised record types, using up resources processing them.
4181 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4183 * Missing CRL sanity check
4185 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4186 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4187 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4189 This issue only affects the OpenSSL 1.0.2i
4194 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4196 * OCSP Status Request extension unbounded memory growth
4198 A malicious client can send an excessively large OCSP Status Request
4199 extension. If that client continually requests renegotiation, sending a
4200 large OCSP Status Request extension each time, then there will be unbounded
4201 memory growth on the server. This will eventually lead to a Denial Of
4202 Service attack through memory exhaustion. Servers with a default
4203 configuration are vulnerable even if they do not support OCSP. Builds using
4204 the "no-ocsp" build time option are not affected.
4206 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4211 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4214 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4220 * OOB write in MDC2_Update()
4222 An overflow can occur in MDC2_Update() either if called directly or
4223 through the EVP_DigestUpdate() function using MDC2. If an attacker
4224 is able to supply very large amounts of input data after a previous
4225 call to EVP_EncryptUpdate() with a partial block then a length check
4226 can overflow resulting in a heap corruption.
4228 The amount of data needed is comparable to SIZE_MAX which is impractical
4231 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4236 * Malformed SHA512 ticket DoS
4238 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4239 DoS attack where a malformed ticket will result in an OOB read which will
4242 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4243 a custom server callback and ticket lookup mechanism.
4245 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4250 * OOB write in BN_bn2dec()
4252 The function BN_bn2dec() does not check the return value of BN_div_word().
4253 This can cause an OOB write if an application uses this function with an
4254 overly large BIGNUM. This could be a problem if an overly large certificate
4255 or CRL is printed out from an untrusted source. TLS is not affected because
4256 record limits will reject an oversized certificate before it is parsed.
4258 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4263 * OOB read in TS_OBJ_print_bio()
4265 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4266 the total length the OID text representation would use and not the amount
4267 of data written. This will result in OOB reads when large OIDs are
4270 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4275 * Pointer arithmetic undefined behaviour
4277 Avoid some undefined pointer arithmetic
4279 A common idiom in the codebase is to check limits in the following manner:
4282 Where "p" points to some malloc'd data of SIZE bytes and
4285 "len" here could be from some externally supplied data (e.g. from a TLS
4288 The rules of C pointer arithmetic are such that "p + len" is only well
4289 defined where len <= SIZE. Therefore the above idiom is actually
4290 undefined behaviour.
4292 For example this could cause problems if some malloc implementation
4293 provides an address for "p" such that "p + len" actually overflows for
4294 values of len that are too big and therefore p + len < limit.
4296 This issue was reported to OpenSSL by Guido Vranken
4301 * Constant time flag not preserved in DSA signing
4303 Operations in the DSA signing algorithm should run in constant time in
4304 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4305 implementation means that a non-constant time codepath is followed for
4306 certain operations. This has been demonstrated through a cache-timing
4307 attack to be sufficient for an attacker to recover the private DSA key.
4309 This issue was reported by César Pereida (Aalto University), Billy Brumley
4310 (Tampere University of Technology), and Yuval Yarom (The University of
4311 Adelaide and NICTA).
4316 * DTLS buffered message DoS
4318 In a DTLS connection where handshake messages are delivered out-of-order
4319 those messages that OpenSSL is not yet ready to process will be buffered
4320 for later use. Under certain circumstances, a flaw in the logic means that
4321 those messages do not get removed from the buffer even though the handshake
4322 has been completed. An attacker could force up to approx. 15 messages to
4323 remain in the buffer when they are no longer required. These messages will
4324 be cleared when the DTLS connection is closed. The default maximum size for
4325 a message is 100k. Therefore the attacker could force an additional 1500k
4326 to be consumed per connection. By opening many simulataneous connections an
4327 attacker could cause a DoS attack through memory exhaustion.
4329 This issue was reported to OpenSSL by Quan Luo.
4334 * DTLS replay protection DoS
4336 A flaw in the DTLS replay attack protection mechanism means that records
4337 that arrive for future epochs update the replay protection "window" before
4338 the MAC for the record has been validated. This could be exploited by an
4339 attacker by sending a record for the next epoch (which does not have to
4340 decrypt or have a valid MAC), with a very large sequence number. This means
4341 that all subsequent legitimate packets are dropped causing a denial of
4342 service for a specific DTLS connection.
4344 This issue was reported to OpenSSL by the OCAP audit team.
4349 * Certificate message OOB reads
4351 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4352 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4353 theoretical DoS risk but this has not been observed in practice on common
4356 The messages affected are client certificate, client certificate request
4357 and server certificate. As a result the attack can only be performed
4358 against a client or a server which enables client authentication.
4360 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4365 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4367 * Prevent padding oracle in AES-NI CBC MAC check
4369 A MITM attacker can use a padding oracle attack to decrypt traffic
4370 when the connection uses an AES CBC cipher and the server support
4373 This issue was introduced as part of the fix for Lucky 13 padding
4374 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4375 constant time by making sure that always the same bytes are read and
4376 compared against either the MAC or padding bytes. But it no longer
4377 checked that there was enough data to have both the MAC and padding
4380 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4385 * Fix EVP_EncodeUpdate overflow
4387 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4388 Base64 encoding of binary data. If an attacker is able to supply very large
4389 amounts of input data then a length check can overflow resulting in a heap
4392 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4393 the `PEM_write_bio*` family of functions. These are mainly used within the
4394 OpenSSL command line applications, so any application which processes data
4395 from an untrusted source and outputs it as a PEM file should be considered
4396 vulnerable to this issue. User applications that call these APIs directly
4397 with large amounts of untrusted data may also be vulnerable.
4399 This issue was reported by Guido Vranken.
4404 * Fix EVP_EncryptUpdate overflow
4406 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4407 is able to supply very large amounts of input data after a previous call to
4408 EVP_EncryptUpdate() with a partial block then a length check can overflow
4409 resulting in a heap corruption. Following an analysis of all OpenSSL
4410 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4411 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4412 the first called function after an EVP_EncryptInit(), and therefore that
4413 specific call must be safe. The second form is where the length passed to
4414 EVP_EncryptUpdate() can be seen from the code to be some small value and
4415 therefore there is no possibility of an overflow. Since all instances are
4416 one of these two forms, it is believed that there can be no overflows in
4417 internal code due to this problem. It should be noted that
4418 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4419 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4420 of these calls have also been analysed too and it is believed there are no
4421 instances in internal usage where an overflow could occur.
4423 This issue was reported by Guido Vranken.
4428 * Prevent ASN.1 BIO excessive memory allocation
4430 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4431 a short invalid encoding can cause allocation of large amounts of memory
4432 potentially consuming excessive resources or exhausting memory.
4434 Any application parsing untrusted data through d2i BIO functions is
4435 affected. The memory based functions such as d2i_X509() are *not* affected.
4436 Since the memory based functions are used by the TLS library, TLS
4437 applications are not affected.
4439 This issue was reported by Brian Carpenter.
4446 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4447 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4448 in arbitrary stack data being returned in the buffer.
4450 This issue was reported by Guido Vranken.
4455 * Modify behavior of ALPN to invoke callback after SNI/servername
4456 callback, such that updates to the SSL_CTX affect ALPN.
4460 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4465 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4466 methods are enabled and ssl2 is disabled the methods return NULL.
4470 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4472 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4473 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4474 provide any "EXPORT" or "LOW" strength ciphers.
4478 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4479 is by default disabled at build-time. Builds that are not configured with
4480 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4481 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4482 will need to explicitly call either of:
4484 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4486 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4488 as appropriate. Even if either of those is used, or the application
4489 explicitly uses the version-specific SSLv2_method() or its client and
4490 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4491 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4492 ciphers, and SSLv2 56-bit DES are no longer available.
4497 * Fix a double-free in DSA code
4499 A double free bug was discovered when OpenSSL parses malformed DSA private
4500 keys and could lead to a DoS attack or memory corruption for applications
4501 that receive DSA private keys from untrusted sources. This scenario is
4504 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4510 * Disable SRP fake user seed to address a server memory leak.
4512 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4514 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4515 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4516 was changed to ignore the "fake user" SRP seed, even if the seed
4519 Users should use SRP_VBASE_get1_by_user instead. Note that in
4520 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4521 also that even though configuring the SRP seed attempts to hide
4522 invalid usernames by continuing the handshake with fake
4523 credentials, this behaviour is not constant time and no strong
4524 guarantees are made that the handshake is indistinguishable from
4525 that of a valid user.
4530 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4532 In the BN_hex2bn function the number of hex digits is calculated using an
4533 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4534 large values of `i` this can result in `bn_expand` not allocating any
4535 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4536 field as NULL leading to a subsequent NULL ptr deref. For very large values
4537 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4538 In this case memory is allocated to the internal BIGNUM data field, but it
4539 is insufficiently sized leading to heap corruption. A similar issue exists
4540 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4541 is ever called by user applications with very large untrusted hex/dec data.
4542 This is anticipated to be a rare occurrence.
4544 All OpenSSL internal usage of these functions use data that is not expected
4545 to be untrusted, e.g. config file data or application command line
4546 arguments. If user developed applications generate config file data based
4547 on untrusted data then it is possible that this could also lead to security
4548 consequences. This is also anticipated to be rare.
4550 This issue was reported to OpenSSL by Guido Vranken.
4555 * Fix memory issues in `BIO_*printf` functions
4557 The internal `fmtstr` function used in processing a "%s" format string in
4558 the `BIO_*printf` functions could overflow while calculating the length of a
4559 string and cause an OOB read when printing very long strings.
4561 Additionally the internal `doapr_outch` function can attempt to write to an
4562 OOB memory location (at an offset from the NULL pointer) in the event of a
4563 memory allocation failure. In 1.0.2 and below this could be caused where
4564 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4565 could be in processing a very long "%s" format string. Memory leaks can
4568 The first issue may mask the second issue dependent on compiler behaviour.
4569 These problems could enable attacks where large amounts of untrusted data
4570 is passed to the `BIO_*printf` functions. If applications use these functions
4571 in this way then they could be vulnerable. OpenSSL itself uses these
4572 functions when printing out human-readable dumps of ASN.1 data. Therefore
4573 applications that print this data could be vulnerable if the data is from
4574 untrusted sources. OpenSSL command line applications could also be
4575 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4576 as command line arguments.
4578 Libssl is not considered directly vulnerable. Additionally certificates etc
4579 received via remote connections via libssl are also unlikely to be able to
4580 trigger these issues because of message size limits enforced within libssl.
4582 This issue was reported to OpenSSL Guido Vranken.
4587 * Side channel attack on modular exponentiation
4589 A side-channel attack was found which makes use of cache-bank conflicts on
4590 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4591 of RSA keys. The ability to exploit this issue is limited as it relies on
4592 an attacker who has control of code in a thread running on the same
4593 hyper-threaded core as the victim thread which is performing decryptions.
4595 This issue was reported to OpenSSL by Yuval Yarom, The University of
4596 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4597 Nadia Heninger, University of Pennsylvania with more information at
4598 <http://cachebleed.info>.
4603 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4604 if no keysize is specified with default_bits. This fixes an
4605 omission in an earlier change that changed all RSA/DSA key generation
4606 apps to use 2048 bits by default.
4610 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4612 * DH small subgroups
4614 Historically OpenSSL only ever generated DH parameters based on "safe"
4615 primes. More recently (in version 1.0.2) support was provided for
4616 generating X9.42 style parameter files such as those required for RFC 5114
4617 support. The primes used in such files may not be "safe". Where an
4618 application is using DH configured with parameters based on primes that are
4619 not "safe" then an attacker could use this fact to find a peer's private
4620 DH exponent. This attack requires that the attacker complete multiple
4621 handshakes in which the peer uses the same private DH exponent. For example
4622 this could be used to discover a TLS server's private DH exponent if it's
4623 reusing the private DH exponent or it's using a static DH ciphersuite.
4625 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4626 TLS. It is not on by default. If the option is not set then the server
4627 reuses the same private DH exponent for the life of the server process and
4628 would be vulnerable to this attack. It is believed that many popular
4629 applications do set this option and would therefore not be at risk.
4631 The fix for this issue adds an additional check where a "q" parameter is
4632 available (as is the case in X9.42 based parameters). This detects the
4633 only known attack, and is the only possible defense for static DH
4634 ciphersuites. This could have some performance impact.
4636 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4637 default and cannot be disabled. This could have some performance impact.
4639 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4644 * SSLv2 doesn't block disabled ciphers
4646 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4647 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4648 been disabled, provided that the SSLv2 protocol was not also disabled via
4651 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4652 and Sebastian Schinzel.
4657 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4659 * BN_mod_exp may produce incorrect results on x86_64
4661 There is a carry propagating bug in the x86_64 Montgomery squaring
4662 procedure. No EC algorithms are affected. Analysis suggests that attacks
4663 against RSA and DSA as a result of this defect would be very difficult to
4664 perform and are not believed likely. Attacks against DH are considered just
4665 feasible (although very difficult) because most of the work necessary to
4666 deduce information about a private key may be performed offline. The amount
4667 of resources required for such an attack would be very significant and
4668 likely only accessible to a limited number of attackers. An attacker would
4669 additionally need online access to an unpatched system using the target
4670 private key in a scenario with persistent DH parameters and a private
4671 key that is shared between multiple clients. For example this can occur by
4672 default in OpenSSL DHE based SSL/TLS ciphersuites.
4674 This issue was reported to OpenSSL by Hanno Böck.
4679 * Certificate verify crash with missing PSS parameter
4681 The signature verification routines will crash with a NULL pointer
4682 dereference if presented with an ASN.1 signature using the RSA PSS
4683 algorithm and absent mask generation function parameter. Since these
4684 routines are used to verify certificate signature algorithms this can be
4685 used to crash any certificate verification operation and exploited in a
4686 DoS attack. Any application which performs certificate verification is
4687 vulnerable including OpenSSL clients and servers which enable client
4690 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4695 * X509_ATTRIBUTE memory leak
4697 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4698 memory. This structure is used by the PKCS#7 and CMS routines so any
4699 application which reads PKCS#7 or CMS data from untrusted sources is
4700 affected. SSL/TLS is not affected.
4702 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4708 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4709 This changes the decoding behaviour for some invalid messages,
4710 though the change is mostly in the more lenient direction, and
4711 legacy behaviour is preserved as much as possible.
4715 * In DSA_generate_parameters_ex, if the provided seed is too short,
4718 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4720 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4722 * Alternate chains certificate forgery
4724 During certificate verification, OpenSSL will attempt to find an
4725 alternative certificate chain if the first attempt to build such a chain
4726 fails. An error in the implementation of this logic can mean that an
4727 attacker could cause certain checks on untrusted certificates to be
4728 bypassed, such as the CA flag, enabling them to use a valid leaf
4729 certificate to act as a CA and "issue" an invalid certificate.
4731 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4736 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4738 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4739 incompatibility in the handling of HMAC. The previous ABI has now been
4744 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4746 * Malformed ECParameters causes infinite loop
4748 When processing an ECParameters structure OpenSSL enters an infinite loop
4749 if the curve specified is over a specially malformed binary polynomial
4752 This can be used to perform denial of service against any
4753 system which processes public keys, certificate requests or
4754 certificates. This includes TLS clients and TLS servers with
4755 client authentication enabled.
4757 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4762 * Exploitable out-of-bounds read in X509_cmp_time
4764 X509_cmp_time does not properly check the length of the ASN1_TIME
4765 string and can read a few bytes out of bounds. In addition,
4766 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4769 An attacker can use this to craft malformed certificates and CRLs of
4770 various sizes and potentially cause a segmentation fault, resulting in
4771 a DoS on applications that verify certificates or CRLs. TLS clients
4772 that verify CRLs are affected. TLS clients and servers with client
4773 authentication enabled may be affected if they use custom verification
4776 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4777 independently by Hanno Böck.
4782 * PKCS7 crash with missing EnvelopedContent
4784 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4785 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4786 with missing content and trigger a NULL pointer dereference on parsing.
4788 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4789 structures from untrusted sources are affected. OpenSSL clients and
4790 servers are not affected.
4792 This issue was reported to OpenSSL by Michal Zalewski (Google).
4797 * CMS verify infinite loop with unknown hash function
4799 When verifying a signedData message the CMS code can enter an infinite loop
4800 if presented with an unknown hash function OID. This can be used to perform
4801 denial of service against any system which verifies signedData messages using
4803 This issue was reported to OpenSSL by Johannes Bauer.
4808 * Race condition handling NewSessionTicket
4810 If a NewSessionTicket is received by a multi-threaded client when attempting to
4811 reuse a previous ticket then a race condition can occur potentially leading to
4812 a double free of the ticket data.
4817 * Only support 256-bit or stronger elliptic curves with the
4818 'ecdh_auto' setting (server) or by default (client). Of supported
4819 curves, prefer P-256 (both).
4823 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4825 * ClientHello sigalgs DoS fix
4827 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4828 invalid signature algorithms extension a NULL pointer dereference will
4829 occur. This can be exploited in a DoS attack against the server.
4831 This issue was was reported to OpenSSL by David Ramos of Stanford
4835 *Stephen Henson and Matt Caswell*
4837 * Multiblock corrupted pointer fix
4839 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4840 feature only applies on 64 bit x86 architecture platforms that support AES
4841 NI instructions. A defect in the implementation of "multiblock" can cause
4842 OpenSSL's internal write buffer to become incorrectly set to NULL when
4843 using non-blocking IO. Typically, when the user application is using a
4844 socket BIO for writing, this will only result in a failed connection.
4845 However if some other BIO is used then it is likely that a segmentation
4846 fault will be triggered, thus enabling a potential DoS attack.
4848 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4853 * Segmentation fault in DTLSv1_listen fix
4855 The DTLSv1_listen function is intended to be stateless and processes the
4856 initial ClientHello from many peers. It is common for user code to loop
4857 over the call to DTLSv1_listen until a valid ClientHello is received with
4858 an associated cookie. A defect in the implementation of DTLSv1_listen means
4859 that state is preserved in the SSL object from one invocation to the next
4860 that can lead to a segmentation fault. Errors processing the initial
4861 ClientHello can trigger this scenario. An example of such an error could be
4862 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4865 This issue was reported to OpenSSL by Per Allansson.
4870 * Segmentation fault in ASN1_TYPE_cmp fix
4872 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4873 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4874 certificate signature algorithm consistency this can be used to crash any
4875 certificate verification operation and exploited in a DoS attack. Any
4876 application which performs certificate verification is vulnerable including
4877 OpenSSL clients and servers which enable client authentication.
4882 * Segmentation fault for invalid PSS parameters fix
4884 The signature verification routines will crash with a NULL pointer
4885 dereference if presented with an ASN.1 signature using the RSA PSS
4886 algorithm and invalid parameters. Since these routines are used to verify
4887 certificate signature algorithms this can be used to crash any
4888 certificate verification operation and exploited in a DoS attack. Any
4889 application which performs certificate verification is vulnerable including
4890 OpenSSL clients and servers which enable client authentication.
4892 This issue was was reported to OpenSSL by Brian Carpenter.
4897 * ASN.1 structure reuse memory corruption fix
4899 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4900 memory corruption via an invalid write. Such reuse is and has been
4901 strongly discouraged and is believed to be rare.
4903 Applications that parse structures containing CHOICE or ANY DEFINED BY
4904 components may be affected. Certificate parsing (d2i_X509 and related
4905 functions) are however not affected. OpenSSL clients and servers are
4911 * PKCS7 NULL pointer dereferences fix
4913 The PKCS#7 parsing code does not handle missing outer ContentInfo
4914 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4915 missing content and trigger a NULL pointer dereference on parsing.
4917 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4918 otherwise parse PKCS#7 structures from untrusted sources are
4919 affected. OpenSSL clients and servers are not affected.
4921 This issue was reported to OpenSSL by Michal Zalewski (Google).
4926 * DoS via reachable assert in SSLv2 servers fix
4928 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4929 servers that both support SSLv2 and enable export cipher suites by sending
4930 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4932 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4933 (OpenSSL development team).
4938 * Empty CKE with client auth and DHE fix
4940 If client auth is used then a server can seg fault in the event of a DHE
4941 ciphersuite being selected and a zero length ClientKeyExchange message
4942 being sent by the client. This could be exploited in a DoS attack.
4947 * Handshake with unseeded PRNG fix
4949 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4950 with an unseeded PRNG. The conditions are:
4951 - The client is on a platform where the PRNG has not been seeded
4952 automatically, and the user has not seeded manually
4953 - A protocol specific client method version has been used (i.e. not
4954 SSL_client_methodv23)
4955 - A ciphersuite is used that does not require additional random data from
4956 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4958 If the handshake succeeds then the client random that has been used will
4959 have been generated from a PRNG with insufficient entropy and therefore the
4960 output may be predictable.
4962 For example using the following command with an unseeded openssl will
4963 succeed on an unpatched platform:
4965 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4970 * Use After Free following d2i_ECPrivatekey error fix
4972 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4973 could cause a use after free condition. This, in turn, could cause a double
4974 free in several private key parsing functions (such as d2i_PrivateKey
4975 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4976 for applications that receive EC private keys from untrusted
4977 sources. This scenario is considered rare.
4979 This issue was discovered by the BoringSSL project and fixed in their
4985 * X509_to_X509_REQ NULL pointer deref fix
4987 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4988 the certificate key is invalid. This function is rarely used in practice.
4990 This issue was discovered by Brian Carpenter.
4995 * Removed the export ciphers from the DEFAULT ciphers
4999 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5001 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5002 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5003 So far those who have to target multiple platforms would compromise
5004 and argue that binary targeting say ARMv5 would still execute on
5005 ARMv8. "Universal" build resolves this compromise by providing
5006 near-optimal performance even on newer platforms.
5010 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5011 (other platforms pending).
5013 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5015 * Add support for the SignedCertificateTimestampList certificate and
5016 OCSP response extensions from RFC6962.
5020 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5021 for corner cases. (Certain input points at infinity could lead to
5022 bogus results, with non-infinity inputs mapped to infinity too.)
5026 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5027 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5028 common cases are optimized and there still is room for further
5029 improvements. Vector Permutation AES for Altivec is also added.
5033 * Add support for little-endian ppc64 Linux target.
5035 *Marcelo Cerri (IBM)*
5037 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5038 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5039 are optimized and there still is room for further improvements.
5040 Both 32- and 64-bit modes are supported.
5042 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5044 * Improved ARMv7 NEON support.
5048 * Support for SPARC Architecture 2011 crypto extensions, first
5049 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5050 SHA256/512, MD5, GHASH and modular exponentiation.
5052 *Andy Polyakov, David Miller*
5054 * Accelerated modular exponentiation for Intel processors, a.k.a.
5057 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5059 * Support for new and upcoming Intel processors, including AVX2,
5060 BMI and SHA ISA extensions. This includes additional "stitched"
5061 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5064 This work was sponsored by Intel Corp.
5068 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5069 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5070 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5074 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5075 this fixes a limitation in previous versions of OpenSSL.
5079 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5080 MGF1 digest and OAEP label.
5084 * Add EVP support for key wrapping algorithms, to avoid problems with
5085 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5086 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5087 algorithms and include tests cases.
5091 * Add functions to allocate and set the fields of an ECDSA_METHOD
5094 *Douglas E. Engert, Steve Henson*
5096 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5097 difference in days and seconds between two tm or ASN1_TIME structures.
5101 * Add -rev test option to s_server to just reverse order of characters
5102 received by client and send back to server. Also prints an abbreviated
5103 summary of the connection parameters.
5107 * New option -brief for s_client and s_server to print out a brief summary
5108 of connection parameters.
5112 * Add callbacks for arbitrary TLS extensions.
5114 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5116 * New option -crl_download in several openssl utilities to download CRLs
5117 from CRLDP extension in certificates.
5121 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5125 * New function X509_CRL_diff to generate a delta CRL from the difference
5126 of two full CRLs. Add support to "crl" utility.
5130 * New functions to set lookup_crls function and to retrieve
5131 X509_STORE from X509_STORE_CTX.
5135 * Print out deprecated issuer and subject unique ID fields in
5140 * Extend OCSP I/O functions so they can be used for simple general purpose
5141 HTTP as well as OCSP. New wrapper function which can be used to download
5142 CRLs using the OCSP API.
5146 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5150 * `SSL_CONF*` functions. These provide a common framework for application
5151 configuration using configuration files or command lines.
5155 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5156 message callback and prints the results. Needs compile time option
5157 "enable-ssl-trace". New options to s_client and s_server to enable
5162 * New ctrl and macro to retrieve supported points extensions.
5163 Print out extension in s_server and s_client.
5167 * New functions to retrieve certificate signature and signature
5172 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5177 * New Suite B modes for TLS code. These use and enforce the requirements
5178 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5179 only use Suite B curves. The Suite B modes can be set by using the
5180 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5184 * New chain verification flags for Suite B levels of security. Check
5185 algorithms are acceptable when flags are set in X509_verify_cert.
5189 * Make tls1_check_chain return a set of flags indicating checks passed
5190 by a certificate chain. Add additional tests to handle client
5191 certificates: checks for matching certificate type and issuer name
5196 * If an attempt is made to use a signature algorithm not in the peer
5197 preference list abort the handshake. If client has no suitable
5198 signature algorithms in response to a certificate request do not
5199 use the certificate.
5203 * If server EC tmp key is not in client preference list abort handshake.
5207 * Add support for certificate stores in CERT structure. This makes it
5208 possible to have different stores per SSL structure or one store in
5209 the parent SSL_CTX. Include distinct stores for certificate chain
5210 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5211 to build and store a certificate chain in CERT structure: returning
5212 an error if the chain cannot be built: this will allow applications
5213 to test if a chain is correctly configured.
5215 Note: if the CERT based stores are not set then the parent SSL_CTX
5216 store is used to retain compatibility with existing behaviour.
5220 * New function ssl_set_client_disabled to set a ciphersuite disabled
5221 mask based on the current session, check mask when sending client
5222 hello and checking the requested ciphersuite.
5226 * New ctrls to retrieve and set certificate types in a certificate
5227 request message. Print out received values in s_client. If certificate
5228 types is not set with custom values set sensible values based on
5229 supported signature algorithms.
5233 * Support for distinct client and server supported signature algorithms.
5237 * Add certificate callback. If set this is called whenever a certificate
5238 is required by client or server. An application can decide which
5239 certificate chain to present based on arbitrary criteria: for example
5240 supported signature algorithms. Add very simple example to s_server.
5241 This fixes many of the problems and restrictions of the existing client
5242 certificate callback: for example you can now clear an existing
5243 certificate and specify the whole chain.
5247 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5248 the certificate can be used for (if anything). Set valid_flags field
5249 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5250 to have similar checks in it.
5252 Add new "cert_flags" field to CERT structure and include a "strict mode".
5253 This enforces some TLS certificate requirements (such as only permitting
5254 certificate signature algorithms contained in the supported algorithms
5255 extension) which some implementations ignore: this option should be used
5256 with caution as it could cause interoperability issues.
5260 * Update and tidy signature algorithm extension processing. Work out
5261 shared signature algorithms based on preferences and peer algorithms
5262 and print them out in s_client and s_server. Abort handshake if no
5263 shared signature algorithms.
5267 * Add new functions to allow customised supported signature algorithms
5268 for SSL and SSL_CTX structures. Add options to s_client and s_server
5273 * New function SSL_certs_clear() to delete all references to certificates
5274 from an SSL structure. Before this once a certificate had been added
5275 it couldn't be removed.
5279 * Integrate hostname, email address and IP address checking with certificate
5280 verification. New verify options supporting checking in openssl utility.
5284 * Fixes and wildcard matching support to hostname and email checking
5285 functions. Add manual page.
5287 *Florian Weimer (Red Hat Product Security Team)*
5289 * New functions to check a hostname email or IP address against a
5290 certificate. Add options x509 utility to print results of checks against
5295 * Fix OCSP checking.
5297 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5299 * Initial experimental support for explicitly trusted non-root CAs.
5300 OpenSSL still tries to build a complete chain to a root but if an
5301 intermediate CA has a trust setting included that is used. The first
5302 setting is used: whether to trust (e.g., -addtrust option to the x509
5307 * Add -trusted_first option which attempts to find certificates in the
5308 trusted store even if an untrusted chain is also supplied.
5312 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5313 platform support for Linux and Android.
5317 * Support for linux-x32, ILP32 environment in x86_64 framework.
5321 * Experimental multi-implementation support for FIPS capable OpenSSL.
5322 When in FIPS mode the approved implementations are used as normal,
5323 when not in FIPS mode the internal unapproved versions are used instead.
5324 This means that the FIPS capable OpenSSL isn't forced to use the
5325 (often lower performance) FIPS implementations outside FIPS mode.
5329 * Transparently support X9.42 DH parameters when calling
5330 PEM_read_bio_DHparameters. This means existing applications can handle
5331 the new parameter format automatically.
5335 * Initial experimental support for X9.42 DH parameter format: mainly
5336 to support use of 'q' parameter for RFC5114 parameters.
5340 * Add DH parameters from RFC5114 including test data to dhtest.
5344 * Support for automatic EC temporary key parameter selection. If enabled
5345 the most preferred EC parameters are automatically used instead of
5346 hardcoded fixed parameters. Now a server just has to call:
5347 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5348 support ECDH and use the most appropriate parameters.
5352 * Enhance and tidy EC curve and point format TLS extension code. Use
5353 static structures instead of allocation if default values are used.
5354 New ctrls to set curves we wish to support and to retrieve shared curves.
5355 Print out shared curves in s_server. New options to s_server and s_client
5356 to set list of supported curves.
5360 * New ctrls to retrieve supported signature algorithms and
5361 supported curve values as an array of NIDs. Extend openssl utility
5362 to print out received values.
5366 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5367 between NIDs and the more common NIST names such as "P-256". Enhance
5368 ecparam utility and ECC method to recognise the NIST names for curves.
5372 * Enhance SSL/TLS certificate chain handling to support different
5373 chains for each certificate instead of one chain in the parent SSL_CTX.
5377 * Support for fixed DH ciphersuite client authentication: where both
5378 server and client use DH certificates with common parameters.
5382 * Support for fixed DH ciphersuites: those requiring DH server
5387 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5389 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5390 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5391 X509_CINF_get_signature were reverted post internal team review.
5396 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5398 * OCSP Status Request extension unbounded memory growth
5400 A malicious client can send an excessively large OCSP Status Request
5401 extension. If that client continually requests renegotiation, sending a
5402 large OCSP Status Request extension each time, then there will be unbounded
5403 memory growth on the server. This will eventually lead to a Denial Of
5404 Service attack through memory exhaustion. Servers with a default
5405 configuration are vulnerable even if they do not support OCSP. Builds using
5406 the "no-ocsp" build time option are not affected.
5408 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5413 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5416 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5422 * OOB write in MDC2_Update()
5424 An overflow can occur in MDC2_Update() either if called directly or
5425 through the EVP_DigestUpdate() function using MDC2. If an attacker
5426 is able to supply very large amounts of input data after a previous
5427 call to EVP_EncryptUpdate() with a partial block then a length check
5428 can overflow resulting in a heap corruption.
5430 The amount of data needed is comparable to SIZE_MAX which is impractical
5433 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5438 * Malformed SHA512 ticket DoS
5440 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5441 DoS attack where a malformed ticket will result in an OOB read which will
5444 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5445 a custom server callback and ticket lookup mechanism.
5447 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5452 * OOB write in BN_bn2dec()
5454 The function BN_bn2dec() does not check the return value of BN_div_word().
5455 This can cause an OOB write if an application uses this function with an
5456 overly large BIGNUM. This could be a problem if an overly large certificate
5457 or CRL is printed out from an untrusted source. TLS is not affected because
5458 record limits will reject an oversized certificate before it is parsed.
5460 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5465 * OOB read in TS_OBJ_print_bio()
5467 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5468 the total length the OID text representation would use and not the amount
5469 of data written. This will result in OOB reads when large OIDs are
5472 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5477 * Pointer arithmetic undefined behaviour
5479 Avoid some undefined pointer arithmetic
5481 A common idiom in the codebase is to check limits in the following manner:
5484 Where "p" points to some malloc'd data of SIZE bytes and
5487 "len" here could be from some externally supplied data (e.g. from a TLS
5490 The rules of C pointer arithmetic are such that "p + len" is only well
5491 defined where len <= SIZE. Therefore the above idiom is actually
5492 undefined behaviour.
5494 For example this could cause problems if some malloc implementation
5495 provides an address for "p" such that "p + len" actually overflows for
5496 values of len that are too big and therefore p + len < limit.
5498 This issue was reported to OpenSSL by Guido Vranken
5503 * Constant time flag not preserved in DSA signing
5505 Operations in the DSA signing algorithm should run in constant time in
5506 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5507 implementation means that a non-constant time codepath is followed for
5508 certain operations. This has been demonstrated through a cache-timing
5509 attack to be sufficient for an attacker to recover the private DSA key.
5511 This issue was reported by César Pereida (Aalto University), Billy Brumley
5512 (Tampere University of Technology), and Yuval Yarom (The University of
5513 Adelaide and NICTA).
5518 * DTLS buffered message DoS
5520 In a DTLS connection where handshake messages are delivered out-of-order
5521 those messages that OpenSSL is not yet ready to process will be buffered
5522 for later use. Under certain circumstances, a flaw in the logic means that
5523 those messages do not get removed from the buffer even though the handshake
5524 has been completed. An attacker could force up to approx. 15 messages to
5525 remain in the buffer when they are no longer required. These messages will
5526 be cleared when the DTLS connection is closed. The default maximum size for
5527 a message is 100k. Therefore the attacker could force an additional 1500k
5528 to be consumed per connection. By opening many simulataneous connections an
5529 attacker could cause a DoS attack through memory exhaustion.
5531 This issue was reported to OpenSSL by Quan Luo.
5536 * DTLS replay protection DoS
5538 A flaw in the DTLS replay attack protection mechanism means that records
5539 that arrive for future epochs update the replay protection "window" before
5540 the MAC for the record has been validated. This could be exploited by an
5541 attacker by sending a record for the next epoch (which does not have to
5542 decrypt or have a valid MAC), with a very large sequence number. This means
5543 that all subsequent legitimate packets are dropped causing a denial of
5544 service for a specific DTLS connection.
5546 This issue was reported to OpenSSL by the OCAP audit team.
5551 * Certificate message OOB reads
5553 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5554 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5555 theoretical DoS risk but this has not been observed in practice on common
5558 The messages affected are client certificate, client certificate request
5559 and server certificate. As a result the attack can only be performed
5560 against a client or a server which enables client authentication.
5562 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5567 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5569 * Prevent padding oracle in AES-NI CBC MAC check
5571 A MITM attacker can use a padding oracle attack to decrypt traffic
5572 when the connection uses an AES CBC cipher and the server support
5575 This issue was introduced as part of the fix for Lucky 13 padding
5576 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5577 constant time by making sure that always the same bytes are read and
5578 compared against either the MAC or padding bytes. But it no longer
5579 checked that there was enough data to have both the MAC and padding
5582 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5587 * Fix EVP_EncodeUpdate overflow
5589 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5590 Base64 encoding of binary data. If an attacker is able to supply very large
5591 amounts of input data then a length check can overflow resulting in a heap
5594 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5595 the `PEM_write_bio*` family of functions. These are mainly used within the
5596 OpenSSL command line applications, so any application which processes data
5597 from an untrusted source and outputs it as a PEM file should be considered
5598 vulnerable to this issue. User applications that call these APIs directly
5599 with large amounts of untrusted data may also be vulnerable.
5601 This issue was reported by Guido Vranken.
5606 * Fix EVP_EncryptUpdate overflow
5608 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5609 is able to supply very large amounts of input data after a previous call to
5610 EVP_EncryptUpdate() with a partial block then a length check can overflow
5611 resulting in a heap corruption. Following an analysis of all OpenSSL
5612 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5613 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5614 the first called function after an EVP_EncryptInit(), and therefore that
5615 specific call must be safe. The second form is where the length passed to
5616 EVP_EncryptUpdate() can be seen from the code to be some small value and
5617 therefore there is no possibility of an overflow. Since all instances are
5618 one of these two forms, it is believed that there can be no overflows in
5619 internal code due to this problem. It should be noted that
5620 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5621 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5622 of these calls have also been analysed too and it is believed there are no
5623 instances in internal usage where an overflow could occur.
5625 This issue was reported by Guido Vranken.
5630 * Prevent ASN.1 BIO excessive memory allocation
5632 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5633 a short invalid encoding can casuse allocation of large amounts of memory
5634 potentially consuming excessive resources or exhausting memory.
5636 Any application parsing untrusted data through d2i BIO functions is
5637 affected. The memory based functions such as d2i_X509() are *not* affected.
5638 Since the memory based functions are used by the TLS library, TLS
5639 applications are not affected.
5641 This issue was reported by Brian Carpenter.
5648 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5649 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5650 in arbitrary stack data being returned in the buffer.
5652 This issue was reported by Guido Vranken.
5657 * Modify behavior of ALPN to invoke callback after SNI/servername
5658 callback, such that updates to the SSL_CTX affect ALPN.
5662 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5667 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5668 methods are enabled and ssl2 is disabled the methods return NULL.
5672 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5674 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5675 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5676 provide any "EXPORT" or "LOW" strength ciphers.
5680 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5681 is by default disabled at build-time. Builds that are not configured with
5682 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5683 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5684 will need to explicitly call either of:
5686 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5688 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5690 as appropriate. Even if either of those is used, or the application
5691 explicitly uses the version-specific SSLv2_method() or its client and
5692 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5693 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5694 ciphers, and SSLv2 56-bit DES are no longer available.
5699 * Fix a double-free in DSA code
5701 A double free bug was discovered when OpenSSL parses malformed DSA private
5702 keys and could lead to a DoS attack or memory corruption for applications
5703 that receive DSA private keys from untrusted sources. This scenario is
5706 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5712 * Disable SRP fake user seed to address a server memory leak.
5714 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5716 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5717 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5718 was changed to ignore the "fake user" SRP seed, even if the seed
5721 Users should use SRP_VBASE_get1_by_user instead. Note that in
5722 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5723 also that even though configuring the SRP seed attempts to hide
5724 invalid usernames by continuing the handshake with fake
5725 credentials, this behaviour is not constant time and no strong
5726 guarantees are made that the handshake is indistinguishable from
5727 that of a valid user.
5732 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5734 In the BN_hex2bn function the number of hex digits is calculated using an
5735 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5736 large values of `i` this can result in `bn_expand` not allocating any
5737 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5738 field as NULL leading to a subsequent NULL ptr deref. For very large values
5739 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5740 In this case memory is allocated to the internal BIGNUM data field, but it
5741 is insufficiently sized leading to heap corruption. A similar issue exists
5742 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5743 is ever called by user applications with very large untrusted hex/dec data.
5744 This is anticipated to be a rare occurrence.
5746 All OpenSSL internal usage of these functions use data that is not expected
5747 to be untrusted, e.g. config file data or application command line
5748 arguments. If user developed applications generate config file data based
5749 on untrusted data then it is possible that this could also lead to security
5750 consequences. This is also anticipated to be rare.
5752 This issue was reported to OpenSSL by Guido Vranken.
5757 * Fix memory issues in `BIO_*printf` functions
5759 The internal `fmtstr` function used in processing a "%s" format string in
5760 the `BIO_*printf` functions could overflow while calculating the length of a
5761 string and cause an OOB read when printing very long strings.
5763 Additionally the internal `doapr_outch` function can attempt to write to an
5764 OOB memory location (at an offset from the NULL pointer) in the event of a
5765 memory allocation failure. In 1.0.2 and below this could be caused where
5766 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5767 could be in processing a very long "%s" format string. Memory leaks can
5770 The first issue may mask the second issue dependent on compiler behaviour.
5771 These problems could enable attacks where large amounts of untrusted data
5772 is passed to the `BIO_*printf` functions. If applications use these functions
5773 in this way then they could be vulnerable. OpenSSL itself uses these
5774 functions when printing out human-readable dumps of ASN.1 data. Therefore
5775 applications that print this data could be vulnerable if the data is from
5776 untrusted sources. OpenSSL command line applications could also be
5777 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5778 as command line arguments.
5780 Libssl is not considered directly vulnerable. Additionally certificates etc
5781 received via remote connections via libssl are also unlikely to be able to
5782 trigger these issues because of message size limits enforced within libssl.
5784 This issue was reported to OpenSSL Guido Vranken.
5789 * Side channel attack on modular exponentiation
5791 A side-channel attack was found which makes use of cache-bank conflicts on
5792 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5793 of RSA keys. The ability to exploit this issue is limited as it relies on
5794 an attacker who has control of code in a thread running on the same
5795 hyper-threaded core as the victim thread which is performing decryptions.
5797 This issue was reported to OpenSSL by Yuval Yarom, The University of
5798 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5799 Nadia Heninger, University of Pennsylvania with more information at
5800 <http://cachebleed.info>.
5805 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5806 if no keysize is specified with default_bits. This fixes an
5807 omission in an earlier change that changed all RSA/DSA key generation
5808 apps to use 2048 bits by default.
5812 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5814 * Protection for DH small subgroup attacks
5816 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5817 switched on by default and cannot be disabled. This could have some
5822 * SSLv2 doesn't block disabled ciphers
5824 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5825 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5826 been disabled, provided that the SSLv2 protocol was not also disabled via
5829 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5830 and Sebastian Schinzel.
5835 * Reject DH handshakes with parameters shorter than 1024 bits.
5839 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5841 * Certificate verify crash with missing PSS parameter
5843 The signature verification routines will crash with a NULL pointer
5844 dereference if presented with an ASN.1 signature using the RSA PSS
5845 algorithm and absent mask generation function parameter. Since these
5846 routines are used to verify certificate signature algorithms this can be
5847 used to crash any certificate verification operation and exploited in a
5848 DoS attack. Any application which performs certificate verification is
5849 vulnerable including OpenSSL clients and servers which enable client
5852 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5857 * X509_ATTRIBUTE memory leak
5859 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5860 memory. This structure is used by the PKCS#7 and CMS routines so any
5861 application which reads PKCS#7 or CMS data from untrusted sources is
5862 affected. SSL/TLS is not affected.
5864 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5870 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5871 This changes the decoding behaviour for some invalid messages,
5872 though the change is mostly in the more lenient direction, and
5873 legacy behaviour is preserved as much as possible.
5877 * In DSA_generate_parameters_ex, if the provided seed is too short,
5878 use a random seed, as already documented.
5880 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5882 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5884 * Alternate chains certificate forgery
5886 During certificate verfification, OpenSSL will attempt to find an
5887 alternative certificate chain if the first attempt to build such a chain
5888 fails. An error in the implementation of this logic can mean that an
5889 attacker could cause certain checks on untrusted certificates to be
5890 bypassed, such as the CA flag, enabling them to use a valid leaf
5891 certificate to act as a CA and "issue" an invalid certificate.
5893 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5899 * Race condition handling PSK identify hint
5901 If PSK identity hints are received by a multi-threaded client then
5902 the values are wrongly updated in the parent SSL_CTX structure. This can
5903 result in a race condition potentially leading to a double free of the
5909 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5911 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5912 incompatibility in the handling of HMAC. The previous ABI has now been
5915 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5917 * Malformed ECParameters causes infinite loop
5919 When processing an ECParameters structure OpenSSL enters an infinite loop
5920 if the curve specified is over a specially malformed binary polynomial
5923 This can be used to perform denial of service against any
5924 system which processes public keys, certificate requests or
5925 certificates. This includes TLS clients and TLS servers with
5926 client authentication enabled.
5928 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5933 * Exploitable out-of-bounds read in X509_cmp_time
5935 X509_cmp_time does not properly check the length of the ASN1_TIME
5936 string and can read a few bytes out of bounds. In addition,
5937 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5940 An attacker can use this to craft malformed certificates and CRLs of
5941 various sizes and potentially cause a segmentation fault, resulting in
5942 a DoS on applications that verify certificates or CRLs. TLS clients
5943 that verify CRLs are affected. TLS clients and servers with client
5944 authentication enabled may be affected if they use custom verification
5947 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5948 independently by Hanno Böck.
5953 * PKCS7 crash with missing EnvelopedContent
5955 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5956 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5957 with missing content and trigger a NULL pointer dereference on parsing.
5959 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5960 structures from untrusted sources are affected. OpenSSL clients and
5961 servers are not affected.
5963 This issue was reported to OpenSSL by Michal Zalewski (Google).
5968 * CMS verify infinite loop with unknown hash function
5970 When verifying a signedData message the CMS code can enter an infinite loop
5971 if presented with an unknown hash function OID. This can be used to perform
5972 denial of service against any system which verifies signedData messages using
5974 This issue was reported to OpenSSL by Johannes Bauer.
5979 * Race condition handling NewSessionTicket
5981 If a NewSessionTicket is received by a multi-threaded client when attempting to
5982 reuse a previous ticket then a race condition can occur potentially leading to
5983 a double free of the ticket data.
5988 * Reject DH handshakes with parameters shorter than 768 bits.
5990 *Kurt Roeckx and Emilia Kasper*
5992 * dhparam: generate 2048-bit parameters by default.
5994 *Kurt Roeckx and Emilia Kasper*
5996 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5998 * Segmentation fault in ASN1_TYPE_cmp fix
6000 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6001 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6002 certificate signature algorithm consistency this can be used to crash any
6003 certificate verification operation and exploited in a DoS attack. Any
6004 application which performs certificate verification is vulnerable including
6005 OpenSSL clients and servers which enable client authentication.
6010 * ASN.1 structure reuse memory corruption fix
6012 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6013 memory corruption via an invalid write. Such reuse is and has been
6014 strongly discouraged and is believed to be rare.
6016 Applications that parse structures containing CHOICE or ANY DEFINED BY
6017 components may be affected. Certificate parsing (d2i_X509 and related
6018 functions) are however not affected. OpenSSL clients and servers are
6024 * PKCS7 NULL pointer dereferences fix
6026 The PKCS#7 parsing code does not handle missing outer ContentInfo
6027 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6028 missing content and trigger a NULL pointer dereference on parsing.
6030 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6031 otherwise parse PKCS#7 structures from untrusted sources are
6032 affected. OpenSSL clients and servers are not affected.
6034 This issue was reported to OpenSSL by Michal Zalewski (Google).
6039 * DoS via reachable assert in SSLv2 servers fix
6041 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6042 servers that both support SSLv2 and enable export cipher suites by sending
6043 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6045 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6046 (OpenSSL development team).
6051 * Use After Free following d2i_ECPrivatekey error fix
6053 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6054 could cause a use after free condition. This, in turn, could cause a double
6055 free in several private key parsing functions (such as d2i_PrivateKey
6056 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6057 for applications that receive EC private keys from untrusted
6058 sources. This scenario is considered rare.
6060 This issue was discovered by the BoringSSL project and fixed in their
6066 * X509_to_X509_REQ NULL pointer deref fix
6068 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6069 the certificate key is invalid. This function is rarely used in practice.
6071 This issue was discovered by Brian Carpenter.
6076 * Removed the export ciphers from the DEFAULT ciphers
6080 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6082 * Build fixes for the Windows and OpenVMS platforms
6084 *Matt Caswell and Richard Levitte*
6086 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6088 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6089 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6090 dereference. This could lead to a Denial Of Service attack. Thanks to
6091 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6096 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6097 dtls1_buffer_record function under certain conditions. In particular this
6098 could occur if an attacker sent repeated DTLS records with the same
6099 sequence number but for the next epoch. The memory leak could be exploited
6100 by an attacker in a Denial of Service attack through memory exhaustion.
6101 Thanks to Chris Mueller for reporting this issue.
6106 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6107 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6108 method would be set to NULL which could later result in a NULL pointer
6109 dereference. Thanks to Frank Schmirler for reporting this issue.
6114 * Abort handshake if server key exchange message is omitted for ephemeral
6117 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6118 reporting this issue.
6123 * Remove non-export ephemeral RSA code on client and server. This code
6124 violated the TLS standard by allowing the use of temporary RSA keys in
6125 non-export ciphersuites and could be used by a server to effectively
6126 downgrade the RSA key length used to a value smaller than the server
6127 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6128 INRIA or reporting this issue.
6133 * Fixed issue where DH client certificates are accepted without verification.
6134 An OpenSSL server will accept a DH certificate for client authentication
6135 without the certificate verify message. This effectively allows a client to
6136 authenticate without the use of a private key. This only affects servers
6137 which trust a client certificate authority which issues certificates
6138 containing DH keys: these are extremely rare and hardly ever encountered.
6139 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6145 * Ensure that the session ID context of an SSL is updated when its
6146 SSL_CTX is updated via SSL_set_SSL_CTX.
6148 The session ID context is typically set from the parent SSL_CTX,
6149 and can vary with the CTX.
6153 * Fix various certificate fingerprint issues.
6155 By using non-DER or invalid encodings outside the signed portion of a
6156 certificate the fingerprint can be changed without breaking the signature.
6157 Although no details of the signed portion of the certificate can be changed
6158 this can cause problems with some applications: e.g. those using the
6159 certificate fingerprint for blacklists.
6161 1. Reject signatures with non zero unused bits.
6163 If the BIT STRING containing the signature has non zero unused bits reject
6164 the signature. All current signature algorithms require zero unused bits.
6166 2. Check certificate algorithm consistency.
6168 Check the AlgorithmIdentifier inside TBS matches the one in the
6169 certificate signature. NB: this will result in signature failure
6170 errors for some broken certificates.
6172 Thanks to Konrad Kraszewski from Google for reporting this issue.
6174 3. Check DSA/ECDSA signatures use DER.
6176 Re-encode DSA/ECDSA signatures and compare with the original received
6177 signature. Return an error if there is a mismatch.
6179 This will reject various cases including garbage after signature
6180 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6181 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6182 (negative or with leading zeroes).
6184 Further analysis was conducted and fixes were developed by Stephen Henson
6185 of the OpenSSL core team.
6191 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6192 results on some platforms, including x86_64. This bug occurs at random
6193 with a very low probability, and is not known to be exploitable in any
6194 way, though its exact impact is difficult to determine. Thanks to Pieter
6195 Wuille (Blockstream) who reported this issue and also suggested an initial
6196 fix. Further analysis was conducted by the OpenSSL development team and
6197 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6198 the OpenSSL core team.
6203 * Do not resume sessions on the server if the negotiated protocol
6204 version does not match the session's version. Resuming with a different
6205 version, while not strictly forbidden by the RFC, is of questionable
6206 sanity and breaks all known clients.
6208 *David Benjamin, Emilia Käsper*
6210 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6211 early CCS messages during renegotiation. (Note that because
6212 renegotiation is encrypted, this early CCS was not exploitable.)
6216 * Tighten client-side session ticket handling during renegotiation:
6217 ensure that the client only accepts a session ticket if the server sends
6218 the extension anew in the ServerHello. Previously, a TLS client would
6219 reuse the old extension state and thus accept a session ticket if one was
6220 announced in the initial ServerHello.
6222 Similarly, ensure that the client requires a session ticket if one
6223 was advertised in the ServerHello. Previously, a TLS client would
6224 ignore a missing NewSessionTicket message.
6228 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6232 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6233 sends a carefully crafted handshake message, to cause OpenSSL to fail
6234 to free up to 64k of memory causing a memory leak. This could be
6235 exploited in a Denial Of Service attack. This issue affects OpenSSL
6236 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6237 whether SRTP is used or configured. Implementations of OpenSSL that
6238 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6240 The fix was developed by the OpenSSL team.
6245 * Session Ticket Memory Leak.
6247 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6248 integrity of that ticket is first verified. In the event of a session
6249 ticket integrity check failing, OpenSSL will fail to free memory
6250 causing a memory leak. By sending a large number of invalid session
6251 tickets an attacker could exploit this issue in a Denial Of Service
6257 * Build option no-ssl3 is incomplete.
6259 When OpenSSL is configured with "no-ssl3" as a build option, servers
6260 could accept and complete a SSL 3.0 handshake, and clients could be
6261 configured to send them.
6264 *Akamai and the OpenSSL team*
6266 * Add support for TLS_FALLBACK_SCSV.
6267 Client applications doing fallback retries should call
6268 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6271 *Adam Langley, Bodo Moeller*
6273 * Add additional DigestInfo checks.
6275 Re-encode DigestInto in DER and check against the original when
6276 verifying RSA signature: this will reject any improperly encoded
6277 DigestInfo structures.
6279 Note: this is a precautionary measure and no attacks are currently known.
6283 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6285 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6286 SRP code can be overrun an internal buffer. Add sanity check that
6287 g, A, B < N to SRP code.
6289 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6290 Group for discovering this issue.
6295 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6296 TLS 1.0 instead of higher protocol versions when the ClientHello message
6297 is badly fragmented. This allows a man-in-the-middle attacker to force a
6298 downgrade to TLS 1.0 even if both the server and the client support a
6299 higher protocol version, by modifying the client's TLS records.
6301 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6302 researching this issue.
6307 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6308 to a denial of service attack. A malicious server can crash the client
6309 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6310 ciphersuite and sending carefully crafted handshake messages.
6312 Thanks to Felix Gröbert (Google) for discovering and researching this
6318 * By sending carefully crafted DTLS packets an attacker could cause openssl
6319 to leak memory. This can be exploited through a Denial of Service attack.
6320 Thanks to Adam Langley for discovering and researching this issue.
6325 * An attacker can force openssl to consume large amounts of memory whilst
6326 processing DTLS handshake messages. This can be exploited through a
6327 Denial of Service attack.
6328 Thanks to Adam Langley for discovering and researching this issue.
6333 * An attacker can force an error condition which causes openssl to crash
6334 whilst processing DTLS packets due to memory being freed twice. This
6335 can be exploited through a Denial of Service attack.
6336 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6342 * If a multithreaded client connects to a malicious server using a resumed
6343 session and the server sends an ec point format extension it could write
6344 up to 255 bytes to freed memory.
6346 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6352 * A malicious server can crash an OpenSSL client with a null pointer
6353 dereference (read) by specifying an SRP ciphersuite even though it was not
6354 properly negotiated with the client. This can be exploited through a
6355 Denial of Service attack.
6357 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6358 discovering and researching this issue.
6363 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6364 X509_name_oneline, X509_name_print_ex et al. to leak some information
6365 from the stack. Applications may be affected if they echo pretty printing
6366 output to the attacker.
6368 Thanks to Ivan Fratric (Google) for discovering this issue.
6371 *Emilia Käsper, and Steve Henson*
6373 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6374 for corner cases. (Certain input points at infinity could lead to
6375 bogus results, with non-infinity inputs mapped to infinity too.)
6379 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6381 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6382 handshake can force the use of weak keying material in OpenSSL
6383 SSL/TLS clients and servers.
6385 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6386 researching this issue. [CVE-2014-0224][]
6388 *KIKUCHI Masashi, Steve Henson*
6390 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6391 OpenSSL DTLS client the code can be made to recurse eventually crashing
6394 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6397 *Imre Rad, Steve Henson*
6399 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6400 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6401 client or server. This is potentially exploitable to run arbitrary
6402 code on a vulnerable client or server.
6404 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6406 *Jüri Aedla, Steve Henson*
6408 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6409 are subject to a denial of service attack.
6411 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6412 this issue. [CVE-2014-3470][]
6414 *Felix Gröbert, Ivan Fratric, Steve Henson*
6416 * Harmonize version and its documentation. -f flag is used to display
6419 *mancha <mancha1@zoho.com>*
6421 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6422 in i2d_ECPrivateKey.
6424 *mancha <mancha1@zoho.com>*
6426 * Fix some double frees. These are not thought to be exploitable.
6428 *mancha <mancha1@zoho.com>*
6430 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6432 * A missing bounds check in the handling of the TLS heartbeat extension
6433 can be used to reveal up to 64k of memory to a connected client or
6436 Thanks for Neel Mehta of Google Security for discovering this bug and to
6437 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6438 preparing the fix [CVE-2014-0160][]
6440 *Adam Langley, Bodo Moeller*
6442 * Fix for the attack described in the paper "Recovering OpenSSL
6443 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6444 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6445 <http://eprint.iacr.org/2014/140>
6447 Thanks to Yuval Yarom and Naomi Benger for discovering this
6448 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6450 *Yuval Yarom and Naomi Benger*
6452 * TLS pad extension: draft-agl-tls-padding-03
6454 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6455 TLS client Hello record length value would otherwise be > 255 and
6456 less that 512 pad with a dummy extension containing zeroes so it
6457 is at least 512 bytes long.
6459 *Adam Langley, Steve Henson*
6461 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6463 * Fix for TLS record tampering bug. A carefully crafted invalid
6464 handshake could crash OpenSSL with a NULL pointer exception.
6465 Thanks to Anton Johansson for reporting this issues.
6468 * Keep original DTLS digest and encryption contexts in retransmission
6469 structures so we can use the previous session parameters if they need
6470 to be resent. [CVE-2013-6450][]
6474 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6475 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6476 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6477 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6478 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6479 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6481 *Rob Stradling, Adam Langley*
6483 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6485 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6486 supporting platforms or when small records were transferred.
6488 *Andy Polyakov, Steve Henson*
6490 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6492 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6494 This addresses the flaw in CBC record processing discovered by
6495 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6496 at: <http://www.isg.rhul.ac.uk/tls/>
6498 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6499 Security Group at Royal Holloway, University of London
6500 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6501 Emilia Käsper for the initial patch.
6504 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6506 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6507 ciphersuites which can be exploited in a denial of service attack.
6508 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6509 and detecting this bug and to Wolfgang Ettlinger
6510 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6515 * Return an error when checking OCSP signatures when key is NULL.
6516 This fixes a DoS attack. [CVE-2013-0166][]
6520 * Make openssl verify return errors.
6522 *Chris Palmer <palmer@google.com> and Ben Laurie*
6524 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6525 the right response is stapled. Also change SSL_get_certificate()
6526 so it returns the certificate actually sent.
6527 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6529 *Rob Stradling <rob.stradling@comodo.com>*
6531 * Fix possible deadlock when decoding public keys.
6535 * Don't use TLS 1.0 record version number in initial client hello
6540 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6542 * Sanity check record length before skipping explicit IV in TLS
6543 1.2, 1.1 and DTLS to fix DoS attack.
6545 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6546 fuzzing as a service testing platform.
6551 * Initialise tkeylen properly when encrypting CMS messages.
6552 Thanks to Solar Designer of Openwall for reporting this issue.
6556 * In FIPS mode don't try to use composite ciphers as they are not
6561 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6563 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6564 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6565 mean any application compiled against OpenSSL 1.0.0 headers setting
6566 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6567 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6568 0x10000000L Any application which was previously compiled against
6569 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6570 will need to be recompiled as a result. Letting be results in
6571 inability to disable specifically TLS 1.1 and in client context,
6572 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6576 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6577 disable just protocol X, but all protocols above X *if* there are
6578 protocols *below* X still enabled. In more practical terms it means
6579 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6580 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6581 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6586 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6588 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6589 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6590 in CRYPTO_realloc_clean.
6592 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6593 issue and to Adam Langley <agl@chromium.org> for fixing it.
6596 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6598 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6602 * Workarounds for some broken servers that "hang" if a client hello
6603 record length exceeds 255 bytes.
6605 1. Do not use record version number > TLS 1.0 in initial client
6606 hello: some (but not all) hanging servers will now work.
6607 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6608 the number of ciphers sent in the client hello. This should be
6609 set to an even number, such as 50, for example by passing:
6610 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6611 Most broken servers should now work.
6612 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6613 TLS 1.2 client support entirely.
6617 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6621 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6623 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6624 STRING form instead of a DigestInfo.
6628 * The format used for MDC2 RSA signatures is inconsistent between EVP
6629 and the RSA_sign/RSA_verify functions. This was made more apparent when
6630 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6631 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6632 the correct format in RSA_verify so both forms transparently work.
6636 * Some servers which support TLS 1.0 can choke if we initially indicate
6637 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6638 encrypted premaster secret. As a workaround use the maximum permitted
6639 client version in client hello, this should keep such servers happy
6640 and still work with previous versions of OpenSSL.
6644 * Add support for TLS/DTLS heartbeats.
6646 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6648 * Add support for SCTP.
6650 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6652 * Improved PRNG seeding for VOS.
6654 *Paul Green <Paul.Green@stratus.com>*
6656 * Extensive assembler packs updates, most notably:
6658 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6659 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6660 - x86_64: bit-sliced AES implementation;
6661 - ARM: NEON support, contemporary platforms optimizations;
6662 - s390x: z196 support;
6663 - `*`: GHASH and GF(2^m) multiplication implementations;
6667 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6668 (removal of unnecessary code)
6670 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6672 * Add TLS key material exporter from RFC 5705.
6676 * Add DTLS-SRTP negotiation from RFC 5764.
6680 * Add Next Protocol Negotiation,
6681 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6682 disabled with a no-npn flag to config or Configure. Code donated
6685 *Adam Langley <agl@google.com> and Ben Laurie*
6687 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6688 NIST-P256, NIST-P521, with constant-time single point multiplication on
6689 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6690 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6691 Code made available under Apache License version 2.0.
6693 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6694 line to include this in your build of OpenSSL, and run "make depend" (or
6695 "make update"). This enables the following EC_METHODs:
6697 EC_GFp_nistp224_method()
6698 EC_GFp_nistp256_method()
6699 EC_GFp_nistp521_method()
6701 EC_GROUP_new_by_curve_name() will automatically use these (while
6702 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6705 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6707 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6708 all platforms. Move ssize_t definition from e_os.h to the public
6709 header file e_os2.h as it now appears in public header file cms.h
6713 * New -sigopt option to the ca, req and x509 utilities. Additional
6714 signature parameters can be passed using this option and in
6719 * Add RSA PSS signing function. This will generate and set the
6720 appropriate AlgorithmIdentifiers for PSS based on those in the
6721 corresponding EVP_MD_CTX structure. No application support yet.
6725 * Support for companion algorithm specific ASN1 signing routines.
6726 New function ASN1_item_sign_ctx() signs a pre-initialised
6727 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6728 the appropriate parameters.
6732 * Add new algorithm specific ASN1 verification initialisation function
6733 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6734 handling will be the same no matter what EVP_PKEY_METHOD is used.
6735 Add a PSS handler to support verification of PSS signatures: checked
6736 against a number of sample certificates.
6740 * Add signature printing for PSS. Add PSS OIDs.
6742 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6744 * Add algorithm specific signature printing. An individual ASN1 method
6745 can now print out signatures instead of the standard hex dump.
6747 More complex signatures (e.g. PSS) can print out more meaningful
6748 information. Include DSA version that prints out the signature
6753 * Password based recipient info support for CMS library: implementing
6758 * Split password based encryption into PBES2 and PBKDF2 functions. This
6759 neatly separates the code into cipher and PBE sections and is required
6760 for some algorithms that split PBES2 into separate pieces (such as
6761 password based CMS).
6765 * Session-handling fixes:
6766 - Fix handling of connections that are resuming with a session ID,
6767 but also support Session Tickets.
6768 - Fix a bug that suppressed issuing of a new ticket if the client
6769 presented a ticket with an expired session.
6770 - Try to set the ticket lifetime hint to something reasonable.
6771 - Make tickets shorter by excluding irrelevant information.
6772 - On the client side, don't ignore renewed tickets.
6774 *Adam Langley, Bodo Moeller (Google)*
6776 * Fix PSK session representation.
6780 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6782 This work was sponsored by Intel.
6786 * Add GCM support to TLS library. Some custom code is needed to split
6787 the IV between the fixed (from PRF) and explicit (from TLS record)
6788 portions. This adds all GCM ciphersuites supported by RFC5288 and
6789 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6790 add a special AESGCM string for GCM only.
6794 * Expand range of ctrls for AES GCM. Permit setting invocation
6795 field on decrypt and retrieval of invocation field only on encrypt.
6799 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6800 As required by RFC5289 these ciphersuites cannot be used if for
6801 versions of TLS earlier than 1.2.
6805 * For FIPS capable OpenSSL interpret a NULL default public key method
6806 as unset and return the appropriate default but do *not* set the default.
6807 This means we can return the appropriate method in applications that
6808 switch between FIPS and non-FIPS modes.
6812 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6813 ENGINE is used then we cannot handle that in the FIPS module so we
6814 keep original code iff non-FIPS operations are allowed.
6818 * Add -attime option to openssl utilities.
6820 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6822 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6826 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6827 FIPS EC methods unconditionally for now.
6831 * New build option no-ec2m to disable characteristic 2 code.
6835 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6836 all cases can be covered as some introduce binary incompatibilities.
6840 * Redirect RSA operations to FIPS module including keygen,
6841 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6845 * Add similar low level API blocking to ciphers.
6849 * Low level digest APIs are not approved in FIPS mode: any attempt
6850 to use these will cause a fatal error. Applications that *really* want
6851 to use them can use the `private_*` version instead.
6855 * Redirect cipher operations to FIPS module for FIPS builds.
6859 * Redirect digest operations to FIPS module for FIPS builds.
6863 * Update build system to add "fips" flag which will link in fipscanister.o
6864 for static and shared library builds embedding a signature if needed.
6868 * Output TLS supported curves in preference order instead of numerical
6869 order. This is currently hardcoded for the highest order curves first.
6870 This should be configurable so applications can judge speed vs strength.
6874 * Add TLS v1.2 server support for client authentication.
6878 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6883 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6884 FIPS modules versions.
6888 * Add TLS v1.2 client side support for client authentication. Keep cache
6889 of handshake records longer as we don't know the hash algorithm to use
6890 until after the certificate request message is received.
6894 * Initial TLS v1.2 client support. Add a default signature algorithms
6895 extension including all the algorithms we support. Parse new signature
6896 format in client key exchange. Relax some ECC signing restrictions for
6897 TLS v1.2 as indicated in RFC5246.
6901 * Add server support for TLS v1.2 signature algorithms extension. Switch
6902 to new signature format when needed using client digest preference.
6903 All server ciphersuites should now work correctly in TLS v1.2. No client
6904 support yet and no support for client certificates.
6908 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6909 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6910 ciphersuites. At present only RSA key exchange ciphersuites work with
6911 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6912 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6913 and version checking.
6917 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6918 with this defined it will not be affected by any changes to ssl internal
6919 structures. Add several utility functions to allow openssl application
6920 to work with OPENSSL_NO_SSL_INTERN defined.
6924 * A long standing patch to add support for SRP from EdelWeb (Peter
6925 Sylvester and Christophe Renou) was integrated.
6926 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6927 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6930 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6934 * Permit abbreviated handshakes when renegotiating using the function
6935 SSL_renegotiate_abbreviated().
6937 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6939 * Add call to ENGINE_register_all_complete() to
6940 ENGINE_load_builtin_engines(), so some implementations get used
6941 automatically instead of needing explicit application support.
6945 * Add support for TLS key exporter as described in RFC5705.
6947 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6949 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6950 a few changes are required:
6952 Add SSL_OP_NO_TLSv1_1 flag.
6953 Add TLSv1_1 methods.
6954 Update version checking logic to handle version 1.1.
6955 Add explicit IV handling (ported from DTLS code).
6956 Add command line options to s_client/s_server.
6963 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6965 * X509_ATTRIBUTE memory leak
6967 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6968 memory. This structure is used by the PKCS#7 and CMS routines so any
6969 application which reads PKCS#7 or CMS data from untrusted sources is
6970 affected. SSL/TLS is not affected.
6972 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6978 * Race condition handling PSK identify hint
6980 If PSK identity hints are received by a multi-threaded client then
6981 the values are wrongly updated in the parent SSL_CTX structure. This can
6982 result in a race condition potentially leading to a double free of the
6988 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6990 * Malformed ECParameters causes infinite loop
6992 When processing an ECParameters structure OpenSSL enters an infinite loop
6993 if the curve specified is over a specially malformed binary polynomial
6996 This can be used to perform denial of service against any
6997 system which processes public keys, certificate requests or
6998 certificates. This includes TLS clients and TLS servers with
6999 client authentication enabled.
7001 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7006 * Exploitable out-of-bounds read in X509_cmp_time
7008 X509_cmp_time does not properly check the length of the ASN1_TIME
7009 string and can read a few bytes out of bounds. In addition,
7010 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7013 An attacker can use this to craft malformed certificates and CRLs of
7014 various sizes and potentially cause a segmentation fault, resulting in
7015 a DoS on applications that verify certificates or CRLs. TLS clients
7016 that verify CRLs are affected. TLS clients and servers with client
7017 authentication enabled may be affected if they use custom verification
7020 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7021 independently by Hanno Böck.
7026 * PKCS7 crash with missing EnvelopedContent
7028 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7029 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7030 with missing content and trigger a NULL pointer dereference on parsing.
7032 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7033 structures from untrusted sources are affected. OpenSSL clients and
7034 servers are not affected.
7036 This issue was reported to OpenSSL by Michal Zalewski (Google).
7041 * CMS verify infinite loop with unknown hash function
7043 When verifying a signedData message the CMS code can enter an infinite loop
7044 if presented with an unknown hash function OID. This can be used to perform
7045 denial of service against any system which verifies signedData messages using
7047 This issue was reported to OpenSSL by Johannes Bauer.
7052 * Race condition handling NewSessionTicket
7054 If a NewSessionTicket is received by a multi-threaded client when attempting to
7055 reuse a previous ticket then a race condition can occur potentially leading to
7056 a double free of the ticket data.
7061 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7063 * Segmentation fault in ASN1_TYPE_cmp fix
7065 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7066 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7067 certificate signature algorithm consistency this can be used to crash any
7068 certificate verification operation and exploited in a DoS attack. Any
7069 application which performs certificate verification is vulnerable including
7070 OpenSSL clients and servers which enable client authentication.
7075 * ASN.1 structure reuse memory corruption fix
7077 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7078 memory corruption via an invalid write. Such reuse is and has been
7079 strongly discouraged and is believed to be rare.
7081 Applications that parse structures containing CHOICE or ANY DEFINED BY
7082 components may be affected. Certificate parsing (d2i_X509 and related
7083 functions) are however not affected. OpenSSL clients and servers are
7089 * PKCS7 NULL pointer dereferences fix
7091 The PKCS#7 parsing code does not handle missing outer ContentInfo
7092 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7093 missing content and trigger a NULL pointer dereference on parsing.
7095 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7096 otherwise parse PKCS#7 structures from untrusted sources are
7097 affected. OpenSSL clients and servers are not affected.
7099 This issue was reported to OpenSSL by Michal Zalewski (Google).
7104 * DoS via reachable assert in SSLv2 servers fix
7106 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7107 servers that both support SSLv2 and enable export cipher suites by sending
7108 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7110 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7111 (OpenSSL development team).
7116 * Use After Free following d2i_ECPrivatekey error fix
7118 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7119 could cause a use after free condition. This, in turn, could cause a double
7120 free in several private key parsing functions (such as d2i_PrivateKey
7121 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7122 for applications that receive EC private keys from untrusted
7123 sources. This scenario is considered rare.
7125 This issue was discovered by the BoringSSL project and fixed in their
7131 * X509_to_X509_REQ NULL pointer deref fix
7133 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7134 the certificate key is invalid. This function is rarely used in practice.
7136 This issue was discovered by Brian Carpenter.
7141 * Removed the export ciphers from the DEFAULT ciphers
7145 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7147 * Build fixes for the Windows and OpenVMS platforms
7149 *Matt Caswell and Richard Levitte*
7151 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7153 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7154 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7155 dereference. This could lead to a Denial Of Service attack. Thanks to
7156 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7161 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7162 dtls1_buffer_record function under certain conditions. In particular this
7163 could occur if an attacker sent repeated DTLS records with the same
7164 sequence number but for the next epoch. The memory leak could be exploited
7165 by an attacker in a Denial of Service attack through memory exhaustion.
7166 Thanks to Chris Mueller for reporting this issue.
7171 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7172 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7173 method would be set to NULL which could later result in a NULL pointer
7174 dereference. Thanks to Frank Schmirler for reporting this issue.
7179 * Abort handshake if server key exchange message is omitted for ephemeral
7182 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7183 reporting this issue.
7188 * Remove non-export ephemeral RSA code on client and server. This code
7189 violated the TLS standard by allowing the use of temporary RSA keys in
7190 non-export ciphersuites and could be used by a server to effectively
7191 downgrade the RSA key length used to a value smaller than the server
7192 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7193 INRIA or reporting this issue.
7198 * Fixed issue where DH client certificates are accepted without verification.
7199 An OpenSSL server will accept a DH certificate for client authentication
7200 without the certificate verify message. This effectively allows a client to
7201 authenticate without the use of a private key. This only affects servers
7202 which trust a client certificate authority which issues certificates
7203 containing DH keys: these are extremely rare and hardly ever encountered.
7204 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7210 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7211 results on some platforms, including x86_64. This bug occurs at random
7212 with a very low probability, and is not known to be exploitable in any
7213 way, though its exact impact is difficult to determine. Thanks to Pieter
7214 Wuille (Blockstream) who reported this issue and also suggested an initial
7215 fix. Further analysis was conducted by the OpenSSL development team and
7216 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7217 the OpenSSL core team.
7222 * Fix various certificate fingerprint issues.
7224 By using non-DER or invalid encodings outside the signed portion of a
7225 certificate the fingerprint can be changed without breaking the signature.
7226 Although no details of the signed portion of the certificate can be changed
7227 this can cause problems with some applications: e.g. those using the
7228 certificate fingerprint for blacklists.
7230 1. Reject signatures with non zero unused bits.
7232 If the BIT STRING containing the signature has non zero unused bits reject
7233 the signature. All current signature algorithms require zero unused bits.
7235 2. Check certificate algorithm consistency.
7237 Check the AlgorithmIdentifier inside TBS matches the one in the
7238 certificate signature. NB: this will result in signature failure
7239 errors for some broken certificates.
7241 Thanks to Konrad Kraszewski from Google for reporting this issue.
7243 3. Check DSA/ECDSA signatures use DER.
7245 Reencode DSA/ECDSA signatures and compare with the original received
7246 signature. Return an error if there is a mismatch.
7248 This will reject various cases including garbage after signature
7249 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7250 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7251 (negative or with leading zeroes).
7253 Further analysis was conducted and fixes were developed by Stephen Henson
7254 of the OpenSSL core team.
7260 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7262 * Session Ticket Memory Leak.
7264 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7265 integrity of that ticket is first verified. In the event of a session
7266 ticket integrity check failing, OpenSSL will fail to free memory
7267 causing a memory leak. By sending a large number of invalid session
7268 tickets an attacker could exploit this issue in a Denial Of Service
7274 * Build option no-ssl3 is incomplete.
7276 When OpenSSL is configured with "no-ssl3" as a build option, servers
7277 could accept and complete a SSL 3.0 handshake, and clients could be
7278 configured to send them.
7281 *Akamai and the OpenSSL team*
7283 * Add support for TLS_FALLBACK_SCSV.
7284 Client applications doing fallback retries should call
7285 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7288 *Adam Langley, Bodo Moeller*
7290 * Add additional DigestInfo checks.
7292 Reencode DigestInto in DER and check against the original when
7293 verifying RSA signature: this will reject any improperly encoded
7294 DigestInfo structures.
7296 Note: this is a precautionary measure and no attacks are currently known.
7300 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7302 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7303 to a denial of service attack. A malicious server can crash the client
7304 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7305 ciphersuite and sending carefully crafted handshake messages.
7307 Thanks to Felix Gröbert (Google) for discovering and researching this
7313 * By sending carefully crafted DTLS packets an attacker could cause openssl
7314 to leak memory. This can be exploited through a Denial of Service attack.
7315 Thanks to Adam Langley for discovering and researching this issue.
7320 * An attacker can force openssl to consume large amounts of memory whilst
7321 processing DTLS handshake messages. This can be exploited through a
7322 Denial of Service attack.
7323 Thanks to Adam Langley for discovering and researching this issue.
7328 * An attacker can force an error condition which causes openssl to crash
7329 whilst processing DTLS packets due to memory being freed twice. This
7330 can be exploited through a Denial of Service attack.
7331 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7337 * If a multithreaded client connects to a malicious server using a resumed
7338 session and the server sends an ec point format extension it could write
7339 up to 255 bytes to freed memory.
7341 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7347 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7348 X509_name_oneline, X509_name_print_ex et al. to leak some information
7349 from the stack. Applications may be affected if they echo pretty printing
7350 output to the attacker.
7352 Thanks to Ivan Fratric (Google) for discovering this issue.
7355 *Emilia Käsper, and Steve Henson*
7357 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7358 for corner cases. (Certain input points at infinity could lead to
7359 bogus results, with non-infinity inputs mapped to infinity too.)
7363 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7365 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7366 handshake can force the use of weak keying material in OpenSSL
7367 SSL/TLS clients and servers.
7369 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7370 researching this issue. [CVE-2014-0224][]
7372 *KIKUCHI Masashi, Steve Henson*
7374 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7375 OpenSSL DTLS client the code can be made to recurse eventually crashing
7378 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7381 *Imre Rad, Steve Henson*
7383 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7384 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7385 client or server. This is potentially exploitable to run arbitrary
7386 code on a vulnerable client or server.
7388 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7390 *Jüri Aedla, Steve Henson*
7392 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7393 are subject to a denial of service attack.
7395 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7396 this issue. [CVE-2014-3470][]
7398 *Felix Gröbert, Ivan Fratric, Steve Henson*
7400 * Harmonize version and its documentation. -f flag is used to display
7403 *mancha <mancha1@zoho.com>*
7405 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7406 in i2d_ECPrivateKey.
7408 *mancha <mancha1@zoho.com>*
7410 * Fix some double frees. These are not thought to be exploitable.
7412 *mancha <mancha1@zoho.com>*
7414 * Fix for the attack described in the paper "Recovering OpenSSL
7415 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7416 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7417 <http://eprint.iacr.org/2014/140>
7419 Thanks to Yuval Yarom and Naomi Benger for discovering this
7420 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7422 *Yuval Yarom and Naomi Benger*
7424 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7426 * Keep original DTLS digest and encryption contexts in retransmission
7427 structures so we can use the previous session parameters if they need
7428 to be resent. [CVE-2013-6450][]
7432 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7433 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7434 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7435 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7436 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7437 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7439 *Rob Stradling, Adam Langley*
7441 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7443 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7445 This addresses the flaw in CBC record processing discovered by
7446 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7447 at: <http://www.isg.rhul.ac.uk/tls/>
7449 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7450 Security Group at Royal Holloway, University of London
7451 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7452 Emilia Käsper for the initial patch.
7455 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7457 * Return an error when checking OCSP signatures when key is NULL.
7458 This fixes a DoS attack. [CVE-2013-0166][]
7462 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7463 the right response is stapled. Also change SSL_get_certificate()
7464 so it returns the certificate actually sent.
7465 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7466 (This is a backport)
7468 *Rob Stradling <rob.stradling@comodo.com>*
7470 * Fix possible deadlock when decoding public keys.
7474 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7476 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7479 * Sanity check record length before skipping explicit IV in DTLS
7482 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7483 fuzzing as a service testing platform.
7488 * Initialise tkeylen properly when encrypting CMS messages.
7489 Thanks to Solar Designer of Openwall for reporting this issue.
7493 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7495 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7496 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7497 in CRYPTO_realloc_clean.
7499 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7500 issue and to Adam Langley <agl@chromium.org> for fixing it.
7503 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7505 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7507 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7508 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7509 content decryption and always return the same error. Note: this attack
7510 needs on average 2^20 messages so it only affects automated senders. The
7511 old behaviour can be re-enabled in the CMS code by setting the
7512 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7513 an MMA defence is not necessary.
7514 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7515 this issue. [CVE-2012-0884][]
7519 * Fix CVE-2011-4619: make sure we really are receiving a
7520 client hello before rejecting multiple SGC restarts. Thanks to
7521 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7525 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7527 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7528 Thanks to Antonio Martin, Enterprise Secure Access Research and
7529 Development, Cisco Systems, Inc. for discovering this bug and
7530 preparing a fix. [CVE-2012-0050][]
7534 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7536 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7537 of the Vaudenay padding oracle attack on CBC mode encryption
7538 which enables an efficient plaintext recovery attack against
7539 the OpenSSL implementation of DTLS. Their attack exploits timing
7540 differences arising during decryption processing. A research
7541 paper describing this attack can be found at:
7542 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7543 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7544 Security Group at Royal Holloway, University of London
7545 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7546 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7547 for preparing the fix. [CVE-2011-4108][]
7549 *Robin Seggelmann, Michael Tuexen*
7551 * Clear bytes used for block padding of SSL 3.0 records.
7554 *Adam Langley (Google)*
7556 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7557 Kadianakis <desnacked@gmail.com> for discovering this issue and
7558 Adam Langley for preparing the fix. [CVE-2011-4619][]
7560 *Adam Langley (Google)*
7562 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7564 *Andrey Kulikov <amdeich@gmail.com>*
7566 * Prevent malformed RFC3779 data triggering an assertion failure.
7567 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7568 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7570 *Rob Austein <sra@hactrn.net>*
7572 * Improved PRNG seeding for VOS.
7574 *Paul Green <Paul.Green@stratus.com>*
7576 * Fix ssl_ciph.c set-up race.
7578 *Adam Langley (Google)*
7580 * Fix spurious failures in ecdsatest.c.
7582 *Emilia Käsper (Google)*
7584 * Fix the BIO_f_buffer() implementation (which was mixing different
7585 interpretations of the `..._len` fields).
7587 *Adam Langley (Google)*
7589 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7590 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7591 threads won't reuse the same blinding coefficients.
7593 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7594 lock to call BN_BLINDING_invert_ex, and avoids one use of
7595 BN_BLINDING_update for each BN_BLINDING structure (previously,
7596 the last update always remained unused).
7598 *Emilia Käsper (Google)*
7600 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7602 *Bob Buckholz (Google)*
7604 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7606 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7607 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7609 *Kaspar Brand <ossl@velox.ch>*
7611 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7612 for multi-threaded use of ECDH. [CVE-2011-3210][]
7614 *Adam Langley (Google)*
7616 * Fix x509_name_ex_d2i memory leak on bad inputs.
7620 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7621 signature public key algorithm by using OID xref utilities instead.
7622 Before this you could only use some ECC ciphersuites with SHA1 only.
7626 * Add protection against ECDSA timing attacks as mentioned in the paper
7627 by Billy Bob Brumley and Nicola Tuveri, see:
7628 <http://eprint.iacr.org/2011/232.pdf>
7630 *Billy Bob Brumley and Nicola Tuveri*
7632 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7634 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7636 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7638 * Fix bug in string printing code: if *any* escaping is enabled we must
7639 escape the escape character (backslash) or the resulting string is
7644 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7646 * Disable code workaround for ancient and obsolete Netscape browsers
7647 and servers: an attacker can use it in a ciphersuite downgrade attack.
7648 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7652 * Fixed J-PAKE implementation error, originally discovered by
7653 Sebastien Martini, further info and confirmation from Stefan
7654 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7658 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7660 * Fix extension code to avoid race conditions which can result in a buffer
7661 overrun vulnerability: resumed sessions must not be modified as they can
7662 be shared by multiple threads. CVE-2010-3864
7666 * Fix WIN32 build system to correctly link an ENGINE directory into
7671 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7673 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7676 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7678 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7680 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7681 context. The operation can be customised via the ctrl mechanism in
7682 case ENGINEs want to include additional functionality.
7686 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7690 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7691 output hashes compatible with older versions of OpenSSL.
7693 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7695 * Fix compression algorithm handling: if resuming a session use the
7696 compression algorithm of the resumed session instead of determining
7697 it from client hello again. Don't allow server to change algorithm.
7701 * Add load_crls() function to apps tidying load_certs() too. Add option
7702 to verify utility to allow additional CRLs to be included.
7706 * Update OCSP request code to permit adding custom headers to the request:
7707 some responders need this.
7711 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7714 *Julia Lawall <julia@diku.dk>*
7716 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7717 needlessly dereferenced structures, used obsolete functions and
7718 didn't handle all updated verify codes correctly.
7722 * Disable MD2 in the default configuration.
7726 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7727 indicate the initial BIO being pushed or popped. This makes it possible
7728 to determine whether the BIO is the one explicitly called or as a result
7729 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7730 it handles reference counts correctly and doesn't zero out the I/O bio
7731 when it is not being explicitly popped. WARNING: applications which
7732 included workarounds for the old buggy behaviour will need to be modified
7733 or they could free up already freed BIOs.
7737 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7738 renaming to all platforms (within the 0.9.8 branch, this was
7739 done conditionally on Netware platforms to avoid a name clash).
7741 *Guenter <lists@gknw.net>*
7743 * Add ECDHE and PSK support to DTLS.
7745 *Michael Tuexen <tuexen@fh-muenster.de>*
7747 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7752 * Add "missing" function EVP_MD_flags() (without this the only way to
7753 retrieve a digest flags is by accessing the structure directly. Update
7754 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7755 or cipher is registered as in the "from" argument. Print out all
7756 registered digests in the dgst usage message instead of manually
7757 attempting to work them out.
7761 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7762 this allows the use of compression and extensions. Change default cipher
7763 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7764 by default unless an application cipher string requests it.
7768 * Alter match criteria in PKCS12_parse(). It used to try to use local
7769 key ids to find matching certificates and keys but some PKCS#12 files
7770 don't follow the (somewhat unwritten) rules and this strategy fails.
7771 Now just gather all certificates together and the first private key
7772 then look for the first certificate that matches the key.
7776 * Support use of registered digest and cipher names for dgst and cipher
7777 commands instead of having to add each one as a special case. So now
7784 openssl dgst -sha256 foo
7786 and this works for ENGINE based algorithms too.
7790 * Update Gost ENGINE to support parameter files.
7792 *Victor B. Wagner <vitus@cryptocom.ru>*
7794 * Support GeneralizedTime in ca utility.
7796 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7798 * Enhance the hash format used for certificate directory links. The new
7799 form uses the canonical encoding (meaning equivalent names will work
7800 even if they aren't identical) and uses SHA1 instead of MD5. This form
7801 is incompatible with the older format and as a result c_rehash should
7802 be used to rebuild symbolic links.
7806 * Make PKCS#8 the default write format for private keys, replacing the
7807 traditional format. This form is standardised, more secure and doesn't
7808 include an implicit MD5 dependency.
7812 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7813 committed to OpenSSL should pass this lot as a minimum.
7817 * Add session ticket override functionality for use by EAP-FAST.
7819 *Jouni Malinen <j@w1.fi>*
7821 * Modify HMAC functions to return a value. Since these can be implemented
7822 in an ENGINE errors can occur.
7826 * Type-checked OBJ_bsearch_ex.
7830 * Type-checked OBJ_bsearch. Also some constification necessitated
7831 by type-checking. Still to come: TXT_DB, bsearch(?),
7832 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7837 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7838 seconds to a tm structure directly, instead of going through OS
7839 specific date routines. This avoids any issues with OS routines such
7840 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7841 and X509_time_adj_ex() to cover the extended range. The existing
7842 X509_time_adj() is still usable and will no longer have any date issues.
7846 * Delta CRL support. New use deltas option which will attempt to locate
7847 and search any appropriate delta CRLs available.
7849 This work was sponsored by Google.
7853 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7854 code and add additional score elements. Validate alternate CRL paths
7855 as part of the CRL checking and indicate a new error "CRL path validation
7856 error" in this case. Applications wanting additional details can use
7857 the verify callback and check the new "parent" field. If this is not
7858 NULL CRL path validation is taking place. Existing applications won't
7859 see this because it requires extended CRL support which is off by
7862 This work was sponsored by Google.
7866 * Support for freshest CRL extension.
7868 This work was sponsored by Google.
7872 * Initial indirect CRL support. Currently only supported in the CRLs
7873 passed directly and not via lookup. Process certificate issuer
7874 CRL entry extension and lookup CRL entries by bother issuer name
7875 and serial number. Check and process CRL issuer entry in IDP extension.
7877 This work was sponsored by Google.
7881 * Add support for distinct certificate and CRL paths. The CRL issuer
7882 certificate is validated separately in this case. Only enabled if
7883 an extended CRL support flag is set: this flag will enable additional
7884 CRL functionality in future.
7886 This work was sponsored by Google.
7890 * Add support for policy mappings extension.
7892 This work was sponsored by Google.
7896 * Fixes to pathlength constraint, self issued certificate handling,
7897 policy processing to align with RFC3280 and PKITS tests.
7899 This work was sponsored by Google.
7903 * Support for name constraints certificate extension. DN, email, DNS
7904 and URI types are currently supported.
7906 This work was sponsored by Google.
7910 * To cater for systems that provide a pointer-based thread ID rather
7911 than numeric, deprecate the current numeric thread ID mechanism and
7912 replace it with a structure and associated callback type. This
7913 mechanism allows a numeric "hash" to be extracted from a thread ID in
7914 either case, and on platforms where pointers are larger than 'long',
7915 mixing is done to help ensure the numeric 'hash' is usable even if it
7916 can't be guaranteed unique. The default mechanism is to use "&errno"
7917 as a pointer-based thread ID to distinguish between threads.
7919 Applications that want to provide their own thread IDs should now use
7920 CRYPTO_THREADID_set_callback() to register a callback that will call
7921 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7923 Note that ERR_remove_state() is now deprecated, because it is tied
7924 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7925 to free the current thread's error state should be replaced by
7926 ERR_remove_thread_state(NULL).
7928 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7929 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7930 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7931 application was previously providing a numeric thread callback that
7932 was inappropriate for distinguishing threads, then uniqueness might
7933 have been obtained with &errno that happened immediately in the
7934 intermediate development versions of OpenSSL; this is no longer the
7935 case, the numeric thread callback will now override the automatic use
7938 *Geoff Thorpe, with help from Bodo Moeller*
7940 * Initial support for different CRL issuing certificates. This covers a
7941 simple case where the self issued certificates in the chain exist and
7942 the real CRL issuer is higher in the existing chain.
7944 This work was sponsored by Google.
7948 * Removed effectively defunct crypto/store from the build.
7952 * Revamp of STACK to provide stronger type-checking. Still to come:
7953 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7954 ASN1_STRING, CONF_VALUE.
7958 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7959 RAM on SSL connections. This option can save about 34k per idle SSL.
7963 * Revamp of LHASH to provide stronger type-checking. Still to come:
7964 STACK, TXT_DB, bsearch, qsort.
7968 * Initial support for Cryptographic Message Syntax (aka CMS) based
7969 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7970 support for data, signedData, compressedData, digestedData and
7971 encryptedData, envelopedData types included. Scripts to check against
7972 RFC4134 examples draft and interop and consistency checks of many
7973 content types and variants.
7977 * Add options to enc utility to support use of zlib compression BIO.
7981 * Extend mk1mf to support importing of options and assembly language
7982 files from Configure script, currently only included in VC-WIN32.
7983 The assembly language rules can now optionally generate the source
7984 files from the associated perl scripts.
7988 * Implement remaining functionality needed to support GOST ciphersuites.
7989 Interop testing has been performed using CryptoPro implementations.
7991 *Victor B. Wagner <vitus@cryptocom.ru>*
7993 * s390x assembler pack.
7997 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8002 * Implement Opaque PRF Input TLS extension as specified in
8003 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8004 official specification yet and no extension type assignment by
8005 IANA exists, this extension (for now) will have to be explicitly
8006 enabled when building OpenSSL by providing the extension number
8007 to use. For example, specify an option
8009 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8011 to the "config" or "Configure" script to enable the extension,
8012 assuming extension number 0x9527 (which is a completely arbitrary
8013 and unofficial assignment based on the MD5 hash of the Internet
8014 Draft). Note that by doing so, you potentially lose
8015 interoperability with other TLS implementations since these might
8016 be using the same extension number for other purposes.
8018 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8019 opaque PRF input value to use in the handshake. This will create
8020 an internal copy of the length-'len' string at 'src', and will
8021 return non-zero for success.
8023 To get more control and flexibility, provide a callback function
8026 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8027 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8031 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8034 Callback function 'cb' will be called in handshakes, and is
8035 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8036 Argument 'arg' is for application purposes (the value as given to
8037 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8038 be provided to the callback function). The callback function
8039 has to return non-zero to report success: usually 1 to use opaque
8040 PRF input just if possible, or 2 to enforce use of the opaque PRF
8041 input. In the latter case, the library will abort the handshake
8042 if opaque PRF input is not successfully negotiated.
8044 Arguments 'peerinput' and 'len' given to the callback function
8045 will always be NULL and 0 in the case of a client. A server will
8046 see the client's opaque PRF input through these variables if
8047 available (NULL and 0 otherwise). Note that if the server
8048 provides an opaque PRF input, the length must be the same as the
8049 length of the client's opaque PRF input.
8051 Note that the callback function will only be called when creating
8052 a new session (session resumption can resume whatever was
8053 previously negotiated), and will not be called in SSL 2.0
8054 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8055 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8056 for applications that need to enforce opaque PRF input.
8060 * Update ssl code to support digests other than SHA1+MD5 for handshake
8063 *Victor B. Wagner <vitus@cryptocom.ru>*
8065 * Add RFC4507 support to OpenSSL. This includes the corrections in
8066 RFC4507bis. The encrypted ticket format is an encrypted encoded
8067 SSL_SESSION structure, that way new session features are automatically
8070 If a client application caches session in an SSL_SESSION structure
8071 support is transparent because tickets are now stored in the encoded
8074 The SSL_CTX structure automatically generates keys for ticket
8075 protection in servers so again support should be possible
8076 with no application modification.
8078 If a client or server wishes to disable RFC4507 support then the option
8079 SSL_OP_NO_TICKET can be set.
8081 Add a TLS extension debugging callback to allow the contents of any client
8082 or server extensions to be examined.
8084 This work was sponsored by Google.
8088 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8089 OpenSSL should now compile cleanly on gcc 4.2
8091 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8093 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8094 support including streaming MAC support: this is required for GOST
8095 ciphersuite support.
8097 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8099 * Add option -stream to use PKCS#7 streaming in smime utility. New
8100 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8101 to output in BER and PEM format.
8105 * Experimental support for use of HMAC via EVP_PKEY interface. This
8106 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8107 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8108 ENGINE support for HMAC keys which are unextractable. New -mac and
8109 -macopt options to dgst utility.
8113 * New option -sigopt to dgst utility. Update dgst to use
8114 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8115 alternative signing parameters such as X9.31 or PSS in the dgst
8120 * Change ssl_cipher_apply_rule(), the internal function that does
8121 the work each time a ciphersuite string requests enabling
8122 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8123 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8124 the order of disabled ciphersuites such that those ciphersuites
8125 that most recently went from enabled to disabled not only stay
8126 in order with respect to each other, but also have higher priority
8127 than other disabled ciphersuites the next time ciphersuites are
8130 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8131 the same ciphersuites as with "HIGH" alone, but in a specific
8132 order where the PSK ciphersuites come first (since they are the
8133 most recently disabled ciphersuites when "HIGH" is parsed).
8135 Also, change ssl_create_cipher_list() (using this new
8136 functionality) such that between otherwise identical
8137 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8142 * Change ssl_create_cipher_list() so that it automatically
8143 arranges the ciphersuites in reasonable order before starting
8144 to process the rule string. Thus, the definition for "DEFAULT"
8145 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8146 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8147 This makes it much easier to arrive at a reasonable default order
8148 in applications for which anonymous ciphers are OK (meaning
8149 that you can't actually use DEFAULT).
8151 *Bodo Moeller; suggested by Victor Duchovni*
8153 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8154 processing) into multiple integers instead of setting
8155 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8156 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8157 (These masks as well as the individual bit definitions are hidden
8158 away into the non-exported interface ssl/ssl_locl.h, so this
8159 change to the definition of the SSL_CIPHER structure shouldn't
8160 affect applications.) This give us more bits for each of these
8161 categories, so there is no longer a need to coagulate AES128 and
8162 AES256 into a single algorithm bit, and to coagulate Camellia128
8163 and Camellia256 into a single algorithm bit, which has led to all
8166 Thus, among other things, the kludge introduced in 0.9.7m and
8167 0.9.8e for masking out AES256 independently of AES128 or masking
8168 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8170 With the change, we also introduce new ciphersuite aliases that
8171 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8176 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8177 Use the leftmost N bytes of the signature input if the input is
8178 larger than the prime q (with N being the size in bytes of q).
8182 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8183 it yet and it is largely untested.
8187 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8191 * Initial incomplete changes to avoid need for function casts in OpenSSL
8192 some compilers (gcc 4.2 and later) reject their use. Safestack is
8193 reimplemented. Update ASN1 to avoid use of legacy functions.
8197 * Win32/64 targets are linked with Winsock2.
8201 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8202 to external functions. This can be used to increase CRL handling
8203 efficiency especially when CRLs are very large by (for example) storing
8204 the CRL revoked certificates in a database.
8208 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8209 new CRLs added to a directory can be used. New command line option
8210 -verify_return_error to s_client and s_server. This causes real errors
8211 to be returned by the verify callback instead of carrying on no matter
8212 what. This reflects the way a "real world" verify callback would behave.
8216 * GOST engine, supporting several GOST algorithms and public key formats.
8217 Kindly donated by Cryptocom.
8221 * Partial support for Issuing Distribution Point CRL extension. CRLs
8222 partitioned by DP are handled but no indirect CRL or reason partitioning
8223 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8224 selected via a scoring technique which handles IDP and AKID in CRLs.
8228 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8229 will ultimately be used for all verify operations: this will remove the
8230 X509_STORE dependency on certificate verification and allow alternative
8231 lookup methods. X509_STORE based implementations of these two callbacks.
8235 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8236 Modify get_crl() to find a valid (unexpired) CRL if possible.
8240 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8241 this would be called X509_CRL_cmp() but that name is already used by
8242 a function that just compares CRL issuer names. Cache several CRL
8243 extensions in X509_CRL structure and cache CRLDP in X509.
8247 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8248 this maps equivalent X509_NAME structures into a consistent structure.
8249 Name comparison can then be performed rapidly using memcmp().
8253 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8258 * Allow digests to supply their own micalg string for S/MIME type using
8259 the ctrl EVP_MD_CTRL_MICALG.
8263 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8264 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8265 ctrl. It can then customise the structure before and/or after signing
8270 * New function OBJ_add_sigid() to allow application defined signature OIDs
8271 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8272 to free up any added signature OIDs.
8276 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8277 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8278 digest and cipher tables. New options added to openssl utility:
8279 list-message-digest-algorithms and list-cipher-algorithms.
8283 * Change the array representation of binary polynomials: the list
8284 of degrees of non-zero coefficients is now terminated with -1.
8285 Previously it was terminated with 0, which was also part of the
8286 value; thus, the array representation was not applicable to
8287 polynomials where t^0 has coefficient zero. This change makes
8288 the array representation useful in a more general context.
8292 * Various modifications and fixes to SSL/TLS cipher string
8293 handling. For ECC, the code now distinguishes between fixed ECDH
8294 with RSA certificates on the one hand and with ECDSA certificates
8295 on the other hand, since these are separate ciphersuites. The
8296 unused code for Fortezza ciphersuites has been removed.
8298 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8299 (not "ECDHE"). For consistency with the code for DH
8300 certificates, use of ECDH certificates is now considered ECDH
8301 authentication, not RSA or ECDSA authentication (the latter is
8302 merely the CA's signing algorithm and not actively used in the
8305 The temporary ciphersuite alias "ECCdraft" is no longer
8306 available, and ECC ciphersuites are no longer excluded from "ALL"
8307 and "DEFAULT". The following aliases now exist for RFC 4492
8308 ciphersuites, most of these by analogy with the DH case:
8310 kECDHr - ECDH cert, signed with RSA
8311 kECDHe - ECDH cert, signed with ECDSA
8312 kECDH - ECDH cert (signed with either RSA or ECDSA)
8313 kEECDH - ephemeral ECDH
8314 ECDH - ECDH cert or ephemeral ECDH
8320 AECDH - anonymous ECDH
8321 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8325 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8326 Use correct micalg parameters depending on digest(s) in signed message.
8330 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8331 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8335 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8336 an engine to register a method. Add ENGINE lookups for methods and
8337 functional reference processing.
8341 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8342 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8347 * New -resign option to smime utility. This adds one or more signers
8348 to an existing PKCS#7 signedData structure. Also -md option to use an
8349 alternative message digest algorithm for signing.
8353 * Tidy up PKCS#7 routines and add new functions to make it easier to
8354 create PKCS7 structures containing multiple signers. Update smime
8355 application to support multiple signers.
8359 * New -macalg option to pkcs12 utility to allow setting of an alternative
8364 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8365 Reorganize PBE internals to lookup from a static table using NIDs,
8366 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8367 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8368 PRF which will be automatically used with PBES2.
8372 * Replace the algorithm specific calls to generate keys in "req" with the
8377 * Update PKCS#7 enveloped data routines to use new API. This is now
8378 supported by any public key method supporting the encrypt operation. A
8379 ctrl is added to allow the public key algorithm to examine or modify
8380 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8385 * Add a ctrl to asn1 method to allow a public key algorithm to express
8386 a default digest type to use. In most cases this will be SHA1 but some
8387 algorithms (such as GOST) need to specify an alternative digest. The
8388 return value indicates how strong the preference is 1 means optional and
8389 2 is mandatory (that is it is the only supported type). Modify
8390 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8391 use the default md. Update openssl utilities to use the default digest
8392 type for signing if it is not explicitly indicated.
8396 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8397 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8398 signing method from the key type. This effectively removes the link
8399 between digests and public key types.
8403 * Add an OID cross reference table and utility functions. Its purpose is to
8404 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8405 rsaEncryption. This will allow some of the algorithm specific hackery
8406 needed to use the correct OID to be removed.
8410 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8411 structures for PKCS7_sign(). They are now set up by the relevant public
8416 * Add provisional EC pkey method with support for ECDSA and ECDH.
8420 * Add support for key derivation (agreement) in the API, DH method and
8425 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8426 public and private key formats. As a side effect these add additional
8427 command line functionality not previously available: DSA signatures can be
8428 generated and verified using pkeyutl and DH key support and generation in
8435 *Oliver Tappe <zooey@hirschkaefer.de>*
8437 * New make target "install_html_docs" installs HTML renditions of the
8440 *Oliver Tappe <zooey@hirschkaefer.de>*
8442 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8443 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8444 support key and parameter generation and add initial key generation
8445 functionality for RSA.
8449 * Add functions for main EVP_PKEY_method operations. The undocumented
8450 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8451 `EVP_PKEY_{encrypt,decrypt}_old`.
8455 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8456 key API, doesn't do much yet.
8460 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8461 public key algorithms. New option to openssl utility:
8462 "list-public-key-algorithms" to print out info.
8466 * Implement the Supported Elliptic Curves Extension for
8467 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8471 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8472 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8476 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8477 utilities such as rsa, dsa, dsaparam etc except they process any key
8482 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8483 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8484 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8489 * Initial support for pluggable public key ASN1.
8490 De-spaghettify the public key ASN1 handling. Move public and private
8491 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8492 algorithm specific handling to a single module within the relevant
8493 algorithm directory. Add functions to allow (near) opaque processing
8494 of public and private key structures.
8498 * Implement the Supported Point Formats Extension for
8499 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8503 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8504 for the psk identity [hint] and the psk callback functions to the
8505 SSL_SESSION, SSL and SSL_CTX structure.
8508 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8512 SSL_CTX_use_psk_identity_hint
8513 SSL_get_psk_identity_hint
8514 SSL_get_psk_identity
8515 SSL_use_psk_identity_hint
8517 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8519 * Add RFC 3161 compliant time stamp request creation, response generation
8520 and response verification functionality.
8522 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8524 * Add initial support for TLS extensions, specifically for the server_name
8525 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8526 have new members for a host name. The SSL data structure has an
8527 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8528 stored in that context to allow for session resumption, even after the
8529 SSL has been switched to a new SSL_CTX in reaction to a client's
8530 server_name extension.
8532 New functions (subject to change):
8534 SSL_get_servername()
8535 SSL_get_servername_type()
8538 New CTRL codes and macros (subject to change):
8540 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8541 - SSL_CTX_set_tlsext_servername_callback()
8542 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8543 - SSL_CTX_set_tlsext_servername_arg()
8544 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8546 openssl s_client has a new '-servername ...' option.
8548 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8549 '-key2 ...', '-servername_fatal' (subject to change). This allows
8550 testing the HostName extension for a specific single host name ('-cert'
8551 and '-key' remain fallbacks for handshakes without HostName
8552 negotiation). If the unrecognized_name alert has to be sent, this by
8553 default is a warning; it becomes fatal with the '-servername_fatal'
8556 *Peter Sylvester, Remy Allais, Christophe Renou*
8558 * Whirlpool hash implementation is added.
8562 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8563 bn(64,32). Because of instruction set limitations it doesn't have
8564 any negative impact on performance. This was done mostly in order
8565 to make it possible to share assembler modules, such as bn_mul_mont
8566 implementations, between 32- and 64-bit builds without hassle.
8570 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8571 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8576 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8577 dedicated Montgomery multiplication procedure, is introduced.
8578 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8579 "64-bit" performance on certain 32-bit targets.
8583 * New option SSL_OP_NO_COMP to disable use of compression selectively
8584 in SSL structures. New SSL ctrl to set maximum send fragment size.
8585 Save memory by setting the I/O buffer sizes dynamically instead of
8586 using the maximum available value.
8590 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8591 in addition to the text details.
8595 * Very, very preliminary EXPERIMENTAL support for printing of general
8596 ASN1 structures. This currently produces rather ugly output and doesn't
8597 handle several customised structures at all.
8601 * Integrated support for PVK file format and some related formats such
8602 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8603 these in the 'rsa' and 'dsa' utilities.
8607 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8611 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8612 place for the (very old) "NETSCAPE" format certificates which are now
8613 handled using new ASN1 code equivalents.
8617 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8618 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8619 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8623 * Modify CRL distribution points extension code to print out previously
8624 unsupported fields. Enhance extension setting code to allow setting of
8629 * Add print and set support for Issuing Distribution Point CRL extension.
8633 * Change 'Configure' script to enable Camellia by default.
8640 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8642 * When rejecting SSL/TLS records due to an incorrect version number, never
8643 update s->server with a new major version number. As of
8644 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8645 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8646 the previous behavior could result in a read attempt at NULL when
8647 receiving specific incorrect SSL/TLS records once record payload
8648 protection is active. [CVE-2010-0740][]
8650 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8652 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8653 could be crashed if the relevant tables were not present (e.g. chrooted).
8655 *Tomas Hoger <thoger@redhat.com>*
8657 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8659 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8661 *Martin Olsson, Neel Mehta*
8663 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8664 accommodate for stack sorting, always a write lock!).
8668 * On some versions of WIN32 Heap32Next is very slow. This can cause
8669 excessive delays in the RAND_poll(): over a minute. As a workaround
8670 include a time check in the inner Heap32Next loop too.
8674 * The code that handled flushing of data in SSL/TLS originally used the
8675 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8676 the problem outlined in PR#1949. The fix suggested there however can
8677 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8678 of Apache). So instead simplify the code to flush unconditionally.
8679 This should be fine since flushing with no data to flush is a no op.
8683 * Handle TLS versions 2.0 and later properly and correctly use the
8684 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8685 off ancient servers have a habit of sticking around for a while...
8689 * Modify compression code so it frees up structures without using the
8690 ex_data callbacks. This works around a problem where some applications
8691 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8692 restarting) then use compression (e.g. SSL with compression) later.
8693 This results in significant per-connection memory leaks and
8694 has caused some security issues including CVE-2008-1678 and
8699 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8700 change when encrypting or decrypting.
8704 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8705 connect and renegotiate with servers which do not support RI.
8706 Until RI is more widely deployed this option is enabled by default.
8710 * Add "missing" ssl ctrls to clear options and mode.
8714 * If client attempts to renegotiate and doesn't support RI respond with
8715 a no_renegotiation alert as required by RFC5746. Some renegotiating
8716 TLS clients will continue a connection gracefully when they receive
8717 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8718 waiting for a server hello which it will never receive. Now we treat a
8719 received no_renegotiation alert as a fatal error. This is because
8720 applications requesting a renegotiation might well expect it to succeed
8721 and would have no code in place to handle the server denying it so the
8722 only safe thing to do is to terminate the connection.
8726 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8727 peer supports secure renegotiation and 0 otherwise. Print out peer
8728 renegotiation support in s_client/s_server.
8732 * Replace the highly broken and deprecated SPKAC certification method with
8733 the updated NID creation version. This should correctly handle UTF8.
8737 * Implement RFC5746. Re-enable renegotiation but require the extension
8738 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8739 turns out to be a bad idea. It has been replaced by
8740 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8741 SSL_CTX_set_options(). This is really not recommended unless you
8742 know what you are doing.
8744 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8746 * Fixes to stateless session resumption handling. Use initial_ctx when
8747 issuing and attempting to decrypt tickets in case it has changed during
8748 servername handling. Use a non-zero length session ID when attempting
8749 stateless session resumption: this makes it possible to determine if
8750 a resumption has occurred immediately after receiving server hello
8751 (several places in OpenSSL subtly assume this) instead of later in
8756 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8757 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8758 fixes for a few places where the return code is not checked
8761 *Julia Lawall <julia@diku.dk>*
8763 * Add --strict-warnings option to Configure script to include devteam
8764 warnings in other configurations.
8768 * Add support for --libdir option and LIBDIR variable in makefiles. This
8769 makes it possible to install openssl libraries in locations which
8770 have names other than "lib", for example "/usr/lib64" which some
8773 *Steve Henson, based on patch from Jeremy Utley*
8775 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8776 X690 8.9.12 and can produce some misleading textual output of OIDs.
8778 *Steve Henson, reported by Dan Kaminsky*
8780 * Delete MD2 from algorithm tables. This follows the recommendation in
8781 several standards that it is not used in new applications due to
8782 several cryptographic weaknesses. For binary compatibility reasons
8783 the MD2 API is still compiled in by default.
8787 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8792 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8793 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8796 *Guenter <lists@gknw.net>*
8798 * Fix the server certificate chain building code to use X509_verify_cert(),
8799 it used to have an ad-hoc builder which was unable to cope with anything
8800 other than a simple chain.
8802 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8804 * Don't check self signed certificate signatures in X509_verify_cert()
8805 by default (a flag can override this): it just wastes time without
8806 adding any security. As a useful side effect self signed root CAs
8807 with non-FIPS digests are now usable in FIPS mode.
8811 * In dtls1_process_out_of_seq_message() the check if the current message
8812 is already buffered was missing. For every new message was memory
8813 allocated, allowing an attacker to perform an denial of service attack
8814 with sending out of seq handshake messages until there is no memory
8815 left. Additionally every future message was buffered, even if the
8816 sequence number made no sense and would be part of another handshake.
8817 So only messages with sequence numbers less than 10 in advance will be
8818 buffered. [CVE-2009-1378][]
8820 *Robin Seggelmann, discovered by Daniel Mentz*
8822 * Records are buffered if they arrive with a future epoch to be
8823 processed after finishing the corresponding handshake. There is
8824 currently no limitation to this buffer allowing an attacker to perform
8825 a DOS attack with sending records with future epochs until there is no
8826 memory left. This patch adds the pqueue_size() function to determine
8827 the size of a buffer and limits the record buffer to 100 entries.
8830 *Robin Seggelmann, discovered by Daniel Mentz*
8832 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8833 parent structure is freed. [CVE-2009-1379][]
8837 * Handle non-blocking I/O properly in SSL_shutdown() call.
8839 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8841 * Add `2.5.4.*` OIDs
8843 *Ilya O. <vrghost@gmail.com>*
8845 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8847 * Disable renegotiation completely - this fixes a severe security
8848 problem [CVE-2009-3555][] at the cost of breaking all
8849 renegotiation. Renegotiation can be re-enabled by setting
8850 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8851 run-time. This is really not recommended unless you know what
8856 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8858 * Don't set val to NULL when freeing up structures, it is freed up by
8859 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8860 zeroing past the valid field. [CVE-2009-0789][]
8862 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8864 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8865 checked correctly. This would allow some invalid signed attributes to
8866 appear to verify correctly. [CVE-2009-0591][]
8868 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8870 * Reject UniversalString and BMPString types with invalid lengths. This
8871 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8872 a legal length. [CVE-2009-0590][]
8876 * Set S/MIME signing as the default purpose rather than setting it
8877 unconditionally. This allows applications to override it at the store
8882 * Permit restricted recursion of ASN1 strings. This is needed in practice
8883 to handle some structures.
8887 * Improve efficiency of mem_gets: don't search whole buffer each time
8890 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8892 * New -hex option for openssl rand.
8896 * Print out UTF8String and NumericString when parsing ASN1.
8900 * Support NumericString type for name components.
8904 * Allow CC in the environment to override the automatically chosen
8905 compiler. Note that nothing is done to ensure flags work with the
8910 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8912 * Properly check EVP_VerifyFinal() and similar return values
8915 *Ben Laurie, Bodo Moeller, Google Security Team*
8917 * Enable TLS extensions by default.
8921 * Allow the CHIL engine to be loaded, whether the application is
8922 multithreaded or not. (This does not release the developer from the
8923 obligation to set up the dynamic locking callbacks.)
8925 *Sander Temme <sander@temme.net>*
8927 * Use correct exit code if there is an error in dgst command.
8929 *Steve Henson; problem pointed out by Roland Dirlewanger*
8931 * Tweak Configure so that you need to say "experimental-jpake" to enable
8932 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8936 * Add experimental JPAKE support, including demo authentication in
8937 s_client and s_server.
8941 * Set the comparison function in v3_addr_canonize().
8943 *Rob Austein <sra@hactrn.net>*
8945 * Add support for XMPP STARTTLS in s_client.
8947 *Philip Paeps <philip@freebsd.org>*
8949 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8950 to ensure that even with this option, only ciphersuites in the
8951 server's preference list will be accepted. (Note that the option
8952 applies only when resuming a session, so the earlier behavior was
8953 just about the algorithm choice for symmetric cryptography.)
8957 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8959 * Fix NULL pointer dereference if a DTLS server received
8960 ChangeCipherSpec as first record [CVE-2009-1386][].
8964 * Fix a state transition in s3_srvr.c and d1_srvr.c
8965 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8969 * The fix in 0.9.8c that supposedly got rid of unsafe
8970 double-checked locking was incomplete for RSA blinding,
8971 addressing just one layer of what turns out to have been
8972 doubly unsafe triple-checked locking.
8974 So now fix this for real by retiring the MONT_HELPER macro
8975 in crypto/rsa/rsa_eay.c.
8977 *Bodo Moeller; problem pointed out by Marius Schilder*
8979 * Various precautionary measures:
8981 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8983 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8984 (NB: This would require knowledge of the secret session ticket key
8985 to exploit, in which case you'd be SOL either way.)
8987 - Change bn_nist.c so that it will properly handle input BIGNUMs
8988 outside the expected range.
8990 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8993 *Neel Mehta, Bodo Moeller*
8995 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8996 the load fails. Useful for distros.
8998 *Ben Laurie and the FreeBSD team*
9000 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9004 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9008 * Expand ENGINE to support engine supplied SSL client certificate functions.
9010 This work was sponsored by Logica.
9014 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9015 keystores. Support for SSL/TLS client authentication too.
9016 Not compiled unless enable-capieng specified to Configure.
9018 This work was sponsored by Logica.
9022 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9023 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9024 attribute creation routines such as certificate requests and PKCS#12
9029 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9031 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9032 handshake which could lead to a client crash as found using the
9033 Codenomicon TLS test suite [CVE-2008-1672][]
9035 *Steve Henson, Mark Cox*
9037 * Fix double free in TLS server name extensions which could lead to
9038 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
9042 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9044 Clear the error queue to ensure that error entries left from
9045 older function calls do not interfere with the correct operation.
9047 *Lutz Jaenicke, Erik de Castro Lopo*
9049 * Remove root CA certificates of commercial CAs:
9051 The OpenSSL project does not recommend any specific CA and does not
9052 have any policy with respect to including or excluding any CA.
9053 Therefore it does not make any sense to ship an arbitrary selection
9054 of root CA certificates with the OpenSSL software.
9058 * RSA OAEP patches to fix two separate invalid memory reads.
9059 The first one involves inputs when 'lzero' is greater than
9060 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9061 before the beginning of from). The second one involves inputs where
9062 the 'db' section contains nothing but zeroes (there is a one-byte
9063 invalid read after the end of 'db').
9065 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9067 * Partial backport from 0.9.9-dev:
9069 Introduce bn_mul_mont (dedicated Montgomery multiplication
9070 procedure) as a candidate for BIGNUM assembler implementation.
9071 While 0.9.9-dev uses assembler for various architectures, only
9072 x86_64 is available by default here in the 0.9.8 branch, and
9073 32-bit x86 is available through a compile-time setting.
9075 To try the 32-bit x86 assembler implementation, use Configure
9076 option "enable-montasm" (which exists only for this backport).
9078 As "enable-montasm" for 32-bit x86 disclaims code stability
9079 anyway, in this constellation we activate additional code
9080 backported from 0.9.9-dev for further performance improvements,
9081 namely BN_from_montgomery_word. (To enable this otherwise,
9082 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9084 *Andy Polyakov (backport partially by Bodo Moeller)*
9086 * Add TLS session ticket callback. This allows an application to set
9087 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9088 values. This is useful for key rollover for example where several key
9089 sets may exist with different names.
9093 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9094 This was broken until now in 0.9.8 releases, such that the only way
9095 a registered ENGINE could be used (assuming it initialises
9096 successfully on the host) was to explicitly set it as the default
9097 for the relevant algorithms. This is in contradiction with 0.9.7
9098 behaviour and the documentation. With this fix, when an ENGINE is
9099 registered into a given algorithm's table of implementations, the
9100 'uptodate' flag is reset so that auto-discovery will be used next
9101 time a new context for that algorithm attempts to select an
9104 *Ian Lister (tweaked by Geoff Thorpe)*
9106 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9107 implementation in the following ways:
9109 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9112 Lack of BER streaming support means one pass streaming processing is
9113 only supported if data is detached: setting the streaming flag is
9114 ignored for embedded content.
9116 CMS support is disabled by default and must be explicitly enabled
9117 with the enable-cms configuration option.
9121 * Update the GMP engine glue to do direct copies between BIGNUM and
9122 mpz_t when openssl and GMP use the same limb size. Otherwise the
9123 existing "conversion via a text string export" trick is still used.
9125 *Paul Sheer <paulsheer@gmail.com>*
9127 * Zlib compression BIO. This is a filter BIO which compressed and
9128 uncompresses any data passed through it.
9132 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9133 RFC3394 compatible AES key wrapping.
9137 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9138 sets string data without copying. X509_ALGOR_set0() and
9139 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9140 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9141 from an X509_ATTRIBUTE structure optionally checking it occurs only
9142 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9147 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9148 to get the expected BN_FLG_CONSTTIME behavior.
9150 *Bodo Moeller (Google)*
9154 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9155 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9156 - added some more tests to do_tests.pl
9157 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9158 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9159 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9160 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9161 - various changes to netware.pl to enable gcc-cross builds on Win32
9163 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9164 - various changes to fix missing prototype warnings
9165 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9166 - added AES, WHIRLPOOL and CPUID assembler code to build files
9167 - added missing AES assembler make rules to mk1mf.pl
9168 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9170 *Guenter Knauf <eflash@gmx.net>*
9172 * Implement certificate status request TLS extension defined in RFC3546.
9173 A client can set the appropriate parameters and receive the encoded
9174 OCSP response via a callback. A server can query the supplied parameters
9175 and set the encoded OCSP response in the callback. Add simplified examples
9176 to s_client and s_server.
9180 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9183 + Binary incompatibility of ssl_ctx_st structure
9184 + DTLS interoperation with non-compliant servers
9185 + Don't call get_session_cb() without proposed session
9186 + Fix ia64 assembler code
9188 *Andy Polyakov, Steve Henson*
9190 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9192 * DTLS Handshake overhaul. There were longstanding issues with
9193 OpenSSL DTLS implementation, which were making it impossible for
9194 RFC 4347 compliant client to communicate with OpenSSL server.
9195 Unfortunately just fixing these incompatibilities would "cut off"
9196 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9197 server keeps tolerating non RFC compliant syntax. The opposite is
9198 not true, 0.9.8f client can not communicate with earlier server.
9199 This update even addresses CVE-2007-4995.
9203 * Changes to avoid need for function casts in OpenSSL: some compilers
9204 (gcc 4.2 and later) reject their use.
9205 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9208 * Add RFC4507 support to OpenSSL. This includes the corrections in
9209 RFC4507bis. The encrypted ticket format is an encrypted encoded
9210 SSL_SESSION structure, that way new session features are automatically
9213 If a client application caches session in an SSL_SESSION structure
9214 support is transparent because tickets are now stored in the encoded
9217 The SSL_CTX structure automatically generates keys for ticket
9218 protection in servers so again support should be possible
9219 with no application modification.
9221 If a client or server wishes to disable RFC4507 support then the option
9222 SSL_OP_NO_TICKET can be set.
9224 Add a TLS extension debugging callback to allow the contents of any client
9225 or server extensions to be examined.
9227 This work was sponsored by Google.
9231 * Add initial support for TLS extensions, specifically for the server_name
9232 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9233 have new members for a host name. The SSL data structure has an
9234 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9235 stored in that context to allow for session resumption, even after the
9236 SSL has been switched to a new SSL_CTX in reaction to a client's
9237 server_name extension.
9239 New functions (subject to change):
9241 SSL_get_servername()
9242 SSL_get_servername_type()
9245 New CTRL codes and macros (subject to change):
9247 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9248 - SSL_CTX_set_tlsext_servername_callback()
9249 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9250 - SSL_CTX_set_tlsext_servername_arg()
9251 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9253 openssl s_client has a new '-servername ...' option.
9255 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9256 '-key2 ...', '-servername_fatal' (subject to change). This allows
9257 testing the HostName extension for a specific single host name ('-cert'
9258 and '-key' remain fallbacks for handshakes without HostName
9259 negotiation). If the unrecognized_name alert has to be sent, this by
9260 default is a warning; it becomes fatal with the '-servername_fatal'
9263 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9265 * Add AES and SSE2 assembly language support to VC++ build.
9269 * Mitigate attack on final subtraction in Montgomery reduction.
9273 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9274 (which previously caused an internal error).
9278 * Squeeze another 10% out of IGE mode when in != out.
9282 * AES IGE mode speedup.
9284 *Dean Gaudet (Google)*
9286 * Add the Korean symmetric 128-bit cipher SEED (see
9287 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9288 add SEED ciphersuites from RFC 4162:
9290 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9291 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9292 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9293 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9295 To minimize changes between patchlevels in the OpenSSL 0.9.8
9296 series, SEED remains excluded from compilation unless OpenSSL
9297 is configured with 'enable-seed'.
9299 *KISA, Bodo Moeller*
9301 * Mitigate branch prediction attacks, which can be practical if a
9302 single processor is shared, allowing a spy process to extract
9303 information. For detailed background information, see
9304 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9305 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9306 and Necessary Software Countermeasures"). The core of the change
9307 are new versions BN_div_no_branch() and
9308 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9309 respectively, which are slower, but avoid the security-relevant
9310 conditional branches. These are automatically called by BN_div()
9311 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9312 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9313 remove a conditional branch.
9315 BN_FLG_CONSTTIME is the new name for the previous
9316 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9317 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9318 in the exponent causes BN_mod_exp_mont() to use the alternative
9319 implementation in BN_mod_exp_mont_consttime().) The old name
9320 remains as a deprecated alias.
9322 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9323 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9324 constant-time implementations for more than just exponentiation.
9325 Here too the old name is kept as a deprecated alias.
9327 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9328 the BN_BLINDING structure gets an independent copy of the
9329 modulus. This means that the previous `BIGNUM *m` argument to
9330 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9331 essentially becomes `const BIGNUM *m`, although we can't actually
9332 change this in the header file before 0.9.9. It allows
9333 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9334 enable BN_FLG_CONSTTIME.
9336 *Matthew D Wood (Intel Corp)*
9338 * In the SSL/TLS server implementation, be strict about session ID
9339 context matching (which matters if an application uses a single
9340 external cache for different purposes). Previously,
9341 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9342 set. This did ensure strict client verification, but meant that,
9343 with applications using a single external cache for quite
9344 different requirements, clients could circumvent ciphersuite
9345 restrictions for a given session ID context by starting a session
9346 in a different context.
9350 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9351 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9352 authentication-only ciphersuites.
9356 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9357 not complete and could lead to a possible single byte overflow
9358 [CVE-2007-5135][] [Ben Laurie]
9360 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9362 * Since AES128 and AES256 (and similarly Camellia128 and
9363 Camellia256) share a single mask bit in the logic of
9364 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9365 kludge to work properly if AES128 is available and AES256 isn't
9366 (or if Camellia128 is available and Camellia256 isn't).
9370 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9371 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9372 When a point or a seed is encoded in a BIT STRING, we need to
9373 prevent the removal of trailing zero bits to get the proper DER
9374 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9375 of a NamedBitList, for which trailing 0 bits need to be removed.)
9379 * Have SSL/TLS server implementation tolerate "mismatched" record
9380 protocol version while receiving ClientHello even if the
9381 ClientHello is fragmented. (The server can't insist on the
9382 particular protocol version it has chosen before the ServerHello
9383 message has informed the client about his choice.)
9387 * Add RFC 3779 support.
9389 *Rob Austein for ARIN, Ben Laurie*
9391 * Load error codes if they are not already present instead of using a
9392 static variable. This allows them to be cleanly unloaded and reloaded.
9393 Improve header file function name parsing.
9397 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9398 or CAPABILITY handshake as required by RFCs.
9402 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9404 * Introduce limits to prevent malicious keys being able to
9405 cause a denial of service. [CVE-2006-2940][]
9407 *Steve Henson, Bodo Moeller*
9409 * Fix ASN.1 parsing of certain invalid structures that can result
9410 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9412 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9413 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9415 * Fix SSL client code which could crash if connecting to a
9416 malicious SSLv2 server. [CVE-2006-4343][]
9418 *Tavis Ormandy and Will Drewry, Google Security Team*
9420 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9421 match only those. Before that, "AES256-SHA" would be interpreted
9422 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9423 the same strength classification in 0.9.7h) as we currently only
9424 have a single AES bit in the ciphersuite description bitmap.
9425 That change, however, also applied to ciphersuite strings such as
9426 "RC4-MD5" that intentionally matched multiple ciphersuites --
9427 namely, SSL 2.0 ciphersuites in addition to the more common ones
9428 from SSL 3.0/TLS 1.0.
9430 So we change the selection algorithm again: Naming an explicit
9431 ciphersuite selects this one ciphersuite, and any other similar
9432 ciphersuite (same bitmap) from *other* protocol versions.
9433 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9434 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9436 Since SSL 2.0 does not have any ciphersuites for which the
9437 128/256 bit distinction would be relevant, this works for now.
9438 The proper fix will be to use different bits for AES128 and
9439 AES256, which would have avoided the problems from the beginning;
9440 however, bits are scarce, so we can only do this in a new release
9441 (not just a patchlevel) when we can change the SSL_CIPHER
9442 definition to split the single 'unsigned long mask' bitmap into
9443 multiple values to extend the available space.
9447 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9449 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9450 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9452 * Add AES IGE and biIGE modes.
9456 * Change the Unix randomness entropy gathering to use poll() when
9457 possible instead of select(), since the latter has some
9458 undesirable limitations.
9460 *Darryl Miles via Richard Levitte and Bodo Moeller*
9462 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9463 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9464 cannot be implicitly activated as part of, e.g., the "AES" alias.
9465 However, please upgrade to OpenSSL 0.9.9[-dev] for
9466 non-experimental use of the ECC ciphersuites to get TLS extension
9467 support, which is required for curve and point format negotiation
9468 to avoid potential handshake problems.
9472 * Disable rogue ciphersuites:
9474 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9475 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9476 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9478 The latter two were purportedly from
9479 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9482 Also deactivate the remaining ciphersuites from
9483 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9484 unofficial, and the ID has long expired.
9488 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9489 dual-core machines) and other potential thread-safety issues.
9493 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9494 versions), which is now available for royalty-free use
9495 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9496 Also, add Camellia TLS ciphersuites from RFC 4132.
9498 To minimize changes between patchlevels in the OpenSSL 0.9.8
9499 series, Camellia remains excluded from compilation unless OpenSSL
9500 is configured with 'enable-camellia'.
9504 * Disable the padding bug check when compression is in use. The padding
9505 bug check assumes the first packet is of even length, this is not
9506 necessarily true if compression is enabled and can result in false
9507 positives causing handshake failure. The actual bug test is ancient
9508 code so it is hoped that implementations will either have fixed it by
9509 now or any which still have the bug do not support compression.
9513 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9515 * When applying a cipher rule check to see if string match is an explicit
9516 cipher suite and only match that one cipher suite if it is.
9520 * Link in manifests for VC++ if needed.
9522 *Austin Ziegler <halostatue@gmail.com>*
9524 * Update support for ECC-based TLS ciphersuites according to
9525 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9526 TLS extensions, which are supported starting with the 0.9.9
9527 branch, not in the OpenSSL 0.9.8 branch).
9531 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9532 opaque EVP_CIPHER_CTX handling.
9536 * Fixes and enhancements to zlib compression code. We now only use
9537 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9538 to conform with the standards mentioned here:
9539 <http://www.zlib.net/DLL_FAQ.txt>
9540 Static zlib linking now works on Windows and the new --with-zlib-include
9541 --with-zlib-lib options to Configure can be used to supply the location
9542 of the headers and library. Gracefully handle case where zlib library
9547 * Several fixes and enhancements to the OID generation code. The old code
9548 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9549 handle numbers larger than ULONG_MAX, truncated printing and had a
9550 non standard OBJ_obj2txt() behaviour.
9554 * Add support for building of engines under engine/ as shared libraries
9555 under VC++ build system.
9559 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9560 Hopefully, we will not see any false combination of paths any more.
9564 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9566 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9567 (part of SSL_OP_ALL). This option used to disable the
9568 countermeasure against man-in-the-middle protocol-version
9569 rollback in the SSL 2.0 server implementation, which is a bad
9570 idea. [CVE-2005-2969][]
9572 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9573 for Information Security, National Institute of Advanced Industrial
9574 Science and Technology [AIST], Japan)*
9576 * Add two function to clear and return the verify parameter flags.
9580 * Keep cipherlists sorted in the source instead of sorting them at
9581 runtime, thus removing the need for a lock.
9585 * Avoid some small subgroup attacks in Diffie-Hellman.
9587 *Nick Mathewson and Ben Laurie*
9589 * Add functions for well-known primes.
9593 * Extended Windows CE support.
9595 *Satoshi Nakamura and Andy Polyakov*
9597 * Initialize SSL_METHOD structures at compile time instead of during
9598 runtime, thus removing the need for a lock.
9602 * Make PKCS7_decrypt() work even if no certificate is supplied by
9603 attempting to decrypt each encrypted key in turn. Add support to
9608 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9610 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9613 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9617 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9618 key into the same file any more.
9622 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9626 * Add -utf8 command line and config file option to 'ca'.
9628 *Stefan <stf@udoma.org*
9630 * Removed the macro des_crypt(), as it seems to conflict with some
9631 libraries. Use DES_crypt().
9635 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9636 involves renaming the source and generated shared-libs for
9637 both. The engines will accept the corrected or legacy ids
9638 ('ncipher' and '4758_cca' respectively) when binding. NB,
9639 this only applies when building 'shared'.
9641 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9643 * Add attribute functions to EVP_PKEY structure. Modify
9644 PKCS12_create() to recognize a CSP name attribute and
9645 use it. Make -CSP option work again in pkcs12 utility.
9649 * Add new functionality to the bn blinding code:
9650 - automatic re-creation of the BN_BLINDING parameters after
9651 a fixed number of uses (currently 32)
9652 - add new function for parameter creation
9653 - introduce flags to control the update behaviour of the
9654 BN_BLINDING parameters
9655 - hide BN_BLINDING structure
9656 Add a second BN_BLINDING slot to the RSA structure to improve
9657 performance when a single RSA object is shared among several
9662 * Add support for DTLS.
9664 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9666 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9667 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9671 * Remove buggy and incomplete DH cert support from
9672 ssl/ssl_rsa.c and ssl/s3_both.c
9676 * Use SHA-1 instead of MD5 as the default digest algorithm for
9677 the apps/openssl applications.
9681 * Compile clean with "-Wall -Wmissing-prototypes
9682 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9683 DEBUG_SAFESTACK must also be set.
9687 * Change ./Configure so that certain algorithms can be disabled by default.
9688 The new counterpiece to "no-xxx" is "enable-xxx".
9690 The patented RC5 and MDC2 algorithms will now be disabled unless
9691 "enable-rc5" and "enable-mdc2", respectively, are specified.
9693 (IDEA remains enabled despite being patented. This is because IDEA
9694 is frequently required for interoperability, and there is no license
9695 fee for non-commercial use. As before, "no-idea" can be used to
9696 avoid this algorithm.)
9700 * Add processing of proxy certificates (see RFC 3820). This work was
9701 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9702 EGEE (Enabling Grids for E-science in Europe).
9706 * RC4 performance overhaul on modern architectures/implementations, such
9707 as Intel P4, IA-64 and AMD64.
9711 * New utility extract-section.pl. This can be used specify an alternative
9712 section number in a pod file instead of having to treat each file as
9713 a separate case in Makefile. This can be done by adding two lines to the
9716 =for comment openssl_section:XXX
9718 The blank line is mandatory.
9722 * New arguments -certform, -keyform and -pass for s_client and s_server
9723 to allow alternative format key and certificate files and passphrase
9728 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9729 update associated structures and add various utility functions.
9731 Add new policy related verify parameters, include policy checking in
9732 standard verify code. Enhance 'smime' application with extra parameters
9733 to support policy checking and print out.
9737 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9738 Nehemiah processors. These extensions support AES encryption in hardware
9739 as well as RNG (though RNG support is currently disabled).
9741 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9743 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9747 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9749 *Andy Polyakov and a number of other people*
9751 * Improved PowerPC platform support. Most notably BIGNUM assembler
9752 implementation contributed by IBM.
9754 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9756 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9757 exponent rather than 'unsigned long'. There is a corresponding change to
9758 the new 'rsa_keygen' element of the RSA_METHOD structure.
9760 *Jelte Jansen, Geoff Thorpe*
9762 * Functionality for creating the initial serial number file is now
9763 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9765 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9766 number file to 1, which is bound to cause problems. To avoid
9767 the problems while respecting compatibility between different 0.9.7
9768 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9769 CA.pl for serial number initialization. With the new release 0.9.8,
9770 we can fix the problem directly in the 'ca' utility.)
9774 * Reduced header interdependencies by declaring more opaque objects in
9775 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9776 give fewer recursive includes, which could break lazy source code - so
9777 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9778 developers should define this symbol when building and using openssl to
9779 ensure they track the recommended behaviour, interfaces, [etc], but
9780 backwards-compatible behaviour prevails when this isn't defined.
9784 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9788 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9789 This will generate a random key of the appropriate length based on the
9790 cipher context. The EVP_CIPHER can provide its own random key generation
9791 routine to support keys of a specific form. This is used in the des and
9792 3des routines to generate a key of the correct parity. Update S/MIME
9793 code to use new functions and hence generate correct parity DES keys.
9794 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9795 valid (weak or incorrect parity).
9799 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9800 as looking them up. This is useful when the verified structure may contain
9801 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9802 present unless the new PKCS7_NO_CRL flag is asserted.
9806 * Extend ASN1 oid configuration module. It now additionally accepts the
9809 shortName = some long name, 1.2.3.4
9813 * Reimplemented the BN_CTX implementation. There is now no more static
9814 limitation on the number of variables it can handle nor the depth of the
9815 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9816 information can now expand as required, and rather than having a single
9817 static array of bignums, BN_CTX now uses a linked-list of such arrays
9818 allowing it to expand on demand whilst maintaining the usefulness of
9819 BN_CTX's "bundling".
9823 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9824 to allow all RSA operations to function using a single BN_CTX.
9828 * Preliminary support for certificate policy evaluation and checking. This
9829 is initially intended to pass the tests outlined in "Conformance Testing
9830 of Relying Party Client Certificate Path Processing Logic" v1.07.
9834 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9835 remained unused and not that useful. A variety of other little bignum
9836 tweaks and fixes have also been made continuing on from the audit (see
9841 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9842 associated ASN1, EVP and SSL functions and old ASN1 macros.
9846 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9847 and this should never fail. So the return value from the use of
9848 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9849 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9853 * BN_CTX_get() should return zero-valued bignums, providing the same
9854 initialised value as BN_new().
9856 *Geoff Thorpe, suggested by Ulf Möller*
9858 * Support for inhibitAnyPolicy certificate extension.
9862 * An audit of the BIGNUM code is underway, for which debugging code is
9863 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9864 is considered valid when processing BIGNUMs, and causes execution to
9865 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9866 further steps are taken to deliberately pollute unused data in BIGNUM
9867 structures to try and expose faulty code further on. For now, openssl will
9868 (in its default mode of operation) continue to tolerate the inconsistent
9869 forms that it has tolerated in the past, but authors and packagers should
9870 consider trying openssl and their own applications when compiled with
9871 these debugging symbols defined. It will help highlight potential bugs in
9872 their own code, and will improve the test coverage for OpenSSL itself. At
9873 some point, these tighter rules will become openssl's default to improve
9874 maintainability, though the assert()s and other overheads will remain only
9875 in debugging configurations. See bn.h for more details.
9877 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9879 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9880 that can only be obtained through BN_CTX_new() (which implicitly
9881 initialises it). The presence of this function only made it possible
9882 to overwrite an existing structure (and cause memory leaks).
9886 * Because of the callback-based approach for implementing LHASH as a
9887 template type, lh_insert() adds opaque objects to hash-tables and
9888 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9889 to clean up those corresponding objects before destroying the hash table
9890 (and losing the object pointers). So some over-zealous constifications in
9891 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9892 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9893 prototyped to have "const" restrictions on the object pointers they are
9894 given (and so aren't required to cast them away any more).
9898 * The tmdiff.h API was so ugly and minimal that our own timing utility
9899 (speed) prefers to use its own implementation. The two implementations
9900 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9901 its object type properly exposed (MS_TM) instead of casting to/from
9902 `char *`. This may still change yet if someone realises MS_TM and
9904 aren't necessarily the greatest nomenclatures - but this is what was used
9905 internally to the implementation so I've used that for now.
9909 * Ensure that deprecated functions do not get compiled when
9910 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9911 the self-tests were still using deprecated key-generation functions so
9912 these have been updated also.
9916 * Reorganise PKCS#7 code to separate the digest location functionality
9917 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9918 New function PKCS7_set_digest() to set the digest type for PKCS#7
9919 digestedData type. Add additional code to correctly generate the
9920 digestedData type and add support for this type in PKCS7 initialization
9925 * New function PKCS7_set0_type_other() this initializes a PKCS7
9926 structure of type "other".
9930 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9931 sure the loop does correctly stop and breaking ("division by zero")
9932 modulus operations are not performed. The (pre-generated) prime
9933 table crypto/bn/bn_prime.h was already correct, but it could not be
9934 re-generated on some platforms because of the "division by zero"
9935 situation in the script.
9937 *Ralf S. Engelschall*
9939 * Update support for ECC-based TLS ciphersuites according to
9940 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9941 SHA-1 now is only used for "small" curves (where the
9942 representation of a field element takes up to 24 bytes); for
9943 larger curves, the field element resulting from ECDH is directly
9944 used as premaster secret.
9946 *Douglas Stebila (Sun Microsystems Laboratories)*
9948 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9949 curve secp160r1 to the tests.
9951 *Douglas Stebila (Sun Microsystems Laboratories)*
9953 * Add the possibility to load symbols globally with DSO.
9955 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9957 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9958 control of the error stack.
9962 * Add support for STORE in ENGINE.
9966 * Add the STORE type. The intention is to provide a common interface
9967 to certificate and key stores, be they simple file-based stores, or
9968 HSM-type store, or LDAP stores, or...
9969 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9973 * Add a generic structure called OPENSSL_ITEM. This can be used to
9974 pass a list of arguments to any function as well as provide a way
9975 for a function to pass data back to the caller.
9979 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9980 works like BUF_strdup() but can be used to duplicate a portion of
9981 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9986 * Add the function sk_find_ex() which works like sk_find(), but will
9987 return an index to an element even if an exact match couldn't be
9988 found. The index is guaranteed to point at the element where the
9989 searched-for key would be inserted to preserve sorting order.
9993 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9994 takes an extra flags argument for optional functionality. Currently,
9995 the following flags are defined:
9997 OBJ_BSEARCH_VALUE_ON_NOMATCH
9998 This one gets OBJ_bsearch_ex() to return a pointer to the first
9999 element where the comparing function returns a negative or zero
10002 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10003 This one gets OBJ_bsearch_ex() to return a pointer to the first
10004 element where the comparing function returns zero. This is useful
10005 if there are more than one element where the comparing function
10010 * Make it possible to create self-signed certificates with 'openssl ca'
10011 in such a way that the self-signed certificate becomes part of the
10012 CA database and uses the same mechanisms for serial number generation
10013 as all other certificate signing. The new flag '-selfsign' enables
10014 this functionality. Adapt CA.sh and CA.pl.in.
10018 * Add functionality to check the public key of a certificate request
10019 against a given private. This is useful to check that a certificate
10020 request can be signed by that key (self-signing).
10024 * Make it possible to have multiple active certificates with the same
10025 subject in the CA index file. This is done only if the keyword
10026 'unique_subject' is set to 'no' in the main CA section (default
10027 if 'CA_default') of the configuration file. The value is saved
10028 with the database itself in a separate index attribute file,
10029 named like the index file with '.attr' appended to the name.
10033 * Generate multi-valued AVAs using '+' notation in config files for
10038 * Support for nameConstraints certificate extension.
10042 * Support for policyConstraints certificate extension.
10046 * Support for policyMappings certificate extension.
10050 * Make sure the default DSA_METHOD implementation only uses its
10051 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10052 and change its own handlers to be NULL so as to remove unnecessary
10053 indirection. This lets alternative implementations fallback to the
10054 default implementation more easily.
10058 * Support for directoryName in GeneralName related extensions
10063 * Make it possible to link applications using Makefile.shared.
10064 Make that possible even when linking against static libraries!
10068 * Support for single pass processing for S/MIME signing. This now
10069 means that S/MIME signing can be done from a pipe, in addition
10070 cleartext signing (multipart/signed type) is effectively streaming
10071 and the signed data does not need to be all held in memory.
10073 This is done with a new flag PKCS7_STREAM. When this flag is set
10074 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10075 is done after the data is output (and digests calculated) in
10076 SMIME_write_PKCS7().
10080 * Add full support for -rpath/-R, both in shared libraries and
10081 applications, at least on the platforms where it's known how
10086 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10087 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10088 will now compute a table of multiples of the generator that
10089 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10090 faster (notably in the case of a single point multiplication,
10091 scalar * generator).
10093 *Nils Larsch, Bodo Moeller*
10095 * IPv6 support for certificate extensions. The various extensions
10096 which use the IP:a.b.c.d can now take IPv6 addresses using the
10097 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10102 * Added an ENGINE that implements RSA by performing private key
10103 exponentiations with the GMP library. The conversions to and from
10104 GMP's mpz_t format aren't optimised nor are any montgomery forms
10105 cached, and on x86 it appears OpenSSL's own performance has caught up.
10106 However there are likely to be other architectures where GMP could
10107 provide a boost. This ENGINE is not built in by default, but it can be
10108 specified at Configure time and should be accompanied by the necessary
10109 linker additions, eg;
10110 ./config -DOPENSSL_USE_GMP -lgmp
10114 * "openssl engine" will not display ENGINE/DSO load failure errors when
10115 testing availability of engines with "-t" - the old behaviour is
10116 produced by increasing the feature's verbosity with "-tt".
10120 * ECDSA routines: under certain error conditions uninitialized BN objects
10121 could be freed. Solution: make sure initialization is performed early
10122 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10127 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10128 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10129 software implementations. For DSA and DH, parameter generation can
10130 also be overridden by providing the appropriate method callbacks.
10134 * Change the "progress" mechanism used in key-generation and
10135 primality testing to functions that take a new BN_GENCB pointer in
10136 place of callback/argument pairs. The new API functions have `_ex`
10137 postfixes and the older functions are reimplemented as wrappers for
10138 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10139 declarations of the old functions to help (graceful) attempts to
10140 migrate to the new functions. Also, the new key-generation API
10141 functions operate on a caller-supplied key-structure and return
10142 success/failure rather than returning a key or NULL - this is to
10143 help make "keygen" another member function of RSA_METHOD etc.
10145 Example for using the new callback interface:
10147 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10148 void *my_arg = ...;
10151 BN_GENCB_set(&my_cb, my_callback, my_arg);
10153 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10154 /* For the meaning of a, b in calls to my_callback(), see the
10155 * documentation of the function that calls the callback.
10156 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10157 * my_callback should return 1 if it wants BN_is_prime_ex()
10158 * to continue, or 0 to stop.
10163 * Change the ZLIB compression method to be stateful, and make it
10164 available to TLS with the number defined in
10165 draft-ietf-tls-compression-04.txt.
10169 * Add the ASN.1 structures and functions for CertificatePair, which
10170 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10172 CertificatePair ::= SEQUENCE {
10173 forward [0] Certificate OPTIONAL,
10174 reverse [1] Certificate OPTIONAL,
10175 -- at least one of the pair shall be present -- }
10177 Also implement the PEM functions to read and write certificate
10178 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10180 This needed to be defined, mostly for the sake of the LDAP
10181 attribute crossCertificatePair, but may prove useful elsewhere as
10186 * Make it possible to inhibit symlinking of shared libraries in
10187 Makefile.shared, for Cygwin's sake.
10191 * Extend the BIGNUM API by creating a function
10192 void BN_set_negative(BIGNUM *a, int neg);
10193 and a macro that behave like
10194 int BN_is_negative(const BIGNUM *a);
10196 to avoid the need to access 'a->neg' directly in applications.
10200 * Implement fast modular reduction for pseudo-Mersenne primes
10201 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10202 EC_GROUP_new_curve_GFp() will now automatically use this
10205 *Nils Larsch <nla@trustcenter.de>*
10207 * Add new lock type (CRYPTO_LOCK_BN).
10211 * Change the ENGINE framework to automatically load engines
10212 dynamically from specific directories unless they could be
10213 found to already be built in or loaded. Move all the
10214 current engines except for the cryptodev one to a new
10215 directory engines/.
10216 The engines in engines/ are built as shared libraries if
10217 the "shared" options was given to ./Configure or ./config.
10218 Otherwise, they are inserted in libcrypto.a.
10219 /usr/local/ssl/engines is the default directory for dynamic
10220 engines, but that can be overridden at configure time through
10221 the usual use of --prefix and/or --openssldir, and at run
10222 time with the environment variable OPENSSL_ENGINES.
10224 *Geoff Thorpe and Richard Levitte*
10226 * Add Makefile.shared, a helper makefile to build shared
10227 libraries. Adapt Makefile.org.
10231 * Add version info to Win32 DLLs.
10233 *Peter 'Luna' Runestig" <peter@runestig.com>*
10235 * Add new 'medium level' PKCS#12 API. Certificates and keys
10236 can be added using this API to created arbitrary PKCS#12
10237 files while avoiding the low level API.
10239 New options to PKCS12_create(), key or cert can be NULL and
10240 will then be omitted from the output file. The encryption
10241 algorithm NIDs can be set to -1 for no encryption, the mac
10242 iteration count can be set to 0 to omit the mac.
10244 Enhance pkcs12 utility by making the -nokeys and -nocerts
10245 options work when creating a PKCS#12 file. New option -nomac
10246 to omit the mac, NONE can be set for an encryption algorithm.
10247 New code is modified to use the enhanced PKCS12_create()
10248 instead of the low level API.
10252 * Extend ASN1 encoder to support indefinite length constructed
10253 encoding. This can output sequences tags and octet strings in
10254 this form. Modify pk7_asn1.c to support indefinite length
10255 encoding. This is experimental and needs additional code to
10256 be useful, such as an ASN1 bio and some enhanced streaming
10259 Extend template encode functionality so that tagging is passed
10260 down to the template encoder.
10264 * Let 'openssl req' fail if an argument to '-newkey' is not
10265 recognized instead of using RSA as a default.
10269 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10270 As these are not official, they are not included in "ALL";
10271 the "ECCdraft" ciphersuite group alias can be used to select them.
10273 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10275 * Add ECDH engine support.
10277 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10279 * Add ECDH in new directory crypto/ecdh/.
10281 *Douglas Stebila (Sun Microsystems Laboratories)*
10283 * Let BN_rand_range() abort with an error after 100 iterations
10284 without success (which indicates a broken PRNG).
10288 * Change BN_mod_sqrt() so that it verifies that the input value
10289 is really the square of the return value. (Previously,
10290 BN_mod_sqrt would show GIGO behaviour.)
10294 * Add named elliptic curves over binary fields from X9.62, SECG,
10295 and WAP/WTLS; add OIDs that were still missing.
10297 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10299 * Extend the EC library for elliptic curves over binary fields
10300 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10303 EC_GF2m_simple_method
10307 EC_GROUP_new_curve_GF2m
10308 EC_GROUP_set_curve_GF2m
10309 EC_GROUP_get_curve_GF2m
10310 EC_POINT_set_affine_coordinates_GF2m
10311 EC_POINT_get_affine_coordinates_GF2m
10312 EC_POINT_set_compressed_coordinates_GF2m
10314 Point compression for binary fields is disabled by default for
10315 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10318 As binary polynomials are represented as BIGNUMs, various members
10319 of the EC_GROUP and EC_POINT data structures can be shared
10320 between the implementations for prime fields and binary fields;
10321 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10322 are essentially identical to their `..._GFp` counterparts.
10323 (For simplicity, the `..._GFp` prefix has been dropped from
10324 various internal method names.)
10326 An internal 'field_div' method (similar to 'field_mul' and
10327 'field_sqr') has been added; this is used only for binary fields.
10329 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10331 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10332 through methods ('mul', 'precompute_mult').
10334 The generic implementations (now internally called 'ec_wNAF_mul'
10335 and 'ec_wNAF_precomputed_mult') remain the default if these
10336 methods are undefined.
10338 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10340 * New function EC_GROUP_get_degree, which is defined through
10341 EC_METHOD. For curves over prime fields, this returns the bit
10342 length of the modulus.
10344 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10346 * New functions EC_GROUP_dup, EC_POINT_dup.
10347 (These simply call ..._new and ..._copy).
10349 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10351 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10352 Polynomials are represented as BIGNUMs (where the sign bit is not
10353 used) in the following functions [macros]:
10356 BN_GF2m_sub [= BN_GF2m_add]
10357 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10358 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10359 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10361 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10362 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10363 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10364 BN_GF2m_cmp [= BN_ucmp]
10366 (Note that only the 'mod' functions are actually for fields GF(2^m).
10367 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10369 For some functions, an the irreducible polynomial defining a
10370 field can be given as an 'unsigned int[]' with strictly
10371 decreasing elements giving the indices of those bits that are set;
10372 i.e., p[] represents the polynomial
10373 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10375 p[0] > p[1] > ... > p[k] = 0.
10376 This applies to the following functions:
10379 BN_GF2m_mod_mul_arr
10380 BN_GF2m_mod_sqr_arr
10381 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10382 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10383 BN_GF2m_mod_exp_arr
10384 BN_GF2m_mod_sqrt_arr
10385 BN_GF2m_mod_solve_quad_arr
10389 Conversion can be performed by the following functions:
10394 bntest.c has additional tests for binary polynomial arithmetic.
10396 Two implementations for BN_GF2m_mod_div() are available.
10397 The default algorithm simply uses BN_GF2m_mod_inv() and
10398 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10399 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10400 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10402 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10404 * Add new error code 'ERR_R_DISABLED' that can be used when some
10405 functionality is disabled at compile-time.
10407 *Douglas Stebila <douglas.stebila@sun.com>*
10409 * Change default behaviour of 'openssl asn1parse' so that more
10410 information is visible when viewing, e.g., a certificate:
10412 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10413 mode the content of non-printable OCTET STRINGs is output in a
10414 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10415 avoid the appearance of a printable string.
10417 *Nils Larsch <nla@trustcenter.de>*
10419 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10421 EC_GROUP_set_asn1_flag()
10422 EC_GROUP_get_asn1_flag()
10423 EC_GROUP_set_point_conversion_form()
10424 EC_GROUP_get_point_conversion_form()
10425 These control ASN1 encoding details:
10426 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10427 has been set to OPENSSL_EC_NAMED_CURVE.
10428 - Points are encoded in uncompressed form by default; options for
10429 asn1_for are as for point2oct, namely
10430 POINT_CONVERSION_COMPRESSED
10431 POINT_CONVERSION_UNCOMPRESSED
10432 POINT_CONVERSION_HYBRID
10434 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10436 EC_GROUP_set_seed()
10437 EC_GROUP_get0_seed()
10438 EC_GROUP_get_seed_len()
10439 This is used only for ASN1 purposes (so far).
10441 *Nils Larsch <nla@trustcenter.de>*
10443 * Add 'field_type' member to EC_METHOD, which holds the NID
10444 of the appropriate field type OID. The new function
10445 EC_METHOD_get_field_type() returns this value.
10447 *Nils Larsch <nla@trustcenter.de>*
10450 EC_POINT_point2bn()
10451 EC_POINT_bn2point()
10452 EC_POINT_point2hex()
10453 EC_POINT_hex2point()
10454 providing useful interfaces to EC_POINT_point2oct() and
10455 EC_POINT_oct2point().
10457 *Nils Larsch <nla@trustcenter.de>*
10459 * Change internals of the EC library so that the functions
10460 EC_GROUP_set_generator()
10461 EC_GROUP_get_generator()
10462 EC_GROUP_get_order()
10463 EC_GROUP_get_cofactor()
10464 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10465 to methods, which would lead to unnecessary code duplication when
10466 adding different types of curves.
10468 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10470 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10471 arithmetic, and such that modified wNAFs are generated
10472 (which avoid length expansion in many cases).
10476 * Add a function EC_GROUP_check_discriminant() (defined via
10477 EC_METHOD) that verifies that the curve discriminant is non-zero.
10479 Add a function EC_GROUP_check() that makes some sanity tests
10480 on a EC_GROUP, its generator and order. This includes
10481 EC_GROUP_check_discriminant().
10483 *Nils Larsch <nla@trustcenter.de>*
10485 * Add ECDSA in new directory crypto/ecdsa/.
10487 Add applications 'openssl ecparam' and 'openssl ecdsa'
10488 (these are based on 'openssl dsaparam' and 'openssl dsa').
10490 ECDSA support is also included in various other files across the
10491 library. Most notably,
10492 - 'openssl req' now has a '-newkey ecdsa:file' option;
10493 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10494 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10495 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10496 them suitable for ECDSA where domain parameters must be
10497 extracted before the specific public key;
10498 - ECDSA engine support has been added.
10500 *Nils Larsch <nla@trustcenter.de>*
10502 * Include some named elliptic curves, and add OIDs from X9.62,
10503 SECG, and WAP/WTLS. Each curve can be obtained from the new
10505 EC_GROUP_new_by_curve_name(),
10506 and the list of available named curves can be obtained with
10507 EC_get_builtin_curves().
10508 Also add a 'curve_name' member to EC_GROUP objects, which can be
10510 EC_GROUP_set_curve_name()
10511 EC_GROUP_get_curve_name()
10513 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10515 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10516 was actually never needed) and in BN_mul(). The removal in BN_mul()
10517 required a small change in bn_mul_part_recursive() and the addition
10518 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10519 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10520 bn_sub_words() and bn_add_words() except they take arrays with
10525 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10527 * Cleanse PEM buffers before freeing them since they may contain
10530 *Benjamin Bennett <ben@psc.edu>*
10532 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10533 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10534 authentication-only ciphersuites.
10538 * Since AES128 and AES256 share a single mask bit in the logic of
10539 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10540 kludge to work properly if AES128 is available and AES256 isn't.
10544 * Expand security boundary to match 1.1.1 module.
10548 * Remove redundant features: hash file source, editing of test vectors
10549 modify fipsld to use external fips_premain.c signature.
10553 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10554 run algorithm test programs.
10558 * Make algorithm test programs more tolerant of whitespace.
10562 * Have SSL/TLS server implementation tolerate "mismatched" record
10563 protocol version while receiving ClientHello even if the
10564 ClientHello is fragmented. (The server can't insist on the
10565 particular protocol version it has chosen before the ServerHello
10566 message has informed the client about his choice.)
10570 * Load error codes if they are not already present instead of using a
10571 static variable. This allows them to be cleanly unloaded and reloaded.
10575 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10577 * Introduce limits to prevent malicious keys being able to
10578 cause a denial of service. [CVE-2006-2940][]
10580 *Steve Henson, Bodo Moeller*
10582 * Fix ASN.1 parsing of certain invalid structures that can result
10583 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10585 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10586 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10588 * Fix SSL client code which could crash if connecting to a
10589 malicious SSLv2 server. [CVE-2006-4343][]
10591 *Tavis Ormandy and Will Drewry, Google Security Team*
10593 * Change ciphersuite string processing so that an explicit
10594 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10595 will no longer include "AES128-SHA"), and any other similar
10596 ciphersuite (same bitmap) from *other* protocol versions (so that
10597 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10598 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10599 changes from 0.9.8b and 0.9.8d.
10603 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10605 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10606 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10608 * Change the Unix randomness entropy gathering to use poll() when
10609 possible instead of select(), since the latter has some
10610 undesirable limitations.
10612 *Darryl Miles via Richard Levitte and Bodo Moeller*
10614 * Disable rogue ciphersuites:
10616 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10617 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10618 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10620 The latter two were purportedly from
10621 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10624 Also deactivate the remaining ciphersuites from
10625 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10626 unofficial, and the ID has long expired.
10630 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10631 dual-core machines) and other potential thread-safety issues.
10635 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10637 * Adapt fipsld and the build system to link against the validated FIPS
10638 module in FIPS mode.
10642 * Fixes for VC++ 2005 build under Windows.
10646 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10647 from a Windows bash shell such as MSYS. It is autodetected from the
10648 "config" script when run from a VC++ environment. Modify standard VC++
10649 build to use fipscanister.o from the GNU make build.
10653 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10655 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10656 The value now differs depending on if you build for FIPS or not.
10657 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10658 safely run with a non-FIPSed libcrypto, as it may crash because of
10659 the difference induced by this change.
10663 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10665 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10666 (part of SSL_OP_ALL). This option used to disable the
10667 countermeasure against man-in-the-middle protocol-version
10668 rollback in the SSL 2.0 server implementation, which is a bad
10669 idea. [CVE-2005-2969][]
10671 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10672 for Information Security, National Institute of Advanced Industrial
10673 Science and Technology [AIST, Japan)]*
10675 * Minimal support for X9.31 signatures and PSS padding modes. This is
10676 mainly for FIPS compliance and not fully integrated at this stage.
10680 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10681 the exponentiation using a fixed-length exponent. (Otherwise,
10682 the information leaked through timing could expose the secret key
10683 after many signatures; cf. Bleichenbacher's attack on DSA with
10688 * Make a new fixed-window mod_exp implementation the default for
10689 RSA, DSA, and DH private-key operations so that the sequence of
10690 squares and multiplies and the memory access pattern are
10691 independent of the particular secret key. This will mitigate
10692 cache-timing and potential related attacks.
10694 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10695 and this is automatically used by BN_mod_exp_mont() if the new flag
10696 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10697 will use this BN flag for private exponents unless the flag
10698 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10699 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10701 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10703 * Change the client implementation for SSLv23_method() and
10704 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10705 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10706 (Previously, the SSL 2.0 backwards compatible Client Hello
10707 message format would be used even with SSL_OP_NO_SSLv2.)
10711 * Add support for smime-type MIME parameter in S/MIME messages which some
10716 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10717 a threadsafe manner. Modify rsa code to use new function and add calls
10718 to dsa and dh code (which had race conditions before).
10722 * Include the fixed error library code in the C error file definitions
10723 instead of fixing them up at runtime. This keeps the error code
10724 structures constant.
10728 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10730 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10733 * Fixes for newer kerberos headers. NB: the casts are needed because
10734 the 'length' field is signed on one version and unsigned on another
10735 with no (?) obvious way to tell the difference, without these VC++
10736 complains. Also the "definition" of FAR (blank) is no longer included
10737 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10738 some needed definitions.
10742 * Undo Cygwin change.
10746 * Added support for proxy certificates according to RFC 3820.
10747 Because they may be a security thread to unaware applications,
10748 they must be explicitly allowed in run-time. See
10749 docs/HOWTO/proxy_certificates.txt for further information.
10753 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10755 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10756 server and client random values. Previously
10757 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10758 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10760 This change has negligible security impact because:
10762 1. Server and client random values still have 24 bytes of pseudo random
10765 2. Server and client random values are sent in the clear in the initial
10768 3. The master secret is derived using the premaster secret (48 bytes in
10769 size for static RSA ciphersuites) as well as client server and random
10772 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10775 *Stephen Henson, reported by UK NISCC*
10777 * Use Windows randomness collection on Cygwin.
10781 * Fix hang in EGD/PRNGD query when communication socket is closed
10782 prematurely by EGD/PRNGD.
10784 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10786 * Prompt for pass phrases when appropriate for PKCS12 input format.
10790 * Back-port of selected performance improvements from development
10791 branch, as well as improved support for PowerPC platforms.
10795 * Add lots of checks for memory allocation failure, error codes to indicate
10796 failure and freeing up memory if a failure occurs.
10798 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10800 * Add new -passin argument to dgst.
10804 * Perform some character comparisons of different types in X509_NAME_cmp:
10805 this is needed for some certificates that re-encode DNs into UTF8Strings
10806 (in violation of RFC3280) and can't or won't issue name rollover
10811 * Make an explicit check during certificate validation to see that
10812 the CA setting in each certificate on the chain is correct. As a
10813 side effect always do the following basic checks on extensions,
10814 not just when there's an associated purpose to the check:
10816 - if there is an unhandled critical extension (unless the user
10817 has chosen to ignore this fault)
10818 - if the path length has been exceeded (if one is set at all)
10819 - that certain extensions fit the associated purpose (if one has
10824 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10826 * Avoid a race condition when CRLs are checked in a multi threaded
10827 environment. This would happen due to the reordering of the revoked
10828 entries during signature checking and serial number lookup. Now the
10829 encoding is cached and the serial number sort performed under a lock.
10830 Add new STACK function sk_is_sorted().
10834 * Add Delta CRL to the extension code.
10838 * Various fixes to s3_pkt.c so alerts are sent properly.
10840 *David Holmes <d.holmes@f5.com>*
10842 * Reduce the chances of duplicate issuer name and serial numbers (in
10843 violation of RFC3280) using the OpenSSL certificate creation utilities.
10844 This is done by creating a random 64 bit value for the initial serial
10845 number when a serial number file is created or when a self signed
10846 certificate is created using 'openssl req -x509'. The initial serial
10847 number file is created using 'openssl x509 -next_serial' in CA.pl
10848 rather than being initialized to 1.
10852 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10854 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10855 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10857 *Joe Orton, Steve Henson*
10859 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10862 *Joe Orton, Steve Henson*
10864 * Make it possible to have multiple active certificates with the same
10865 subject in the CA index file. This is done only if the keyword
10866 'unique_subject' is set to 'no' in the main CA section (default
10867 if 'CA_default') of the configuration file. The value is saved
10868 with the database itself in a separate index attribute file,
10869 named like the index file with '.attr' appended to the name.
10873 * X509 verify fixes. Disable broken certificate workarounds when
10874 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10875 keyUsage extension present. Don't accept CRLs with unhandled critical
10876 extensions: since verify currently doesn't process CRL extensions this
10877 rejects a CRL with *any* critical extensions. Add new verify error codes
10882 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10883 A clarification of RFC2560 will require the use of OCTET STRINGs and
10884 some implementations cannot handle the current raw format. Since OpenSSL
10885 copies and compares OCSP nonces as opaque blobs without any attempt at
10886 parsing them this should not create any compatibility issues.
10890 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10891 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10892 this HMAC (and other) operations are several times slower than OpenSSL
10897 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10899 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10901 * Use the correct content when signing type "other".
10905 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10907 * Fix various bugs revealed by running the NISCC test suite:
10909 Stop out of bounds reads in the ASN1 code when presented with
10910 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10912 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10914 If verify callback ignores invalid public key errors don't try to check
10915 certificate signature with the NULL public key.
10919 * New -ignore_err option in ocsp application to stop the server
10920 exiting on the first error in a request.
10924 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10925 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10930 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10931 extra data after the compression methods not only for TLS 1.0
10932 but also for SSL 3.0 (as required by the specification).
10934 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10936 * Change X509_certificate_type() to mark the key as exported/exportable
10937 when it's 512 *bits* long, not 512 bytes.
10941 * Change AES_cbc_encrypt() so it outputs exact multiple of
10942 blocks during encryption.
10946 * Various fixes to base64 BIO and non blocking I/O. On write
10947 flushes were not handled properly if the BIO retried. On read
10948 data was not being buffered properly and had various logic bugs.
10949 This also affects blocking I/O when the data being decoded is a
10954 * Various S/MIME bugfixes and compatibility changes:
10955 output correct application/pkcs7 MIME type if
10956 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10957 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10958 of files as .eml work). Correctly handle very long lines in MIME
10963 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10965 * Countermeasure against the Klima-Pokorny-Rosa extension of
10966 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10967 a protocol version number mismatch like a decryption error
10968 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10972 * Turn on RSA blinding by default in the default implementation
10973 to avoid a timing attack. Applications that don't want it can call
10974 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10975 They would be ill-advised to do so in most cases.
10977 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10979 * Change RSA blinding code so that it works when the PRNG is not
10980 seeded (in this case, the secret RSA exponent is abused as
10981 an unpredictable seed -- if it is not unpredictable, there
10982 is no point in blinding anyway). Make RSA blinding thread-safe
10983 by remembering the creator's thread ID in rsa->blinding and
10984 having all other threads use local one-time blinding factors
10985 (this requires more computation than sharing rsa->blinding, but
10986 avoids excessive locking; and if an RSA object is not shared
10987 between threads, blinding will still be very fast).
10991 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10992 ENGINE as defaults for all supported algorithms irrespective of
10993 the 'flags' parameter. 'flags' is now honoured, so applications
10994 should make sure they are passing it correctly.
10998 * Target "mingw" now allows native Windows code to be generated in
10999 the Cygwin environment as well as with the MinGW compiler.
11003 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11005 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11006 via timing by performing a MAC computation even if incorrect
11007 block cipher padding has been found. This is a countermeasure
11008 against active attacks where the attacker has to distinguish
11009 between bad padding and a MAC verification error. [CVE-2003-0078][]
11011 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11012 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11013 Martin Vuagnoux (EPFL, Ilion)*
11015 * Make the no-err option work as intended. The intention with no-err
11016 is not to have the whole error stack handling routines removed from
11017 libcrypto, it's only intended to remove all the function name and
11018 reason texts, thereby removing some of the footprint that may not
11019 be interesting if those errors aren't displayed anyway.
11021 NOTE: it's still possible for any application or module to have its
11022 own set of error texts inserted. The routines are there, just not
11023 used by default when no-err is given.
11027 * Add support for FreeBSD on IA64.
11029 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11031 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11032 Kerberos function mit_des_cbc_cksum(). Before this change,
11033 the value returned by DES_cbc_cksum() was like the one from
11034 mit_des_cbc_cksum(), except the bytes were swapped.
11036 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11038 * Allow an application to disable the automatic SSL chain building.
11039 Before this a rather primitive chain build was always performed in
11040 ssl3_output_cert_chain(): an application had no way to send the
11041 correct chain if the automatic operation produced an incorrect result.
11043 Now the chain builder is disabled if either:
11045 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11047 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11049 The reasoning behind this is that an application would not want the
11050 auto chain building to take place if extra chain certificates are
11051 present and it might also want a means of sending no additional
11052 certificates (for example the chain has two certificates and the
11057 * Add the possibility to build without the ENGINE framework.
11059 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11061 * Under Win32 gmtime() can return NULL: check return value in
11062 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11066 * DSA routines: under certain error conditions uninitialized BN objects
11067 could be freed. Solution: make sure initialization is performed early
11068 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11069 Nils Larsch <nla@trustcenter.de> via PR#459)
11073 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11074 checked on reconnect on the client side, therefore session resumption
11075 could still fail with a "ssl session id is different" error. This
11076 behaviour is masked when SSL_OP_ALL is used due to
11077 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11078 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11079 followup to PR #377.
11083 * IA-32 assembler support enhancements: unified ELF targets, support
11084 for SCO/Caldera platforms, fix for Cygwin shared build.
11088 * Add support for FreeBSD on sparc64. As a consequence, support for
11089 FreeBSD on non-x86 processors is separate from x86 processors on
11090 the config script, much like the NetBSD support.
11092 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11094 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11096 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11099 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11100 code (06) was taken as the first octet of the session ID and the last
11101 octet was ignored consequently. As a result SSLv2 client side session
11102 caching could not have worked due to the session ID mismatch between
11104 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11109 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11110 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11115 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11116 seems that in spite of existing for more than a year, many application
11117 author have done nothing to provide the necessary callbacks, which
11118 means that this particular engine will not work properly anywhere.
11119 This is a very unfortunate situation which forces us, in the name
11120 of usability, to give the hw_ncipher.c a static lock, which is part
11122 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11123 appear in 0.9.8 or later. We EXPECT application authors to have
11124 dealt properly with this when 0.9.8 is released (unless we actually
11125 make such changes in the libcrypto locking code that changes will
11126 have to be made anyway).
11130 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11131 octets have been read, EOF or an error occurs. Without this change
11132 some truncated ASN1 structures will not produce an error.
11136 * Disable Heimdal support, since it hasn't been fully implemented.
11137 Still give the possibility to force the use of Heimdal, but with
11138 warnings and a request that patches get sent to openssl-dev.
11142 * Add the VC-CE target, introduce the WINCE sysname, and add
11143 INSTALL.WCE and appropriate conditionals to make it build.
11145 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11147 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11148 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11149 edit numbers of the version.
11151 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11153 * Introduce safe string copy and catenation functions
11154 (BUF_strlcpy() and BUF_strlcat()).
11156 *Ben Laurie (CHATS) and Richard Levitte*
11158 * Avoid using fixed-size buffers for one-line DNs.
11160 *Ben Laurie (CHATS)*
11162 * Add BUF_MEM_grow_clean() to avoid information leakage when
11163 resizing buffers containing secrets, and use where appropriate.
11165 *Ben Laurie (CHATS)*
11167 * Avoid using fixed size buffers for configuration file location.
11169 *Ben Laurie (CHATS)*
11171 * Avoid filename truncation for various CA files.
11173 *Ben Laurie (CHATS)*
11175 * Use sizeof in preference to magic numbers.
11177 *Ben Laurie (CHATS)*
11179 * Avoid filename truncation in cert requests.
11181 *Ben Laurie (CHATS)*
11183 * Add assertions to check for (supposedly impossible) buffer
11186 *Ben Laurie (CHATS)*
11188 * Don't cache truncated DNS entries in the local cache (this could
11189 potentially lead to a spoofing attack).
11191 *Ben Laurie (CHATS)*
11193 * Fix various buffers to be large enough for hex/decimal
11194 representations in a platform independent manner.
11196 *Ben Laurie (CHATS)*
11198 * Add CRYPTO_realloc_clean() to avoid information leakage when
11199 resizing buffers containing secrets, and use where appropriate.
11201 *Ben Laurie (CHATS)*
11203 * Add BIO_indent() to avoid much slightly worrying code to do
11206 *Ben Laurie (CHATS)*
11208 * Convert sprintf()/BIO_puts() to BIO_printf().
11210 *Ben Laurie (CHATS)*
11212 * buffer_gets() could terminate with the buffer only half
11215 *Ben Laurie (CHATS)*
11217 * Add assertions to prevent user-supplied crypto functions from
11218 overflowing internal buffers by having large block sizes, etc.
11220 *Ben Laurie (CHATS)*
11222 * New OPENSSL_assert() macro (similar to assert(), but enabled
11225 *Ben Laurie (CHATS)*
11227 * Eliminate unused copy of key in RC4.
11229 *Ben Laurie (CHATS)*
11231 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11233 *Ben Laurie (CHATS)*
11235 * Fix off-by-one error in EGD path.
11237 *Ben Laurie (CHATS)*
11239 * If RANDFILE path is too long, ignore instead of truncating.
11241 *Ben Laurie (CHATS)*
11243 * Eliminate unused and incorrectly sized X.509 structure
11246 *Ben Laurie (CHATS)*
11248 * Eliminate unused and dangerous function knumber().
11250 *Ben Laurie (CHATS)*
11252 * Eliminate unused and dangerous structure, KSSL_ERR.
11254 *Ben Laurie (CHATS)*
11256 * Protect against overlong session ID context length in an encoded
11257 session object. Since these are local, this does not appear to be
11260 *Ben Laurie (CHATS)*
11262 * Change from security patch (see 0.9.6e below) that did not affect
11263 the 0.9.6 release series:
11265 Remote buffer overflow in SSL3 protocol - an attacker could
11266 supply an oversized master key in Kerberos-enabled versions.
11269 *Ben Laurie (CHATS)*
11271 * Change the SSL kerb5 codes to match RFC 2712.
11275 * Make -nameopt work fully for req and add -reqopt switch.
11277 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11279 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11281 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11283 * Make sure tests can be performed even if the corresponding algorithms
11284 have been removed entirely. This was also the last step to make
11285 OpenSSL compilable with DJGPP under all reasonable conditions.
11287 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11289 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11290 to allow version independent disabling of normally unselected ciphers,
11291 which may be activated as a side-effect of selecting a single cipher.
11293 (E.g., cipher list string "RSA" enables ciphersuites that are left
11294 out of "ALL" because they do not provide symmetric encryption.
11295 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11297 *Lutz Jaenicke, Bodo Moeller*
11299 * Add appropriate support for separate platform-dependent build
11300 directories. The recommended way to make a platform-dependent
11301 build directory is the following (tested on Linux), maybe with
11304 # Place yourself outside of the OpenSSL source tree. In
11305 # this example, the environment variable OPENSSL_SOURCE
11306 # is assumed to contain the absolute OpenSSL source directory.
11307 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11308 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11309 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11310 mkdir -p `dirname $F`
11311 ln -s $OPENSSL_SOURCE/$F $F
11314 To be absolutely sure not to disturb the source tree, a "make clean"
11315 is a good thing. If it isn't successful, don't worry about it,
11316 it probably means the source directory is very clean.
11320 * Make sure any ENGINE control commands make local copies of string
11321 pointers passed to them whenever necessary. Otherwise it is possible
11322 the caller may have overwritten (or deallocated) the original string
11323 data when a later ENGINE operation tries to use the stored values.
11325 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11327 * Improve diagnostics in file reading and command-line digests.
11329 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11331 * Add AES modes CFB and OFB to the object database. Correct an
11332 error in AES-CFB decryption.
11336 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11337 allows existing EVP_CIPHER_CTX structures to be reused after
11338 calling `EVP_*Final()`. This behaviour is used by encryption
11339 BIOs and some applications. This has the side effect that
11340 applications must explicitly clean up cipher contexts with
11341 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11345 * Check the values of dna and dnb in bn_mul_recursive before calling
11346 bn_mul_comba (a non zero value means the a or b arrays do not contain
11347 n2 elements) and fallback to bn_mul_normal if either is not zero.
11351 * Fix escaping of non-ASCII characters when using the -subj option
11352 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11356 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11357 form for "surname", serialNumber has no short form.
11358 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11359 therefore remove "mail" short name for "internet 7".
11360 The OID for unique identifiers in X509 certificates is
11361 x500UniqueIdentifier, not uniqueIdentifier.
11362 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11366 * Add an "init" command to the ENGINE config module and auto initialize
11367 ENGINEs. Without any "init" command the ENGINE will be initialized
11368 after all ctrl commands have been executed on it. If init=1 the
11369 ENGINE is initialized at that point (ctrls before that point are run
11370 on the uninitialized ENGINE and after on the initialized one). If
11371 init=0 then the ENGINE will not be initialized at all.
11375 * Fix the 'app_verify_callback' interface so that the user-defined
11376 argument is actually passed to the callback: In the
11377 SSL_CTX_set_cert_verify_callback() prototype, the callback
11378 declaration has been changed from
11381 int (*cb)(X509_STORE_CTX *,void *);
11382 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11383 i=s->ctx->app_verify_callback(&ctx)
11384 has been changed into
11385 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11387 To update applications using SSL_CTX_set_cert_verify_callback(),
11388 a dummy argument can be added to their callback functions.
11390 *D. K. Smetters <smetters@parc.xerox.com>*
11392 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11394 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11396 * Add and OPENSSL_LOAD_CONF define which will cause
11397 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11398 This allows older applications to transparently support certain
11399 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11400 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11401 load the config file and OPENSSL_add_all_algorithms_conf() which will
11402 always load it have also been added.
11406 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11407 Adjust NIDs and EVP layer.
11409 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11411 * Config modules support in openssl utility.
11413 Most commands now load modules from the config file,
11414 though in a few (such as version) this isn't done
11415 because it couldn't be used for anything.
11417 In the case of ca and req the config file used is
11418 the same as the utility itself: that is the -config
11419 command line option can be used to specify an
11424 * Move default behaviour from OPENSSL_config(). If appname is NULL
11425 use "openssl_conf" if filename is NULL use default openssl config file.
11429 * Add an argument to OPENSSL_config() to allow the use of an alternative
11430 config section name. Add a new flag to tolerate a missing config file
11431 and move code to CONF_modules_load_file().
11435 * Support for crypto accelerator cards from Accelerated Encryption
11436 Processing, www.aep.ie. (Use engine 'aep')
11437 The support was copied from 0.9.6c [engine] and adapted/corrected
11438 to work with the new engine framework.
11440 *AEP Inc. and Richard Levitte*
11442 * Support for SureWare crypto accelerator cards from Baltimore
11443 Technologies. (Use engine 'sureware')
11444 The support was copied from 0.9.6c [engine] and adapted
11445 to work with the new engine framework.
11449 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11450 make the newer ENGINE framework commands for the CHIL engine work.
11452 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11454 * Make it possible to produce shared libraries on ReliantUNIX.
11456 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11458 * Add the configuration target debug-linux-ppro.
11459 Make 'openssl rsa' use the general key loading routines
11460 implemented in apps.c, and make those routines able to
11461 handle the key format FORMAT_NETSCAPE and the variant
11464 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11466 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11468 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11470 * Add -keyform to rsautl, and document -engine.
11472 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11474 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11475 BIO_R_NO_SUCH_FILE error code rather than the generic
11476 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11480 * Add new functions
11481 ERR_peek_last_error
11482 ERR_peek_last_error_line
11483 ERR_peek_last_error_line_data.
11484 These are similar to
11486 ERR_peek_error_line
11487 ERR_peek_error_line_data,
11488 but report on the latest error recorded rather than the first one
11489 still in the error queue.
11491 *Ben Laurie, Bodo Moeller*
11493 * default_algorithms option in ENGINE config module. This allows things
11495 default_algorithms = ALL
11496 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11500 * Preliminary ENGINE config module.
11504 * New experimental application configuration code.
11508 * Change the AES code to follow the same name structure as all other
11509 symmetric ciphers, and behave the same way. Move everything to
11510 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11512 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11514 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11516 *Ben Laurie and Theo de Raadt*
11518 * Add option to output public keys in req command.
11520 *Massimiliano Pala madwolf@openca.org*
11522 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11523 (up to about 10% better than before for P-192 and P-224).
11527 * New functions/macros
11529 SSL_CTX_set_msg_callback(ctx, cb)
11530 SSL_CTX_set_msg_callback_arg(ctx, arg)
11531 SSL_set_msg_callback(ssl, cb)
11532 SSL_set_msg_callback_arg(ssl, arg)
11534 to request calling a callback function
11536 void cb(int write_p, int version, int content_type,
11537 const void *buf, size_t len, SSL *ssl, void *arg)
11539 whenever a protocol message has been completely received
11540 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11541 protocol version according to which the SSL library interprets
11542 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11543 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11544 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11545 specification (change_cipher_spec(20), alert(21), handshake(22)).
11546 'buf' and 'len' point to the actual message, 'ssl' to the
11547 SSL object, and 'arg' is the application-defined value set by
11548 SSL[_CTX]_set_msg_callback_arg().
11550 'openssl s_client' and 'openssl s_server' have new '-msg' options
11551 to enable a callback that displays all protocol messages.
11555 * Change the shared library support so shared libraries are built as
11556 soon as the corresponding static library is finished, and thereby get
11557 openssl and the test programs linked against the shared library.
11558 This still only happens when the keyword "shard" has been given to
11559 the configuration scripts.
11561 NOTE: shared library support is still an experimental thing, and
11562 backward binary compatibility is still not guaranteed.
11564 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11566 * Add support for Subject Information Access extension.
11568 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11570 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11571 additional bytes when new memory had to be allocated, not just
11572 when reusing an existing buffer.
11576 * New command line and configuration option 'utf8' for the req command.
11577 This allows field values to be specified as UTF8 strings.
11581 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11582 runs for the former and machine-readable output for the latter.
11586 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11587 of the e-mail address in the DN (i.e., it will go into a certificate
11588 extension only). The new configuration file option 'email_in_dn = no'
11589 has the same effect.
11591 *Massimiliano Pala madwolf@openca.org*
11593 * Change all functions with names starting with `des_` to be starting
11594 with `DES_` instead. Add wrappers that are compatible with libdes,
11595 but are named `_ossl_old_des_*`. Finally, add macros that map the
11596 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11597 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11598 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11601 Since we provide two compatibility mappings, the user needs to
11602 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11603 compatibility is desired. The default (i.e., when that macro
11604 isn't defined) is OpenSSL 0.9.6c compatibility.
11606 There are also macros that enable and disable the support of old
11607 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11608 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11609 are defined, the default will apply: to support the old des routines.
11611 In either case, one must include openssl/des.h to get the correct
11612 definitions. Do not try to just include openssl/des_old.h, that
11615 NOTE: This is a major break of an old API into a new one. Software
11616 authors are encouraged to switch to the `DES_` style functions. Some
11617 time in the future, des_old.h and the libdes compatibility functions
11618 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11619 default), and then completely removed.
11623 * Test for certificates which contain unsupported critical extensions.
11624 If such a certificate is found during a verify operation it is
11625 rejected by default: this behaviour can be overridden by either
11626 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11627 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11628 X509_supported_extension() has also been added which returns 1 if a
11629 particular extension is supported.
11633 * Modify the behaviour of EVP cipher functions in similar way to digests
11634 to retain compatibility with existing code.
11638 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11639 compatibility with existing code. In particular the 'ctx' parameter does
11640 not have to be to be initialized before the call to EVP_DigestInit() and
11641 it is tidied up after a call to EVP_DigestFinal(). New function
11642 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11643 EVP_MD_CTX_copy() changed to not require the destination to be
11644 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11645 requires the destination to be valid.
11647 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11648 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11652 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11653 so that complete 'Handshake' protocol structures are kept in memory
11654 instead of overwriting 'msg_type' and 'length' with 'body' data.
11658 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11660 *Massimo Santin via Richard Levitte*
11662 * Major restructuring to the underlying ENGINE code. This includes
11663 reduction of linker bloat, separation of pure "ENGINE" manipulation
11664 (initialisation, etc) from functionality dealing with implementations
11665 of specific crypto interfaces. This change also introduces integrated
11666 support for symmetric ciphers and digest implementations - so ENGINEs
11667 can now accelerate these by providing EVP_CIPHER and EVP_MD
11668 implementations of their own. This is detailed in
11669 [crypto/engine/README.md](crypto/engine/README.md)
11670 as it couldn't be adequately described here. However, there are a few
11671 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11672 were changed in the original introduction of ENGINE code have now
11673 reverted back - the hooking from this code to ENGINE is now a good
11674 deal more passive and at run-time, operations deal directly with
11675 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11676 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11677 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11678 they were not being used by the framework as there is no concept of a
11679 BIGNUM_METHOD and they could not be generalised to the new
11680 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11681 ENGINE_cpy() has been removed as it cannot be consistently defined in
11686 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11690 * Change mkdef.pl to sort symbols that get the same entry number,
11691 and make sure the automatically generated functions `ERR_load_*`
11692 become part of libeay.num as well.
11696 * New function SSL_renegotiate_pending(). This returns true once
11697 renegotiation has been requested (either SSL_renegotiate() call
11698 or HelloRequest/ClientHello received from the peer) and becomes
11699 false once a handshake has been completed.
11700 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11701 sends a HelloRequest, but does not ensure that a handshake takes
11702 place. SSL_renegotiate_pending() is useful for checking if the
11703 client has followed the request.)
11707 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11708 By default, clients may request session resumption even during
11709 renegotiation (if session ID contexts permit); with this option,
11710 session resumption is possible only in the first handshake.
11712 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11713 more bits available for options that should not be part of
11714 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11718 * Add some demos for certificate and certificate request creation.
11722 * Make maximum certificate chain size accepted from the peer application
11723 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11724 "Douglas E. Engert" <deengert@anl.gov>.
11728 * Add support for shared libraries for Unixware-7
11729 (Boyd Lynn Gerber <gerberb@zenez.com>).
11733 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11734 be done prior to destruction. Use this to unload error strings from
11735 ENGINEs that load their own error strings. NB: This adds two new API
11736 functions to "get" and "set" this destroy handler in an ENGINE.
11740 * Alter all existing ENGINE implementations (except "openssl" and
11741 "openbsd") to dynamically instantiate their own error strings. This
11742 makes them more flexible to be built both as statically-linked ENGINEs
11743 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11744 Also, add stub code to each that makes building them as self-contained
11745 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11749 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11750 implementations into applications that are completely implemented in
11751 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11752 commands that can be used to configure what shared-library to load and
11753 to control aspects of the way it is handled. Also, made an update to
11754 the [README-Engine.md](README-Engine.md) file
11755 that brings its information up-to-date and
11756 provides some information and instructions on the "dynamic" ENGINE
11757 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11761 * Make it possible to unload ranges of ERR strings with a new
11762 "ERR_unload_strings" function.
11766 * Add a copy() function to EVP_MD.
11770 * Make EVP_MD routines take a context pointer instead of just the
11771 md_data void pointer.
11775 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11776 that the digest can only process a single chunk of data
11777 (typically because it is provided by a piece of
11778 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11779 is only going to provide a single chunk of data, and hence the
11780 framework needn't accumulate the data for oneshot drivers.
11784 * As with "ERR", make it possible to replace the underlying "ex_data"
11785 functions. This change also alters the storage and management of global
11786 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11787 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11788 index counters. The API functions that use this state have been changed
11789 to take a "class_index" rather than pointers to the class's local STACK
11790 and counter, and there is now an API function to dynamically create new
11791 classes. This centralisation allows us to (a) plug a lot of the
11792 thread-safety problems that existed, and (b) makes it possible to clean
11793 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11794 such data would previously have always leaked in application code and
11795 workarounds were in place to make the memory debugging turn a blind eye
11796 to it. Application code that doesn't use this new function will still
11797 leak as before, but their memory debugging output will announce it now
11798 rather than letting it slide.
11800 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11801 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11802 has a return value to indicate success or failure.
11806 * Make it possible to replace the underlying "ERR" functions such that the
11807 global state (2 LHASH tables and 2 locks) is only used by the "default"
11808 implementation. This change also adds two functions to "get" and "set"
11809 the implementation prior to it being automatically set the first time
11810 any other ERR function takes place. Ie. an application can call "get",
11811 pass the return value to a module it has just loaded, and that module
11812 can call its own "set" function using that value. This means the
11813 module's "ERR" operations will use (and modify) the error state in the
11814 application and not in its own statically linked copy of OpenSSL code.
11818 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11819 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11820 the operation, and provides a more encapsulated way for external code
11821 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11822 to use these functions rather than manually incrementing the counts.
11824 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11828 * Add EVP test program.
11832 * Add symmetric cipher support to ENGINE. Expect the API to change!
11836 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11837 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11838 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11839 These allow a CRL to be built without having to access X509_CRL fields
11840 directly. Modify 'ca' application to use new functions.
11844 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11845 bug workarounds. Rollback attack detection is a security feature.
11846 The problem will only arise on OpenSSL servers when TLSv1 is not
11847 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11848 Software authors not wanting to support TLSv1 will have special reasons
11849 for their choice and can explicitly enable this option.
11851 *Bodo Moeller, Lutz Jaenicke*
11853 * Rationalise EVP so it can be extended: don't include a union of
11854 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11855 (similar to those existing for EVP_CIPHER_CTX).
11860 EVP_MD_CTX_init(&md); /* new function call */
11861 EVP_DigestInit(&md, EVP_sha1());
11862 EVP_DigestUpdate(&md, in, len);
11863 EVP_DigestFinal(&md, out, NULL);
11864 EVP_MD_CTX_cleanup(&md); /* new function call */
11868 * Make DES key schedule conform to the usual scheme, as well as
11869 correcting its structure. This means that calls to DES functions
11870 now have to pass a pointer to a des_key_schedule instead of a
11871 plain des_key_schedule (which was actually always a pointer
11874 des_key_schedule ks;
11876 des_set_key_checked(..., &ks);
11877 des_ncbc_encrypt(..., &ks, ...);
11879 (Note that a later change renames 'des_...' into 'DES_...'.)
11883 * Initial reduction of linker bloat: the use of some functions, such as
11884 PEM causes large amounts of unused functions to be linked in due to
11885 poor organisation. For example pem_all.c contains every PEM function
11886 which has a knock on effect of linking in large amounts of (unused)
11887 ASN1 code. Grouping together similar functions and splitting unrelated
11888 functions prevents this.
11892 * Cleanup of EVP macros.
11896 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11897 correct `_ecb suffix`.
11901 * Add initial OCSP responder support to ocsp application. The
11902 revocation information is handled using the text based index
11903 use by the ca application. The responder can either handle
11904 requests generated internally, supplied in files (for example
11905 via a CGI script) or using an internal minimal server.
11909 * Add configuration choices to get zlib compression for TLS.
11913 * Changes to Kerberos SSL for RFC 2712 compliance:
11914 1. Implemented real KerberosWrapper, instead of just using
11915 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11916 2. Implemented optional authenticator field of KerberosWrapper.
11918 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11919 and authenticator structs; see crypto/krb5/.
11921 Generalized Kerberos calls to support multiple Kerberos libraries.
11922 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11923 via Richard Levitte*
11925 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11926 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11927 values for each of the key sizes rather than having just
11928 parameters (and 'speed' generating keys each time).
11932 * Speed up EVP routines.
11935 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11936 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11937 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11938 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11940 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11941 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11942 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11945 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11947 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11951 * Added the OS2-EMX target.
11953 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11955 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11956 to support NCONF routines in extension code. New function CONF_set_nconf()
11957 to allow functions which take an NCONF to also handle the old LHASH
11958 structure: this means that the old CONF compatible routines can be
11959 retained (in particular wrt extensions) without having to duplicate the
11960 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11964 * Enhance the general user interface with mechanisms for inner control
11965 and with possibilities to have yes/no kind of prompts.
11969 * Change all calls to low level digest routines in the library and
11970 applications to use EVP. Add missing calls to HMAC_cleanup() and
11971 don't assume HMAC_CTX can be copied using memcpy().
11973 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11975 * Add the possibility to control engines through control names but with
11976 arbitrary arguments instead of just a string.
11977 Change the key loaders to take a UI_METHOD instead of a callback
11978 function pointer. NOTE: this breaks binary compatibility with earlier
11979 versions of OpenSSL [engine].
11980 Adapt the nCipher code for these new conditions and add a card insertion
11985 * Enhance the general user interface with mechanisms to better support
11986 dialog box interfaces, application-defined prompts, the possibility
11987 to use defaults (for example default passwords from somewhere else)
11988 and interrupts/cancellations.
11992 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11993 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11997 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11998 tidy up some unnecessarily weird code in 'sk_new()').
12000 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12002 * Change the key loading routines for ENGINEs to use the same kind
12003 callback (pem_password_cb) as all other routines that need this
12008 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12009 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12010 than this minimum value is recommended.
12014 * New random seeder for OpenVMS, using the system process statistics
12015 that are easily reachable.
12019 * Windows apparently can't transparently handle global
12020 variables defined in DLLs. Initialisations such as:
12022 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12024 won't compile. This is used by the any applications that need to
12025 declare their own ASN1 modules. This was fixed by adding the option
12026 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12027 needed for static libraries under Win32.
12031 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12032 setting of purpose and trust fields. New X509_STORE trust and
12033 purpose functions and tidy up setting in other SSL functions.
12037 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12038 structure. These are inherited by X509_STORE_CTX when it is
12039 initialised. This allows various defaults to be set in the
12040 X509_STORE structure (such as flags for CRL checking and custom
12041 purpose or trust settings) for functions which only use X509_STORE_CTX
12042 internally such as S/MIME.
12044 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12045 trust settings if they are not set in X509_STORE. This allows X509_STORE
12046 purposes and trust (in S/MIME for example) to override any set by default.
12048 Add command line options for CRL checking to smime, s_client and s_server
12053 * Initial CRL based revocation checking. If the CRL checking flag(s)
12054 are set then the CRL is looked up in the X509_STORE structure and
12055 its validity and signature checked, then if the certificate is found
12056 in the CRL the verify fails with a revoked error.
12058 Various new CRL related callbacks added to X509_STORE_CTX structure.
12060 Command line options added to 'verify' application to support this.
12062 This needs some additional work, such as being able to handle multiple
12063 CRLs with different times, extension based lookup (rather than just
12064 by subject name) and ultimately more complete V2 CRL extension
12069 * Add a general user interface API (crypto/ui/). This is designed
12070 to replace things like des_read_password and friends (backward
12071 compatibility functions using this new API are provided).
12072 The purpose is to remove prompting functions from the DES code
12073 section as well as provide for prompting through dialog boxes in
12074 a window system and the like.
12078 * Add "ex_data" support to ENGINE so implementations can add state at a
12079 per-structure level rather than having to store it globally.
12083 * Make it possible for ENGINE structures to be copied when retrieved by
12084 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12085 This causes the "original" ENGINE structure to act like a template,
12086 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12087 operational state can be localised to each ENGINE structure, despite the
12088 fact they all share the same "methods". New ENGINE structures returned in
12089 this case have no functional references and the return value is the single
12090 structural reference. This matches the single structural reference returned
12091 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12096 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12097 needs to match any other type at all we need to manually clear the
12102 * Changes to the "openssl engine" utility to include;
12103 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12104 about an ENGINE's available control commands.
12105 - executing control commands from command line arguments using the
12106 '-pre' and '-post' switches. '-post' is only used if '-t' is
12107 specified and the ENGINE is successfully initialised. The syntax for
12108 the individual commands are colon-separated, for example;
12109 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12113 * New dynamic control command support for ENGINEs. ENGINEs can now
12114 declare their own commands (numbers), names (strings), descriptions,
12115 and input types for run-time discovery by calling applications. A
12116 subset of these commands are implicitly classed as "executable"
12117 depending on their input type, and only these can be invoked through
12118 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12119 can be based on user input, config files, etc). The distinction is
12120 that "executable" commands cannot return anything other than a boolean
12121 result and can only support numeric or string input, whereas some
12122 discoverable commands may only be for direct use through
12123 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12124 pointers, or other custom uses. The "executable" commands are to
12125 support parameterisations of ENGINE behaviour that can be
12126 unambiguously defined by ENGINEs and used consistently across any
12127 OpenSSL-based application. Commands have been added to all the
12128 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12129 control over shared-library paths without source code alterations.
12133 * Changed all ENGINE implementations to dynamically allocate their
12134 ENGINEs rather than declaring them statically. Apart from this being
12135 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12136 this also allows the implementations to compile without using the
12137 internal engine_int.h header.
12141 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12142 'const' value. Any code that should be able to modify a RAND_METHOD
12143 should already have non-const pointers to it (ie. they should only
12144 modify their own ones).
12148 * Made a variety of little tweaks to the ENGINE code.
12149 - "atalla" and "ubsec" string definitions were moved from header files
12150 to C code. "nuron" string definitions were placed in variables
12151 rather than hard-coded - allowing parameterisation of these values
12152 later on via ctrl() commands.
12153 - Removed unused "#if 0"'d code.
12154 - Fixed engine list iteration code so it uses ENGINE_free() to release
12155 structural references.
12156 - Constified the RAND_METHOD element of ENGINE structures.
12157 - Constified various get/set functions as appropriate and added
12158 missing functions (including a catch-all ENGINE_cpy that duplicates
12159 all ENGINE values onto a new ENGINE except reference counts/state).
12160 - Removed NULL parameter checks in get/set functions. Setting a method
12161 or function to NULL is a way of cancelling out a previously set
12162 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12163 and doesn't justify the extra error symbols and code.
12164 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12165 flags from engine_int.h to engine.h.
12166 - Changed prototypes for ENGINE handler functions (init(), finish(),
12167 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12171 * Implement binary inversion algorithm for BN_mod_inverse in addition
12172 to the algorithm using long division. The binary algorithm can be
12173 used only if the modulus is odd. On 32-bit systems, it is faster
12174 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12175 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12176 up to 450 bits. In 64-bit environments, the binary algorithm
12177 appears to be advantageous for much longer moduli; here we use it
12178 for moduli up to 2048 bits.
12182 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12183 could not support the combine flag in choice fields.
12187 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12188 extensions from a certificate request to the certificate.
12192 * Allow multiple 'certopt' and 'nameopt' options to be separated
12193 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12194 file: this allows the display of the certificate about to be
12195 signed to be customised, to allow certain fields to be included
12196 or excluded and extension details. The old system didn't display
12197 multicharacter strings properly, omitted fields not in the policy
12198 and couldn't display additional details such as extensions.
12202 * Function EC_POINTs_mul for multiple scalar multiplication
12203 of an arbitrary number of elliptic curve points
12204 \sum scalars[i]*points[i],
12205 optionally including the generator defined for the EC_GROUP:
12206 scalar*generator + \sum scalars[i]*points[i].
12208 EC_POINT_mul is a simple wrapper function for the typical case
12209 that the point list has just one item (besides the optional
12214 * First EC_METHODs for curves over GF(p):
12216 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12217 operations and provides various method functions that can also
12218 operate with faster implementations of modular arithmetic.
12220 EC_GFp_mont_method() reuses most functions that are part of
12221 EC_GFp_simple_method, but uses Montgomery arithmetic.
12223 *Bodo Moeller; point addition and point doubling
12224 implementation directly derived from source code provided by
12225 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12227 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12228 crypto/ec/ec_lib.c):
12230 Curves are EC_GROUP objects (with an optional group generator)
12231 based on EC_METHODs that are built into the library.
12233 Points are EC_POINT objects based on EC_GROUP objects.
12235 Most of the framework would be able to handle curves over arbitrary
12236 finite fields, but as there are no obvious types for fields other
12237 than GF(p), some functions are limited to that for now.
12241 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12242 that the file contains a complete HTTP response.
12246 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12247 change the def and num file printf format specifier from "%-40sXXX"
12248 to "%-39s XXX". The latter will always guarantee a space after the
12249 field while the former will cause them to run together if the field
12250 is 40 of more characters long.
12254 * Constify the cipher and digest 'method' functions and structures
12255 and modify related functions to take constant EVP_MD and EVP_CIPHER
12260 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12261 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12265 * Modify `EVP_Digest*()` routines so they now return values. Although the
12266 internal software routines can never fail additional hardware versions
12271 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12273 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12274 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12277 ERR_R_NESTED_ASN1_ERROR
12279 ERR_R_MISSING_ASN1_EOS
12280 were 4 .. 9, conflicting with
12281 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12283 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12284 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12286 Add new error code 'ERR_R_INTERNAL_ERROR'.
12290 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12295 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12296 sets the subject name for a new request or supersedes the
12297 subject name in a given request. Formats that can be parsed are
12298 'CN=Some Name, OU=myOU, C=IT'
12300 'CN=Some Name/OU=myOU/C=IT'.
12302 Add options '-batch' and '-verbose' to 'openssl req'.
12304 *Massimiliano Pala <madwolf@hackmasters.net>*
12306 * Introduce the possibility to access global variables through
12307 functions on platform were that's the best way to handle exporting
12308 global variables in shared libraries. To enable this functionality,
12309 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12310 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12311 is normally done by Configure or something similar).
12313 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12314 in the source file (foo.c) like this:
12316 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12317 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12319 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12320 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12322 OPENSSL_DECLARE_GLOBAL(int,foo);
12323 #define foo OPENSSL_GLOBAL_REF(foo)
12324 OPENSSL_DECLARE_GLOBAL(double,bar);
12325 #define bar OPENSSL_GLOBAL_REF(bar)
12327 The #defines are very important, and therefore so is including the
12328 header file everywhere where the defined globals are used.
12330 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12331 of ASN.1 items, but that structure is a bit different.
12333 The largest change is in util/mkdef.pl which has been enhanced with
12334 better and easier to understand logic to choose which symbols should
12335 go into the Windows .def files as well as a number of fixes and code
12336 cleanup (among others, algorithm keywords are now sorted
12337 lexicographically to avoid constant rewrites).
12341 * In BN_div() keep a copy of the sign of 'num' before writing the
12342 result to 'rm' because if rm==num the value will be overwritten
12343 and produce the wrong result if 'num' is negative: this caused
12344 problems with BN_mod() and BN_nnmod().
12348 * Function OCSP_request_verify(). This checks the signature on an
12349 OCSP request and verifies the signer certificate. The signer
12350 certificate is just checked for a generic purpose and OCSP request
12355 * Add OCSP_check_validity() function to check the validity of OCSP
12356 responses. OCSP responses are prepared in real time and may only
12357 be a few seconds old. Simply checking that the current time lies
12358 between thisUpdate and nextUpdate max reject otherwise valid responses
12359 caused by either OCSP responder or client clock inaccuracy. Instead
12360 we allow thisUpdate and nextUpdate to fall within a certain period of
12361 the current time. The age of the response can also optionally be
12362 checked. Two new options -validity_period and -status_age added to
12367 * If signature or public key algorithm is unrecognized print out its
12368 OID rather that just UNKNOWN.
12372 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12373 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12374 ID to be generated from the issuer certificate alone which can then be
12375 passed to OCSP_id_issuer_cmp().
12379 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12380 ASN1 modules to export functions returning ASN1_ITEM pointers
12381 instead of the ASN1_ITEM structures themselves. This adds several
12382 new macros which allow the underlying ASN1 function/structure to
12383 be accessed transparently. As a result code should not use ASN1_ITEM
12384 references directly (such as &X509_it) but instead use the relevant
12385 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12386 use of the new ASN1 code on platforms where exporting structures
12387 is problematical (for example in shared libraries) but exporting
12388 functions returning pointers to structures is not.
12392 * Add support for overriding the generation of SSL/TLS session IDs.
12393 These callbacks can be registered either in an SSL_CTX or per SSL.
12394 The purpose of this is to allow applications to control, if they wish,
12395 the arbitrary values chosen for use as session IDs, particularly as it
12396 can be useful for session caching in multiple-server environments. A
12397 command-line switch for testing this (and any client code that wishes
12398 to use such a feature) has been added to "s_server".
12400 *Geoff Thorpe, Lutz Jaenicke*
12402 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12403 of the form `#if defined(...) || defined(...) || ...` and
12404 `#if !defined(...) && !defined(...) && ...`. This also avoids
12405 the growing number of special cases it was previously handling.
12409 * Make all configuration macros available for application by making
12410 sure they are available in opensslconf.h, by giving them names starting
12411 with `OPENSSL_` to avoid conflicts with other packages and by making
12412 sure e_os2.h will cover all platform-specific cases together with
12414 Additionally, it is now possible to define configuration/platform-
12415 specific names (called "system identities"). In the C code, these
12416 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12417 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12418 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12423 * New option -set_serial to 'req' and 'x509' this allows the serial
12424 number to use to be specified on the command line. Previously self
12425 signed certificates were hard coded with serial number 0 and the
12426 CA options of 'x509' had to use a serial number in a file which was
12431 * New options to 'ca' utility to support V2 CRL entry extensions.
12432 Currently CRL reason, invalidity date and hold instruction are
12433 supported. Add new CRL extensions to V3 code and some new objects.
12437 * New function EVP_CIPHER_CTX_set_padding() this is used to
12438 disable standard block padding (aka PKCS#5 padding) in the EVP
12439 API, which was previously mandatory. This means that the data is
12440 not padded in any way and so the total length much be a multiple
12441 of the block size, otherwise an error occurs.
12445 * Initial (incomplete) OCSP SSL support.
12449 * New function OCSP_parse_url(). This splits up a URL into its host,
12450 port and path components: primarily to parse OCSP URLs. New -url
12451 option to ocsp utility.
12455 * New nonce behavior. The return value of OCSP_check_nonce() now
12456 reflects the various checks performed. Applications can decide
12457 whether to tolerate certain situations such as an absent nonce
12458 in a response when one was present in a request: the ocsp application
12459 just prints out a warning. New function OCSP_add1_basic_nonce()
12460 this is to allow responders to include a nonce in a response even if
12461 the request is nonce-less.
12465 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12466 skipped when using openssl x509 multiple times on a single input file,
12467 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12471 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12472 set string type: to handle setting ASN1_TIME structures. Fix ca
12473 utility to correctly initialize revocation date of CRLs.
12477 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12478 the clients preferred ciphersuites and rather use its own preferences.
12479 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12480 Internet Explorer by ensuring unchanged hash method during stepup.
12481 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12485 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12486 to aes and add a new 'exist' option to print out symbols that don't
12491 * Additional options to ocsp utility to allow flags to be set and
12492 additional certificates supplied.
12496 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12497 OCSP client a number of certificate to only verify the response
12502 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12503 handle the new API. Currently only ECB, CBC modes supported. Add new
12506 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12507 Encryption Standard (AES) Ciphersuites for Transport Layer
12508 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12509 not enabled by default and were not part of the "ALL" ciphersuite
12510 alias because they were not yet official; they could be
12511 explicitly requested by specifying the "AESdraft" ciphersuite
12512 group alias. In the final release of OpenSSL 0.9.7, the group
12513 alias is called "AES" and is part of "ALL".)
12515 *Ben Laurie, Steve Henson, Bodo Moeller*
12517 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12518 request to response.
12522 * Functions for OCSP responders. OCSP_request_onereq_count(),
12523 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12524 extract information from a certificate request. OCSP_response_create()
12525 creates a response and optionally adds a basic response structure.
12526 OCSP_basic_add1_status() adds a complete single response to a basic
12527 response and returns the OCSP_SINGLERESP structure just added (to allow
12528 extensions to be included for example). OCSP_basic_add1_cert() adds a
12529 certificate to a basic response and OCSP_basic_sign() signs a basic
12530 response with various flags. New helper functions ASN1_TIME_check()
12531 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12532 (converts ASN1_TIME to GeneralizedTime).
12536 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12537 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12538 structure from a certificate. X509_pubkey_digest() digests the public_key
12539 contents: this is used in various key identifiers.
12543 * Make sk_sort() tolerate a NULL argument.
12545 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12547 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12548 passed by the function are trusted implicitly. If any of them signed the
12549 response then it is assumed to be valid and is not verified.
12553 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12554 to data. This was previously part of the PKCS7 ASN1 code. This
12555 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12556 *Steve Henson, reported by Kenneth R. Robinette
12557 <support@securenetterm.com>*
12559 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12560 routines: without these tracing memory leaks is very painful.
12561 Fix leaks in PKCS12 and PKCS7 routines.
12565 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12566 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12567 effectively meant GeneralizedTime would never be used. Now it
12568 is initialised to -1 but X509_time_adj() now has to check the value
12569 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12570 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12571 *Steve Henson, reported by Kenneth R. Robinette
12572 <support@securenetterm.com>*
12574 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12575 result in a zero length in the ASN1_INTEGER structure which was
12576 not consistent with the structure when d2i_ASN1_INTEGER() was used
12577 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12578 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12579 where it did not print out a minus for negative ASN1_INTEGER.
12583 * Add summary printout to ocsp utility. The various functions which
12584 convert status values to strings have been renamed to:
12585 OCSP_response_status_str(), OCSP_cert_status_str() and
12586 OCSP_crl_reason_str() and are no longer static. New options
12587 to verify nonce values and to disable verification. OCSP response
12588 printout format cleaned up.
12592 * Add additional OCSP certificate checks. These are those specified
12593 in RFC2560. This consists of two separate checks: the CA of the
12594 certificate being checked must either be the OCSP signer certificate
12595 or the issuer of the OCSP signer certificate. In the latter case the
12596 OCSP signer certificate must contain the OCSP signing extended key
12597 usage. This check is performed by attempting to match the OCSP
12598 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12599 in the OCSP_CERTID structures of the response.
12603 * Initial OCSP certificate verification added to OCSP_basic_verify()
12604 and related routines. This uses the standard OpenSSL certificate
12605 verify routines to perform initial checks (just CA validity) and
12606 to obtain the certificate chain. Then additional checks will be
12607 performed on the chain. Currently the root CA is checked to see
12608 if it is explicitly trusted for OCSP signing. This is used to set
12609 a root CA as a global signing root: that is any certificate that
12610 chains to that CA is an acceptable OCSP signing certificate.
12614 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12615 extensions from a separate configuration file.
12616 As when reading extensions from the main configuration file,
12617 the '-extensions ...' option may be used for specifying the
12620 *Massimiliano Pala <madwolf@comune.modena.it>*
12622 * New OCSP utility. Allows OCSP requests to be generated or
12623 read. The request can be sent to a responder and the output
12624 parsed, outputted or printed in text form. Not complete yet:
12625 still needs to check the OCSP response validity.
12629 * New subcommands for 'openssl ca':
12630 `openssl ca -status <serial>` prints the status of the cert with
12631 the given serial number (according to the index file).
12632 `openssl ca -updatedb` updates the expiry status of certificates
12635 *Massimiliano Pala <madwolf@comune.modena.it>*
12637 * New '-newreq-nodes' command option to CA.pl. This is like
12638 '-newreq', but calls 'openssl req' with the '-nodes' option
12639 so that the resulting key is not encrypted.
12641 *Damien Miller <djm@mindrot.org>*
12643 * New configuration for the GNU Hurd.
12645 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12647 * Initial code to implement OCSP basic response verify. This
12648 is currently incomplete. Currently just finds the signer's
12649 certificate and verifies the signature on the response.
12653 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12654 value of OPENSSLDIR. This is available via the new '-d' option
12655 to 'openssl version', and is also included in 'openssl version -a'.
12659 * Allowing defining memory allocation callbacks that will be given
12660 file name and line number information in additional arguments
12661 (a `const char*` and an int). The basic functionality remains, as
12662 well as the original possibility to just replace malloc(),
12663 realloc() and free() by functions that do not know about these
12664 additional arguments. To register and find out the current
12665 settings for extended allocation functions, the following
12666 functions are provided:
12668 CRYPTO_set_mem_ex_functions
12669 CRYPTO_set_locked_mem_ex_functions
12670 CRYPTO_get_mem_ex_functions
12671 CRYPTO_get_locked_mem_ex_functions
12673 These work the same way as CRYPTO_set_mem_functions and friends.
12674 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12675 extended allocation function is enabled.
12676 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12677 a conventional allocation function is enabled.
12679 *Richard Levitte, Bodo Moeller*
12681 * Finish off removing the remaining LHASH function pointer casts.
12682 There should no longer be any prototype-casting required when using
12683 the LHASH abstraction, and any casts that remain are "bugs". See
12684 the callback types and macros at the head of lhash.h for details
12685 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12689 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12690 If /dev/[u]random devices are not available or do not return enough
12691 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12693 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12694 /etc/entropy will be queried once each in this sequence, querying stops
12695 when enough entropy was collected without querying more sockets.
12699 * Change the Unix RAND_poll() variant to be able to poll several
12700 random devices, as specified by DEVRANDOM, until a sufficient amount
12701 of data has been collected. We spend at most 10 ms on each file
12702 (select timeout) and read in non-blocking mode. DEVRANDOM now
12703 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12704 (previously it was just the string "/dev/urandom"), so on typical
12705 platforms the 10 ms delay will never occur.
12706 Also separate out the Unix variant to its own file, rand_unix.c.
12707 For VMS, there's a currently-empty rand_vms.c.
12711 * Move OCSP client related routines to ocsp_cl.c. These
12712 provide utility functions which an application needing
12713 to issue a request to an OCSP responder and analyse the
12714 response will typically need: as opposed to those which an
12715 OCSP responder itself would need which will be added later.
12717 OCSP_request_sign() signs an OCSP request with an API similar
12718 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12719 response. OCSP_response_get1_basic() extracts basic response
12720 from response. OCSP_resp_find_status(): finds and extracts status
12721 information from an OCSP_CERTID structure (which will be created
12722 when the request structure is built). These are built from lower
12723 level functions which work on OCSP_SINGLERESP structures but
12724 won't normally be used unless the application wishes to examine
12725 extensions in the OCSP response for example.
12727 Replace nonce routines with a pair of functions.
12728 OCSP_request_add1_nonce() adds a nonce value and optionally
12729 generates a random value. OCSP_check_nonce() checks the
12730 validity of the nonce in an OCSP response.
12734 * Change function OCSP_request_add() to OCSP_request_add0_id().
12735 This doesn't copy the supplied OCSP_CERTID and avoids the
12736 need to free up the newly created id. Change return type
12737 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12738 This can then be used to add extensions to the request.
12739 Deleted OCSP_request_new(), since most of its functionality
12740 is now in OCSP_REQUEST_new() (and the case insensitive name
12741 clash) apart from the ability to set the request name which
12742 will be added elsewhere.
12746 * Update OCSP API. Remove obsolete extensions argument from
12747 various functions. Extensions are now handled using the new
12748 OCSP extension code. New simple OCSP HTTP function which
12749 can be used to send requests and parse the response.
12753 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12754 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12755 uses the special reorder version of SET OF to sort the attributes
12756 and reorder them to match the encoded order. This resolves a long
12757 standing problem: a verify on a PKCS7 structure just after signing
12758 it used to fail because the attribute order did not match the
12759 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12760 it uses the received order. This is necessary to tolerate some broken
12761 software that does not order SET OF. This is handled by encoding
12762 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12763 to produce the required SET OF.
12767 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12768 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12769 files to get correct declarations of the ASN.1 item variables.
12773 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12774 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12775 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12776 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12777 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12778 ASN1_ITEM and no wrapper functions.
12782 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12783 replace the old function pointer based I/O routines. Change most of
12784 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12788 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12789 lines, recognize more "algorithms" that can be deselected, and make
12790 it complain about algorithm deselection that isn't recognised.
12794 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12795 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12796 to use new functions. Add NO_ASN1_OLD which can be set to remove
12797 some old style ASN1 functions: this can be used to determine if old
12798 code will still work when these eventually go away.
12802 * New extension functions for OCSP structures, these follow the
12803 same conventions as certificates and CRLs.
12807 * New function X509V3_add1_i2d(). This automatically encodes and
12808 adds an extension. Its behaviour can be customised with various
12809 flags to append, replace or delete. Various wrappers added for
12810 certificates and CRLs.
12814 * Fix to avoid calling the underlying ASN1 print routine when
12815 an extension cannot be parsed. Correct a typo in the
12816 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12820 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12821 entries for variables.
12825 * Add functionality to apps/openssl.c for detecting locking
12826 problems: As the program is single-threaded, all we have
12827 to do is register a locking callback using an array for
12828 storing which locks are currently held by the program.
12832 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12833 SSL_get_ex_data_X509_STORE_idx(), which is used in
12834 ssl_verify_cert_chain() and thus can be called at any time
12835 during TLS/SSL handshakes so that thread-safety is essential.
12836 Unfortunately, the ex_data design is not at all suited
12837 for multi-threaded use, so it probably should be abolished.
12841 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12843 *Broadcom, tweaked and integrated by Geoff Thorpe*
12845 * Move common extension printing code to new function
12846 X509V3_print_extensions(). Reorganise OCSP print routines and
12847 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12851 * New function X509_signature_print() to remove duplication in some
12856 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12857 set (this was treated exactly the same as SET OF previously). This
12858 is used to reorder the STACK representing the structure to match the
12859 encoding. This will be used to get round a problem where a PKCS7
12860 structure which was signed could not be verified because the STACK
12861 order did not reflect the encoded order.
12865 * Reimplement the OCSP ASN1 module using the new code.
12869 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12870 for its ASN1 operations. The old style function pointers still exist
12871 for now but they will eventually go away.
12875 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12876 completely replaces the old ASN1 functionality with a table driven
12877 encoder and decoder which interprets an ASN1_ITEM structure describing
12878 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12879 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12880 has also been converted to the new form.
12884 * Change BN_mod_exp_recp so that negative moduli are tolerated
12885 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12886 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12887 for negative moduli.
12891 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12892 of not touching the result's sign bit.
12896 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12901 * Changed the LHASH code to use prototypes for callbacks, and created
12902 macros to declare and implement thin (optionally static) functions
12903 that provide type-safety and avoid function pointer casting for the
12904 type-specific callbacks.
12908 * Added Kerberos Cipher Suites to be used with TLS, as written in
12910 *Veers Staats <staatsvr@asc.hpc.mil>,
12911 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12913 * Reformat the FAQ so the different questions and answers can be divided
12914 in sections depending on the subject.
12918 * Have the zlib compression code load ZLIB.DLL dynamically under
12923 * New function BN_mod_sqrt for computing square roots modulo a prime
12924 (using the probabilistic Tonelli-Shanks algorithm unless
12925 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12926 be handled deterministically).
12928 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12930 * Make BN_mod_inverse faster by explicitly handling small quotients
12931 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12932 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12936 * New function BN_kronecker.
12940 * Fix BN_gcd so that it works on negative inputs; the result is
12941 positive unless both parameters are zero.
12942 Previously something reasonably close to an infinite loop was
12943 possible because numbers could be growing instead of shrinking
12944 in the implementation of Euclid's algorithm.
12948 * Fix BN_is_word() and BN_is_one() macros to take into account the
12949 sign of the number in question.
12951 Fix BN_is_word(a,w) to work correctly for w == 0.
12953 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12954 because its test if the absolute value of 'a' equals 'w'.
12955 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12956 it exists mostly for use in the implementations of BN_is_zero(),
12957 BN_is_one(), and BN_is_word().
12961 * New function BN_swap.
12965 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12966 the exponentiation functions are more likely to produce reasonable
12967 results on negative inputs.
12971 * Change BN_mod_mul so that the result is always non-negative.
12972 Previously, it could be negative if one of the factors was negative;
12973 I don't think anyone really wanted that behaviour.
12977 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
12978 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
12979 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
12980 and add new functions:
12989 BN_mod_lshift1_quick
12991 BN_mod_lshift_quick
12993 These functions always generate non-negative results.
12995 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
12996 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
12998 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
12999 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13000 be reduced modulo `m`.
13002 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13005 The following entry accidentally appeared in the CHANGES file
13006 distributed with OpenSSL 0.9.7. The modifications described in
13007 it do *not* apply to OpenSSL 0.9.7.
13009 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13010 was actually never needed) and in BN_mul(). The removal in BN_mul()
13011 required a small change in bn_mul_part_recursive() and the addition
13012 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13013 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13014 bn_sub_words() and bn_add_words() except they take arrays with
13020 * In 'openssl passwd', verify passwords read from the terminal
13021 unless the '-salt' option is used (which usually means that
13022 verification would just waste user's time since the resulting
13023 hash is going to be compared with some given password hash)
13024 or the new '-noverify' option is used.
13026 This is an incompatible change, but it does not affect
13027 non-interactive use of 'openssl passwd' (passwords on the command
13028 line, '-stdin' option, '-in ...' option) and thus should not
13029 cause any problems.
13033 * Remove all references to RSAref, since there's no more need for it.
13037 * Make DSO load along a path given through an environment variable
13038 (SHLIB_PATH) with shl_load().
13042 * Constify the ENGINE code as a result of BIGNUM constification.
13043 Also constify the RSA code and most things related to it. In a
13044 few places, most notable in the depth of the ASN.1 code, ugly
13045 casts back to non-const were required (to be solved at a later
13050 * Make it so the openssl application has all engines loaded by default.
13054 * Constify the BIGNUM routines a little more.
13058 * Add the following functions:
13060 ENGINE_load_cswift()
13062 ENGINE_load_atalla()
13063 ENGINE_load_nuron()
13064 ENGINE_load_builtin_engines()
13066 That way, an application can itself choose if external engines that
13067 are built-in in OpenSSL shall ever be used or not. The benefit is
13068 that applications won't have to be linked with libdl or other dso
13069 libraries unless it's really needed.
13071 Changed 'openssl engine' to load all engines on demand.
13072 Changed the engine header files to avoid the duplication of some
13073 declarations (they differed!).
13077 * 'openssl engine' can now list capabilities.
13081 * Better error reporting in 'openssl engine'.
13085 * Never call load_dh_param(NULL) in s_server.
13089 * Add engine application. It can currently list engines by name and
13090 identity, and test if they are actually available.
13094 * Improve RPM specification file by forcing symbolic linking and making
13095 sure the installed documentation is also owned by root.root.
13097 *Damien Miller <djm@mindrot.org>*
13099 * Give the OpenSSL applications more possibilities to make use of
13100 keys (public as well as private) handled by engines.
13104 * Add OCSP code that comes from CertCo.
13108 * Add VMS support for the Rijndael code.
13112 * Added untested support for Nuron crypto accelerator.
13116 * Add support for external cryptographic devices. This code was
13117 previously distributed separately as the "engine" branch.
13119 *Geoff Thorpe, Richard Levitte*
13121 * Rework the filename-translation in the DSO code. It is now possible to
13122 have far greater control over how a "name" is turned into a filename
13123 depending on the operating environment and any oddities about the
13124 different shared library filenames on each system.
13128 * Support threads on FreeBSD-elf in Configure.
13132 * Fix for SHA1 assembly problem with MASM: it produces
13133 warnings about corrupt line number information when assembling
13134 with debugging information. This is caused by the overlapping
13137 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13140 NCONF_get_number() has no error checking at all. As a replacement,
13141 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13142 promoted strongly. The old NCONF_get_number is kept around for
13143 binary backward compatibility.
13144 Make it possible for methods to load from something other than a BIO,
13145 by providing a function pointer that is given a name instead of a BIO.
13146 For example, this could be used to load configuration data from an
13151 * Fix for non blocking accept BIOs. Added new I/O special reason
13152 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13153 with non blocking I/O was not possible because no retry code was
13154 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13159 * Added the beginnings of Rijndael support.
13163 * Fix for bug in DirectoryString mask setting. Add support for
13164 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13165 to allow certificate printing to more controllable, additional
13166 'certopt' option to 'x509' to allow new printing options to be
13171 * Clean old EAY MD5 hack from e_os.h.
13175 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13177 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13178 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13180 *Joe Orton, Steve Henson*
13182 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13184 * Fix additional bug revealed by the NISCC test suite:
13186 Stop bug triggering large recursion when presented with
13187 certain ASN.1 tags [CVE-2003-0851][]
13191 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13193 * Fix various bugs revealed by running the NISCC test suite:
13195 Stop out of bounds reads in the ASN1 code when presented with
13196 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13198 If verify callback ignores invalid public key errors don't try to check
13199 certificate signature with the NULL public key.
13203 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13204 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13209 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13210 extra data after the compression methods not only for TLS 1.0
13211 but also for SSL 3.0 (as required by the specification).
13213 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13215 * Change X509_certificate_type() to mark the key as exported/exportable
13216 when it's 512 *bits* long, not 512 bytes.
13220 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13222 * Countermeasure against the Klima-Pokorny-Rosa extension of
13223 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13224 a protocol version number mismatch like a decryption error
13225 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13229 * Turn on RSA blinding by default in the default implementation
13230 to avoid a timing attack. Applications that don't want it can call
13231 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13232 They would be ill-advised to do so in most cases.
13234 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13236 * Change RSA blinding code so that it works when the PRNG is not
13237 seeded (in this case, the secret RSA exponent is abused as
13238 an unpredictable seed -- if it is not unpredictable, there
13239 is no point in blinding anyway). Make RSA blinding thread-safe
13240 by remembering the creator's thread ID in rsa->blinding and
13241 having all other threads use local one-time blinding factors
13242 (this requires more computation than sharing rsa->blinding, but
13243 avoids excessive locking; and if an RSA object is not shared
13244 between threads, blinding will still be very fast).
13248 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13250 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13251 via timing by performing a MAC computation even if incorrect
13252 block cipher padding has been found. This is a countermeasure
13253 against active attacks where the attacker has to distinguish
13254 between bad padding and a MAC verification error. [CVE-2003-0078][]
13256 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13257 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13258 Martin Vuagnoux (EPFL, Ilion)*
13260 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13262 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13263 memory from its contents. This is done with a counter that will
13264 place alternating values in each byte. This can be used to solve
13265 two issues: 1) the removal of calls to memset() by highly optimizing
13266 compilers, and 2) cleansing with other values than 0, since those can
13267 be read through on certain media, for example a swap space on disk.
13271 * Bugfix: client side session caching did not work with external caching,
13272 because the session->cipher setting was not restored when reloading
13273 from the external cache. This problem was masked, when
13274 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13275 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13279 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13280 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13282 *Zeev Lieber <zeev-l@yahoo.com>*
13284 * Undo an undocumented change introduced in 0.9.6e which caused
13285 repeated calls to OpenSSL_add_all_ciphers() and
13286 OpenSSL_add_all_digests() to be ignored, even after calling
13291 * Change the default configuration reader to deal with last line not
13292 being properly terminated.
13296 * Change X509_NAME_cmp() so it applies the special rules on handling
13297 DN values that are of type PrintableString, as well as RDNs of type
13298 emailAddress where the value has the type ia5String.
13300 *stefank@valicert.com via Richard Levitte*
13302 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13303 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13304 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13305 the bitwise-OR of the two for use by the majority of applications
13306 wanting this behaviour, and update the docs. The documented
13307 behaviour and actual behaviour were inconsistent and had been
13308 changing anyway, so this is more a bug-fix than a behavioural
13311 *Geoff Thorpe, diagnosed by Nadav Har'El*
13313 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13314 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13318 * Fix initialization code race conditions in
13319 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13320 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13321 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13322 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13323 ssl2_get_cipher_by_char(),
13324 ssl3_get_cipher_by_char().
13326 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13328 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13329 the cached sessions are flushed, as the remove_cb() might use ex_data
13330 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13331 (see [openssl.org #212]).
13333 *Geoff Thorpe, Lutz Jaenicke*
13335 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13336 length, instead of the encoding length to d2i_ASN1_OBJECT.
13340 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13342 * [In 0.9.6g-engine release:]
13343 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13345 *Lynn Gazis <lgazis@rainbow.com>*
13347 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13349 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13350 and get fix the header length calculation.
13351 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13352 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13354 * Use proper error handling instead of 'assertions' in buffer
13355 overflow checks added in 0.9.6e. This prevents DoS (the
13356 assertions could call abort()).
13358 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13360 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13362 * Add various sanity checks to asn1_get_length() to reject
13363 the ASN1 length bytes if they exceed sizeof(long), will appear
13364 negative or the content length exceeds the length of the
13367 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13369 * Fix cipher selection routines: ciphers without encryption had no flags
13370 for the cipher strength set and where therefore not handled correctly
13371 by the selection routines (PR #130).
13375 * Fix EVP_dsa_sha macro.
13380 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13381 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13382 that was added in OpenSSL 0.9.6d.
13384 As the countermeasure turned out to be incompatible with some
13385 broken SSL implementations, the new option is part of SSL_OP_ALL.
13386 SSL_OP_ALL is usually employed when compatibility with weird SSL
13387 implementations is desired (e.g. '-bugs' option to 's_client' and
13388 's_server'), so the new option is automatically set in many
13393 * Changes in security patch:
13395 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13396 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13397 Air Force Materiel Command, USAF, under agreement number
13400 * Add various sanity checks to asn1_get_length() to reject
13401 the ASN1 length bytes if they exceed sizeof(long), will appear
13402 negative or the content length exceeds the length of the
13403 supplied buffer. [CVE-2002-0659][]
13405 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13407 * Assertions for various potential buffer overflows, not known to
13408 happen in practice.
13410 *Ben Laurie (CHATS)*
13412 * Various temporary buffers to hold ASCII versions of integers were
13413 too small for 64 bit platforms. [CVE-2002-0655][]
13414 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13416 * Remote buffer overflow in SSL3 protocol - an attacker could
13417 supply an oversized session ID to a client. [CVE-2002-0656][]
13419 *Ben Laurie (CHATS)*
13421 * Remote buffer overflow in SSL2 protocol - an attacker could
13422 supply an oversized client master key. [CVE-2002-0656][]
13424 *Ben Laurie (CHATS)*
13426 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13428 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13429 encoded as NULL) with id-dsa-with-sha1.
13431 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13433 * Check various `X509_...()` return values in apps/req.c.
13435 *Nils Larsch <nla@trustcenter.de>*
13437 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13438 an end-of-file condition would erroneously be flagged, when the CRLF
13439 was just at the end of a processed block. The bug was discovered when
13440 processing data through a buffering memory BIO handing the data to a
13441 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13442 <ptsekov@syntrex.com> and Nedelcho Stanev.
13446 * Implement a countermeasure against a vulnerability recently found
13447 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13448 before application data chunks to avoid the use of known IVs
13449 with data potentially chosen by the attacker.
13453 * Fix length checks in ssl3_get_client_hello().
13457 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13458 to prevent ssl3_read_internal() from incorrectly assuming that
13459 ssl3_read_bytes() found application data while handshake
13460 processing was enabled when in fact s->s3->in_read_app_data was
13461 merely automatically cleared during the initial handshake.
13463 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13465 * Fix object definitions for Private and Enterprise: they were not
13466 recognized in their shortname (=lowercase) representation. Extend
13467 obj_dat.pl to issue an error when using undefined keywords instead
13468 of silently ignoring the problem (Svenning Sorensen
13469 <sss@sss.dnsalias.net>).
13473 * Fix DH_generate_parameters() so that it works for 'non-standard'
13474 generators, i.e. generators other than 2 and 5. (Previously, the
13475 code did not properly initialise the 'add' and 'rem' values to
13476 BN_generate_prime().)
13478 In the new general case, we do not insist that 'generator' is
13479 actually a primitive root: This requirement is rather pointless;
13480 a generator of the order-q subgroup is just as good, if not
13485 * Map new X509 verification errors to alerts. Discovered and submitted by
13486 Tom Wu <tom@arcot.com>.
13490 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13491 returning non-zero before the data has been completely received
13492 when using non-blocking I/O.
13494 *Bodo Moeller; problem pointed out by John Hughes*
13496 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13498 *Ben Laurie, Lutz Jaenicke*
13500 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13501 Yoram Zahavi <YoramZ@gilian.com>).
13505 * Add information about CygWin 1.3 and on, and preserve proper
13506 configuration for the versions before that.
13508 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13510 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13511 check whether we deal with a copy of a session and do not delete from
13512 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13513 <izhar@checkpoint.com>.
13517 * Do not store session data into the internal session cache, if it
13518 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13519 flag is set). Proposed by Aslam <aslam@funk.com>.
13523 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13528 * [In 0.9.6d-engine release:]
13529 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13531 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13533 * Add the configuration target linux-s390x.
13535 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13537 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13538 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13539 variable as an indication that a ClientHello message has been
13540 received. As the flag value will be lost between multiple
13541 invocations of ssl3_accept when using non-blocking I/O, the
13542 function may not be aware that a handshake has actually taken
13543 place, thus preventing a new session from being added to the
13546 To avoid this problem, we now set s->new_session to 2 instead of
13547 using a local variable.
13549 *Lutz Jaenicke, Bodo Moeller*
13551 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13552 if the SSL_R_LENGTH_MISMATCH error is detected.
13554 *Geoff Thorpe, Bodo Moeller*
13556 * New 'shared_ldflag' column in Configure platform table.
13560 * Fix EVP_CIPHER_mode macro.
13562 *"Dan S. Camper" <dan@bti.net>*
13564 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13565 type, we must throw them away by setting rr->length to 0.
13567 *D P Chang <dpc@qualys.com>*
13569 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13571 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13572 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13573 worked incorrectly for those cases where range = `10..._2` and
13574 `3*range` is two bits longer than range.)
13578 * Only add signing time to PKCS7 structures if it is not already
13583 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13584 OBJ_ld_ce should be OBJ_id_ce.
13585 Also some ip-pda OIDs in crypto/objects/objects.txt were
13586 incorrect (cf. RFC 3039).
13588 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13590 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13591 returns early because it has nothing to do.
13593 *Andy Schneider <andy.schneider@bjss.co.uk>*
13595 * [In 0.9.6c-engine release:]
13596 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13598 *Andy Schneider <andy.schneider@bjss.co.uk>*
13600 * [In 0.9.6c-engine release:]
13601 Add support for Cryptographic Appliance's keyserver technology.
13602 (Use engine 'keyclient')
13604 *Cryptographic Appliances and Geoff Thorpe*
13606 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13607 is called via tools/c89.sh because arguments have to be
13608 rearranged (all '-L' options must appear before the first object
13611 *Richard Shapiro <rshapiro@abinitio.com>*
13613 * [In 0.9.6c-engine release:]
13614 Add support for Broadcom crypto accelerator cards, backported
13617 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13619 * [In 0.9.6c-engine release:]
13620 Add support for SureWare crypto accelerator cards from
13621 Baltimore Technologies. (Use engine 'sureware')
13623 *Baltimore Technologies and Mark Cox*
13625 * [In 0.9.6c-engine release:]
13626 Add support for crypto accelerator cards from Accelerated
13627 Encryption Processing, www.aep.ie. (Use engine 'aep')
13629 *AEP Inc. and Mark Cox*
13631 * Add a configuration entry for gcc on UnixWare.
13633 *Gary Benson <gbenson@redhat.com>*
13635 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13636 messages are stored in a single piece (fixed-length part and
13637 variable-length part combined) and fix various bugs found on the way.
13641 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13642 instead. BIO_gethostbyname() does not know what timeouts are
13643 appropriate, so entries would stay in cache even when they have
13645 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13647 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13648 faced with a pathologically small ClientHello fragment that does
13649 not contain client_version: Instead of aborting with an error,
13650 simply choose the highest available protocol version (i.e.,
13651 TLS 1.0 unless it is disabled). In practice, ClientHello
13652 messages are never sent like this, but this change gives us
13653 strictly correct behaviour at least for TLS.
13657 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13658 never resets s->method to s->ctx->method when called from within
13659 one of the SSL handshake functions.
13661 *Bodo Moeller; problem pointed out by Niko Baric*
13663 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13664 (sent using the client's version number) if client_version is
13665 smaller than the protocol version in use. Also change
13666 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13667 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13668 the client will at least see that alert.
13672 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13677 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13678 client receives HelloRequest while in a handshake.
13680 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13682 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13683 should end in 'break', not 'goto end' which circumvents various
13684 cleanups done in state SSL_ST_OK. But session related stuff
13685 must be disabled for SSL_ST_OK in the case that we just sent a
13688 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13689 before just sending a HelloRequest.
13691 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13693 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13694 reveal whether illegal block cipher padding was found or a MAC
13695 verification error occurred. (Neither SSLerr() codes nor alerts
13696 are directly visible to potential attackers, but the information
13697 may leak via logfiles.)
13699 Similar changes are not required for the SSL 2.0 implementation
13700 because the number of padding bytes is sent in clear for SSL 2.0,
13701 and the extra bytes are just ignored. However ssl/s2_pkt.c
13702 failed to verify that the purported number of padding bytes is in
13707 * Add OpenUNIX-8 support including shared libraries
13708 (Boyd Lynn Gerber <gerberb@zenez.com>).
13712 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13713 'wristwatch attack' using huge encoding parameters (cf.
13714 James H. Manger's CRYPTO 2001 paper). Note that the
13715 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13716 encoding parameters and hence was not vulnerable.
13720 * BN_sqr() bug fix.
13722 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13724 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13725 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13726 followed by modular reduction.
13728 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13730 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13731 equivalent based on BN_pseudo_rand() instead of BN_rand().
13735 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13736 This function was broken, as the check for a new client hello message
13737 to handle SGC did not allow these large messages.
13738 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13742 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13746 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13747 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13751 * Rework the configuration and shared library support for Tru64 Unix.
13752 The configuration part makes use of modern compiler features and
13753 still retains old compiler behavior for those that run older versions
13754 of the OS. The shared library support part includes a variant that
13755 uses the RPATH feature, and is available through the special
13756 configuration target "alpha-cc-rpath", which will never be selected
13759 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13761 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13762 with the same message size as in ssl3_get_certificate_request().
13763 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13764 messages might inadvertently be reject as too long.
13766 *Petr Lampa <lampa@fee.vutbr.cz>*
13768 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13772 * Modified SSL library such that the verify_callback that has been set
13773 specifically for an SSL object with SSL_set_verify() is actually being
13774 used. Before the change, a verify_callback set with this function was
13775 ignored and the verify_callback() set in the SSL_CTX at the time of
13776 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13777 to allow the necessary settings.
13781 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13782 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13783 done automatically (in contradiction to the requirements of the C
13784 standard). This made problems when used from OpenSSH.
13788 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13789 dh->length and always used
13791 BN_rand_range(priv_key, dh->p).
13793 BN_rand_range() is not necessary for Diffie-Hellman, and this
13794 specific range makes Diffie-Hellman unnecessarily inefficient if
13795 dh->length (recommended exponent length) is much smaller than the
13796 length of dh->p. We could use BN_rand_range() if the order of
13797 the subgroup was stored in the DH structure, but we only have
13802 BN_rand(priv_key, l, ...)
13804 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13811 RSA_eay_public_encrypt
13812 RSA_eay_private_decrypt
13813 RSA_eay_private_encrypt (signing)
13814 RSA_eay_public_decrypt (signature verification)
13816 (default implementations for RSA_public_encrypt,
13817 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13818 always reject numbers >= n.
13822 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13823 to synchronize access to 'locking_thread'. This is necessary on
13824 systems where access to 'locking_thread' (an 'unsigned long'
13825 variable) is not atomic.
13829 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13830 *before* setting the 'crypto_lock_rand' flag. The previous code had
13831 a race condition if 0 is a valid thread ID.
13833 *Travis Vitek <vitek@roguewave.com>*
13835 * Add support for shared libraries under Irix.
13837 *Albert Chin-A-Young <china@thewrittenword.com>*
13839 * Add configuration option to build on Linux on both big-endian and
13840 little-endian MIPS.
13842 *Ralf Baechle <ralf@uni-koblenz.de>*
13844 * Add the possibility to create shared libraries on HP-UX.
13848 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13850 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13851 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13852 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13853 PRNG state recovery was possible based on the output of
13854 one PRNG request appropriately sized to gain knowledge on
13855 'md' followed by enough consecutive 1-byte PRNG requests
13856 to traverse all of 'state'.
13858 1. When updating 'md_local' (the current thread's copy of 'md')
13859 during PRNG output generation, hash all of the previous
13860 'md_local' value, not just the half used for PRNG output.
13862 2. Make the number of bytes from 'state' included into the hash
13863 independent from the number of PRNG bytes requested.
13865 The first measure alone would be sufficient to avoid
13866 Markku-Juhani's attack. (Actually it had never occurred
13867 to me that the half of 'md_local' used for chaining was the
13868 half from which PRNG output bytes were taken -- I had always
13869 assumed that the secret half would be used.) The second
13870 measure makes sure that additional data from 'state' is never
13871 mixed into 'md_local' in small portions; this heuristically
13872 further strengthens the PRNG.
13876 * Fix crypto/bn/asm/mips3.s.
13880 * When only the key is given to "enc", the IV is undefined. Print out
13881 an error message in this case.
13885 * Handle special case when X509_NAME is empty in X509 printing routines.
13889 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13890 positive and less than q.
13894 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13895 used: it isn't thread safe and the add_lock_callback should handle
13898 *Paul Rose <Paul.Rose@bridge.com>*
13900 * Verify that incoming data obeys the block size in
13901 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13907 *Ulf Möller, Bodo Möller*
13909 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13910 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13911 when fixing the server behaviour for backwards-compatible 'client
13912 hello' messages. (Note that the attack is impractical against
13913 SSL 3.0 and TLS 1.0 anyway because length and version checking
13914 means that the probability of guessing a valid ciphertext is
13915 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13918 Before 0.9.5, the countermeasure (hide the error by generating a
13919 random 'decryption result') did not work properly because
13920 ERR_clear_error() was missing, meaning that SSL_get_error() would
13921 detect the supposedly ignored error.
13923 Both problems are now fixed.
13927 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13928 (previously it was 1024).
13932 * Fix for compatibility mode trust settings: ignore trust settings
13933 unless some valid trust or reject settings are present.
13937 * Fix for blowfish EVP: its a variable length cipher.
13941 * Fix various bugs related to DSA S/MIME verification. Handle missing
13942 parameters in DSA public key structures and return an error in the
13943 DSA routines if parameters are absent.
13947 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13948 in the current directory if neither $RANDFILE nor $HOME was set.
13949 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13950 caused some confusion to Windows users who haven't defined $HOME.
13951 Thus RAND_file_name() is changed again: e_os.h can define a
13952 DEFAULT_HOME, which will be used if $HOME is not set.
13953 For Windows, we use "C:"; on other platforms, we still require
13954 environment variables.
13956 * Move 'if (!initialized) RAND_poll()' into regions protected by
13957 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13958 having multiple threads call RAND_poll() concurrently.
13962 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13963 combination of a flag and a thread ID variable.
13964 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13965 flag), *other* threads can enter ssleay_add_bytes without obeying
13966 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13967 that they do not hold after the first thread unsets add_do_not_lock).
13971 * Change bctest again: '-x' expressions are not available in all
13972 versions of 'test'.
13976 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13978 * Fix a couple of memory leaks in PKCS7_dataDecode()
13980 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13982 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13983 the default extension for executables, if any. Also, make the perl
13984 scripts that use symlink() to test if it really exists and use "cp"
13985 if it doesn't. All this made OpenSSL compilable and installable in
13990 * Fix for asn1_GetSequence() for indefinite length constructed data.
13991 If SEQUENCE is length is indefinite just set c->slen to the total
13992 amount of data available.
13994 *Steve Henson, reported by shige@FreeBSD.org*
13996 *This change does not apply to 0.9.7.*
13998 * Change bctest to avoid here-documents inside command substitution
13999 (workaround for FreeBSD /bin/sh bug).
14000 For compatibility with Ultrix, avoid shell functions (introduced
14001 in the bctest version that searches along $PATH).
14005 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14006 with des_encrypt() defined on some operating systems, like Solaris
14011 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14012 On the Importance of Eliminating Errors in Cryptographic
14013 Computations, J. Cryptology 14 (2001) 2, 101-119,
14014 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14018 * MIPS assembler BIGNUM division bug fix.
14022 * Disabled incorrect Alpha assembler code.
14026 * Fix PKCS#7 decode routines so they correctly update the length
14027 after reading an EOC for the EXPLICIT tag.
14031 *This change does not apply to 0.9.7.*
14033 * Fix bug in PKCS#12 key generation routines. This was triggered
14034 if a 3DES key was generated with a 0 initial byte. Include
14035 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14036 (but broken) behaviour.
14040 * Enhance bctest to search for a working bc along $PATH and print
14043 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14045 * Fix memory leaks in err.c: free err_data string if necessary;
14046 don't write to the wrong index in ERR_set_error_data.
14050 * Implement ssl23_peek (analogous to ssl23_read), which previously
14055 * Replace rdtsc with `_emit` statements for VC++ version 5.
14057 *Jeremy Cooper <jeremy@baymoo.org>*
14059 * Make it possible to reuse SSLv2 sessions.
14063 * In copy_email() check for >= 0 as a return value for
14064 X509_NAME_get_index_by_NID() since 0 is a valid index.
14066 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14068 * Avoid coredump with unsupported or invalid public keys by checking if
14069 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14070 PKCS7_verify() fails with non detached data.
14074 * Don't use getenv in library functions when run as setuid/setgid.
14075 New function OPENSSL_issetugid().
14079 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14080 due to incorrect handling of multi-threading:
14082 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14084 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14086 3. Count how many times MemCheck_off() has been called so that
14087 nested use can be treated correctly. This also avoids
14088 inband-signalling in the previous code (which relied on the
14089 assumption that thread ID 0 is impossible).
14093 * Add "-rand" option also to s_client and s_server.
14097 * Fix CPU detection on Irix 6.x.
14098 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14099 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14101 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14106 *This change does not apply to 0.9.7.*
14108 * Use the cached encoding of an X509_NAME structure rather than
14109 copying it. This is apparently the reason for the libsafe "errors"
14110 but the code is actually correct.
14114 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14115 Bleichenbacher's DSA attack.
14116 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14117 to be set and top=0 forces the highest bit to be set; top=-1 is new
14118 and leaves the highest bit random.
14120 *Ulf Moeller, Bodo Moeller*
14122 * In the `NCONF_...`-based implementations for `CONF_...` queries
14123 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14124 a temporary CONF structure with the data component set to NULL
14125 (which gives segmentation faults in lh_retrieve).
14126 Instead, use NULL for the CONF pointer in CONF_get_string and
14127 CONF_get_number (which may use environment variables) and directly
14128 return NULL from CONF_get_section.
14132 * Fix potential buffer overrun for EBCDIC.
14136 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14137 keyUsage if basicConstraints absent for a CA.
14141 * Make SMIME_write_PKCS7() write mail header values with a format that
14142 is more generally accepted (no spaces before the semicolon), since
14143 some programs can't parse those values properly otherwise. Also make
14144 sure BIO's that break lines after each write do not create invalid
14149 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14150 macros previously used would not encode an empty SEQUENCE OF
14151 and break the signature.
14155 *This change does not apply to 0.9.7.*
14157 * Zero the premaster secret after deriving the master secret in
14162 * Add some EVP_add_digest_alias registrations (as found in
14163 OpenSSL_add_all_digests()) to SSL_library_init()
14164 aka OpenSSL_add_ssl_algorithms(). This provides improved
14165 compatibility with peers using X.509 certificates
14166 with unconventional AlgorithmIdentifier OIDs.
14170 * Fix for Irix with NO_ASM.
14172 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14174 * ./config script fixes.
14176 *Ulf Moeller, Richard Levitte*
14178 * Fix 'openssl passwd -1'.
14182 * Change PKCS12_key_gen_asc() so it can cope with non null
14183 terminated strings whose length is passed in the passlen
14184 parameter, for example from PEM callbacks. This was done
14185 by adding an extra length parameter to asc2uni().
14187 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14189 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14190 call failed, free the DSA structure.
14194 * Fix to uni2asc() to cope with zero length Unicode strings.
14195 These are present in some PKCS#12 files.
14199 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14200 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14201 when writing a 32767 byte record.
14203 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14205 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14206 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14208 (RSA objects have a reference count access to which is protected
14209 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14210 so they are meant to be shared between threads.)
14211 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14212 "Reddie, Steven" <Steven.Reddie@ca.com>*
14214 * Fix a deadlock in CRYPTO_mem_leaks().
14218 * Use better test patterns in bntest.
14222 * rand_win.c fix for Borland C.
14226 * BN_rshift bugfix for n == 0.
14230 * Add a 'bctest' script that checks for some known 'bc' bugs
14231 so that 'make test' does not abort just because 'bc' is broken.
14235 * Store verify_result within SSL_SESSION also for client side to
14236 avoid potential security hole. (Re-used sessions on the client side
14237 always resulted in verify_result==X509_V_OK, not using the original
14238 result of the server certificate verification.)
14242 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14243 SSL3_RT_APPLICATION_DATA, return 0.
14244 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14249 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14250 releases, have been re-implemented by renaming the previous
14251 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14252 and ssl3_read_internal, respectively, and adding 'peek' parameters
14253 to them. The new ssl[23]_{read,peek} functions are calls to
14254 ssl[23]_read_internal with the 'peek' flag set appropriately.
14255 A 'peek' parameter has also been added to ssl3_read_bytes, which
14256 does the actual work for ssl3_read_internal.
14260 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14261 the method-specific "init()" handler. Also clean up ex_data after
14262 calling the method-specific "finish()" handler. Previously, this was
14263 happening the other way round.
14267 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14268 The previous value, 12, was not always sufficient for BN_mod_exp().
14272 * Make sure that shared libraries get the internal name engine with
14273 the full version number and not just 0. This should mark the
14274 shared libraries as not backward compatible. Of course, this should
14275 be changed again when we can guarantee backward binary compatibility.
14279 * Fix typo in get_cert_by_subject() in by_dir.c
14281 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14283 * Rework the system to generate shared libraries:
14285 - Make note of the expected extension for the shared libraries and
14286 if there is a need for symbolic links from for example libcrypto.so.0
14287 to libcrypto.so.0.9.7. There is extended info in Configure for
14290 - Make as few rebuilds of the shared libraries as possible.
14292 - Still avoid linking the OpenSSL programs with the shared libraries.
14294 - When installing, install the shared libraries separately from the
14299 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14301 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14302 and not in SSL_clear because the latter is also used by the
14303 accept/connect functions; previously, the settings made by
14304 SSL_set_read_ahead would be lost during the handshake.
14306 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14308 * Correct util/mkdef.pl to be selective about disabled algorithms.
14309 Previously, it would create entries for disabled algorithms no
14314 * Added several new manual pages for SSL_* function.
14318 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14320 * In ssl23_get_client_hello, generate an error message when faced
14321 with an initial SSL 3.0/TLS record that is too small to contain the
14322 first two bytes of the ClientHello message, i.e. client_version.
14323 (Note that this is a pathologic case that probably has never happened
14324 in real life.) The previous approach was to use the version number
14325 from the record header as a substitute; but our protocol choice
14326 should not depend on that one because it is not authenticated
14327 by the Finished messages.
14331 * More robust randomness gathering functions for Windows.
14333 *Jeffrey Altman <jaltman@columbia.edu>*
14335 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14336 not set then we don't setup the error code for issuer check errors
14337 to avoid possibly overwriting other errors which the callback does
14338 handle. If an application does set the flag then we assume it knows
14339 what it is doing and can handle the new informational codes
14344 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14345 a general "ANY" type, as such it should be able to decode anything
14346 including tagged types. However it didn't check the class so it would
14347 wrongly interpret tagged types in the same way as their universal
14348 counterpart and unknown types were just rejected. Changed so that the
14349 tagged and unknown types are handled in the same way as a SEQUENCE:
14350 that is the encoding is stored intact. There is also a new type
14351 "V_ASN1_OTHER" which is used when the class is not universal, in this
14352 case we have no idea what the actual type is so we just lump them all
14357 * On VMS, stdout may very well lead to a file that is written to
14358 in a record-oriented fashion. That means that every write() will
14359 write a separate record, which will be read separately by the
14360 programs trying to read from it. This can be very confusing.
14362 The solution is to put a BIO filter in the way that will buffer
14363 text until a linefeed is reached, and then write everything a
14364 line at a time, so every record written will be an actual line,
14365 not chunks of lines and not (usually doesn't happen, but I've
14366 seen it once) several lines in one record. BIO_f_linebuffer() is
14369 Currently, it's a VMS-only method, because that's where it has
14370 been tested well enough.
14374 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14375 it can return incorrect results.
14376 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14377 but it was in 0.9.6-beta[12].)
14381 * Disable the check for content being present when verifying detached
14382 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14383 include zero length content when signing messages.
14387 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14388 BIO_ctrl (for BIO pairs).
14392 * Add DSO method for VMS.
14396 * Bug fix: Montgomery multiplication could produce results with the
14401 * Add RPM specification openssl.spec and modify it to build three
14402 packages. The default package contains applications, application
14403 documentation and run-time libraries. The devel package contains
14404 include files, static libraries and function documentation. The
14405 doc package contains the contents of the doc directory. The original
14406 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14410 * Add a large number of documentation files for many SSL routines.
14412 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14414 * Add a configuration entry for Sony News 4.
14416 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14418 * Don't set the two most significant bits to one when generating a
14419 random number < q in the DSA library.
14423 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14424 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14425 the underlying transport is blocking) if a handshake took place.
14426 (The default behaviour is needed by applications such as s_client
14427 and s_server that use select() to determine when to use SSL_read;
14428 but for applications that know in advance when to expect data, it
14429 just makes things more complicated.)
14433 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14438 * Add a few more EBCDIC conditionals that make `req` and `x509`
14439 work better on such systems.
14441 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14443 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14444 Update PKCS12_parse() so it copies the friendlyName and the
14445 keyid to the certificates aux info.
14449 * Fix bug in PKCS7_verify() which caused an infinite loop
14450 if there was more than one signature.
14452 *Sven Uszpelkat <su@celocom.de>*
14454 * Major change in util/mkdef.pl to include extra information
14455 about each symbol, as well as presenting variables as well
14456 as functions. This change means that there's n more need
14457 to rebuild the .num files when some algorithms are excluded.
14461 * Allow the verify time to be set by an application,
14462 rather than always using the current time.
14466 * Phase 2 verify code reorganisation. The certificate
14467 verify code now looks up an issuer certificate by a
14468 number of criteria: subject name, authority key id
14469 and key usage. It also verifies self signed certificates
14470 by the same criteria. The main comparison function is
14471 X509_check_issued() which performs these checks.
14473 Lot of changes were necessary in order to support this
14474 without completely rewriting the lookup code.
14476 Authority and subject key identifier are now cached.
14478 The LHASH 'certs' is X509_STORE has now been replaced
14479 by a STACK_OF(X509_OBJECT). This is mainly because an
14480 LHASH can't store or retrieve multiple objects with
14481 the same hash value.
14483 As a result various functions (which were all internal
14484 use only) have changed to handle the new X509_STORE
14485 structure. This will break anything that messed round
14486 with X509_STORE internally.
14488 The functions X509_STORE_add_cert() now checks for an
14489 exact match, rather than just subject name.
14491 The X509_STORE API doesn't directly support the retrieval
14492 of multiple certificates matching a given criteria, however
14493 this can be worked round by performing a lookup first
14494 (which will fill the cache with candidate certificates)
14495 and then examining the cache for matches. This is probably
14496 the best we can do without throwing out X509_LOOKUP
14497 entirely (maybe later...).
14499 The X509_VERIFY_CTX structure has been enhanced considerably.
14501 All certificate lookup operations now go via a get_issuer()
14502 callback. Although this currently uses an X509_STORE it
14503 can be replaced by custom lookups. This is a simple way
14504 to bypass the X509_STORE hackery necessary to make this
14505 work and makes it possible to use more efficient techniques
14506 in future. A very simple version which uses a simple
14507 STACK for its trusted certificate store is also provided
14508 using X509_STORE_CTX_trusted_stack().
14510 The verify_cb() and verify() callbacks now have equivalents
14511 in the X509_STORE_CTX structure.
14513 X509_STORE_CTX also has a 'flags' field which can be used
14514 to customise the verify behaviour.
14518 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14519 excludes S/MIME capabilities.
14523 * When a certificate request is read in keep a copy of the
14524 original encoding of the signed data and use it when outputting
14525 again. Signatures then use the original encoding rather than
14526 a decoded, encoded version which may cause problems if the
14527 request is improperly encoded.
14531 * For consistency with other BIO_puts implementations, call
14532 buffer_write(b, ...) directly in buffer_puts instead of calling
14535 In BIO_puts, increment b->num_write as in BIO_write.
14537 *Peter.Sylvester@EdelWeb.fr*
14539 * Fix BN_mul_word for the case where the word is 0. (We have to use
14540 BN_zero, we may not return a BIGNUM with an array consisting of
14541 words set to zero.)
14545 * Avoid calling abort() from within the library when problems are
14546 detected, except if preprocessor symbols have been defined
14547 (such as REF_CHECK, BN_DEBUG etc.).
14551 * New openssl application 'rsautl'. This utility can be
14552 used for low level RSA operations. DER public key
14553 BIO/fp routines also added.
14557 * New Configure entry and patches for compiling on QNX 4.
14559 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14561 * A demo state-machine implementation was sponsored by
14562 Nuron (<http://www.nuron.com/>) and is now available in
14563 demos/state_machine.
14567 * New options added to the 'dgst' utility for signature
14568 generation and verification.
14572 * Unrecognized PKCS#7 content types are now handled via a
14573 catch all ASN1_TYPE structure. This allows unsupported
14574 types to be stored as a "blob" and an application can
14575 encode and decode it manually.
14579 * Fix various signed/unsigned issues to make a_strex.c
14580 compile under VC++.
14582 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14584 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14585 length if passed a buffer. ASN1_INTEGER_to_BN failed
14586 if passed a NULL BN and its argument was negative.
14588 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14590 * Modification to PKCS#7 encoding routines to output definite
14591 length encoding. Since currently the whole structures are in
14592 memory there's not real point in using indefinite length
14593 constructed encoding. However if OpenSSL is compiled with
14594 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14598 * Added BIO_vprintf() and BIO_vsnprintf().
14602 * Added more prefixes to parse for in the strings written
14603 through a logging bio, to cover all the levels that are available
14604 through syslog. The prefixes are now:
14606 PANIC, EMERG, EMR => LOG_EMERG
14607 ALERT, ALR => LOG_ALERT
14608 CRIT, CRI => LOG_CRIT
14609 ERROR, ERR => LOG_ERR
14610 WARNING, WARN, WAR => LOG_WARNING
14611 NOTICE, NOTE, NOT => LOG_NOTICE
14612 INFO, INF => LOG_INFO
14613 DEBUG, DBG => LOG_DEBUG
14615 and as before, if none of those prefixes are present at the
14616 beginning of the string, LOG_ERR is chosen.
14618 On Win32, the `LOG_*` levels are mapped according to this:
14620 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14621 LOG_WARNING => EVENTLOG_WARNING_TYPE
14622 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14626 * Made it possible to reconfigure with just the configuration
14627 argument "reconf" or "reconfigure". The command line arguments
14628 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14629 and are retrieved from there when reconfiguring.
14635 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14637 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14641 * The obj_dat.pl script was messing up the sorting of object
14642 names. The reason was that it compared the quoted version
14643 of strings as a result "OCSP" > "OCSP Signing" because
14644 " > SPACE. Changed script to store unquoted versions of
14645 names and add quotes on output. It was also omitting some
14646 names from the lookup table if they were given a default
14647 value (that is if SN is missing it is given the same
14648 value as LN and vice versa), these are now added on the
14649 grounds that if an object has a name we should be able to
14650 look it up. Finally added warning output when duplicate
14651 short or long names are found.
14655 * Changes needed for Tandem NSK.
14657 *Scott Uroff <scott@xypro.com>*
14659 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14660 RSA_padding_check_SSLv23(), special padding was never detected
14661 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14662 version rollback attacks was not effective.
14664 In s23_clnt.c, don't use special rollback-attack detection padding
14665 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14666 client; similarly, in s23_srvr.c, don't do the rollback check if
14667 SSL 2.0 is the only protocol enabled in the server.
14671 * Make it possible to get hexdumps of unprintable data with 'openssl
14672 asn1parse'. By implication, the functions ASN1_parse_dump() and
14673 BIO_dump_indent() are added.
14677 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14678 these print out strings and name structures based on various
14679 flags including RFC2253 support and proper handling of
14680 multibyte characters. Added options to the 'x509' utility
14681 to allow the various flags to be set.
14685 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14686 Also change the functions X509_cmp_current_time() and
14687 X509_gmtime_adj() work with an ASN1_TIME structure,
14688 this will enable certificates using GeneralizedTime in validity
14689 dates to be checked.
14693 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14694 negative public key encodings) on by default,
14695 NO_NEG_PUBKEY_BUG can be set to disable it.
14699 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14700 content octets. An i2c_ASN1_OBJECT is unnecessary because
14701 the encoding can be trivially obtained from the structure.
14705 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14706 not read locks (`CRYPTO_r_[un]lock`).
14710 * A first attempt at creating official support for shared
14711 libraries through configuration. I've kept it so the
14712 default is static libraries only, and the OpenSSL programs
14713 are always statically linked for now, but there are
14714 preparations for dynamic linking in place.
14715 This has been tested on Linux and Tru64.
14719 * Randomness polling function for Win9x, as described in:
14720 Peter Gutmann, Software Generation of Practically Strong
14725 * Fix so PRNG is seeded in req if using an already existing
14730 * New options to smime application. -inform and -outform
14731 allow alternative formats for the S/MIME message including
14732 PEM and DER. The -content option allows the content to be
14733 specified separately. This should allow things like Netscape
14734 form signing output easier to verify.
14738 * Fix the ASN1 encoding of tags using the 'long form'.
14742 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14743 STRING types. These convert content octets to and from the
14744 underlying type. The actual tag and length octets are
14745 already assumed to have been read in and checked. These
14746 are needed because all other string types have virtually
14747 identical handling apart from the tag. By having versions
14748 of the ASN1 functions that just operate on content octets
14749 IMPLICIT tagging can be handled properly. It also allows
14750 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14751 and ASN1_INTEGER are identical apart from the tag.
14755 * Change the handling of OID objects as follows:
14757 - New object identifiers are inserted in objects.txt, following
14758 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14759 - objects.pl is used to process obj_mac.num and create a new
14761 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14764 This is currently kind of a hack, and the perl code in objects.pl
14765 isn't very elegant, but it works as I intended. The simplest way
14766 to check that it worked correctly is to look in obj_dat.h and
14767 check the array nid_objs and make sure the objects haven't moved
14768 around (this is important!). Additions are OK, as well as
14769 consistent name changes.
14773 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14777 * Addition of the command line parameter '-rand file' to 'openssl req'.
14778 The given file adds to whatever has already been seeded into the
14779 random pool through the RANDFILE configuration file option or
14780 environment variable, or the default random state file.
14784 * mkstack.pl now sorts each macro group into lexical order.
14785 Previously the output order depended on the order the files
14786 appeared in the directory, resulting in needless rewriting
14791 * Patches to make OpenSSL compile under Win32 again. Mostly
14792 work arounds for the VC++ problem that it treats func() as
14793 func(void). Also stripped out the parts of mkdef.pl that
14794 added extra typesafe functions: these no longer exist.
14798 * Reorganisation of the stack code. The macros are now all
14799 collected in safestack.h . Each macro is defined in terms of
14800 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14801 DEBUG_SAFESTACK is now handled in terms of function casts,
14802 this has the advantage of retaining type safety without the
14803 use of additional functions. If DEBUG_SAFESTACK is not defined
14804 then the non typesafe macros are used instead. Also modified the
14805 mkstack.pl script to handle the new form. Needs testing to see
14806 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14807 the default if no major problems. Similar behaviour for ASN1_SET_OF
14808 and PKCS12_STACK_OF.
14812 * When some versions of IIS use the 'NET' form of private key the
14813 key derivation algorithm is different. Normally MD5(password) is
14814 used as a 128 bit RC4 key. In the modified case
14815 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14816 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14817 as the old Netscape_RSA functions except they have an additional
14818 'sgckey' parameter which uses the modified algorithm. Also added
14819 an -sgckey command line option to the rsa utility. Thanks to
14820 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14821 algorithm to openssl-dev.
14825 * The evp_local.h macros were using 'c.##kname' which resulted in
14826 invalid expansion on some systems (SCO 5.0.5 for example).
14827 Corrected to 'c.kname'.
14829 *Phillip Porch <root@theporch.com>*
14831 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14832 a STACK of email addresses from a certificate or request, these look
14833 in the subject name and the subject alternative name extensions and
14834 omit any duplicate addresses.
14838 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14839 This makes DSA verification about 2 % faster.
14843 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14844 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14845 plus overhead for 1024 bit moduli).
14846 This makes exponentiations about 0.5 % faster for 1024 bit
14847 exponents (as measured by "openssl speed rsa2048").
14851 * Rename memory handling macros to avoid conflicts with other
14853 Malloc => OPENSSL_malloc
14854 Malloc_locked => OPENSSL_malloc_locked
14855 Realloc => OPENSSL_realloc
14856 Free => OPENSSL_free
14860 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14861 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14865 * CygWin32 support.
14867 *John Jarvie <jjarvie@newsguy.com>*
14869 * The type-safe stack code has been rejigged. It is now only compiled
14870 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14871 by default all type-specific stack functions are "#define"d back to
14872 standard stack functions. This results in more streamlined output
14873 but retains the type-safety checking possibilities of the original
14878 * The STACK code has been cleaned up, and certain type declarations
14879 that didn't make a lot of sense have been brought in line. This has
14880 also involved a cleanup of sorts in safestack.h to more correctly
14881 map type-safe stack functions onto their plain stack counterparts.
14882 This work has also resulted in a variety of "const"ifications of
14883 lots of the code, especially `_cmp` operations which should normally
14884 be prototyped with "const" parameters anyway.
14888 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14889 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14890 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14891 where all of 'md' is used each time the PRNG is used, but 'state'
14892 is used only indexed by a cyclic counter. As entropy may not be
14893 well distributed from the beginning, 'md' is important as a
14894 chaining variable. However, the output function chains only half
14895 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14896 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14897 in all of 'state' being rewritten, with the new values depending
14898 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14902 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14903 the handshake is continued after ssl_verify_cert_chain();
14904 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14905 can lead to 'unexplainable' connection aborts later.
14907 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14909 * Major EVP API cipher revision.
14910 Add hooks for extra EVP features. This allows various cipher
14911 parameters to be set in the EVP interface. Support added for variable
14912 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14913 setting of RC2 and RC5 parameters.
14915 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14918 Remove lots of duplicated code from the EVP library. For example *every*
14919 cipher init() function handles the 'iv' in the same way according to the
14920 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14921 for CFB and OFB modes they zero ctx->num.
14923 New functionality allows removal of S/MIME code RC2 hack.
14925 Most of the routines have the same form and so can be declared in terms
14928 By shifting this to the top level EVP_CipherInit() it can be removed from
14929 all individual ciphers. If the cipher wants to handle IVs or keys
14930 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14933 Change lots of functions like EVP_EncryptUpdate() to now return a
14934 value: although software versions of the algorithms cannot fail
14935 any installed hardware versions can.
14939 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14940 this option is set, tolerate broken clients that send the negotiated
14941 protocol version number instead of the requested protocol version
14946 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14947 i.e. non-zero for export ciphersuites, zero otherwise.
14948 Previous versions had this flag inverted, inconsistent with
14949 rsa_tmp_cb (..._TMP_RSA_CB).
14951 *Bodo Moeller; problem reported by Amit Chopra*
14953 * Add missing DSA library text string. Work around for some IIS
14954 key files with invalid SEQUENCE encoding.
14958 * Add a document (doc/standards.txt) that list all kinds of standards
14959 and so on that are implemented in OpenSSL.
14963 * Enhance c_rehash script. Old version would mishandle certificates
14964 with the same subject name hash and wouldn't handle CRLs at all.
14965 Added -fingerprint option to crl utility, to support new c_rehash
14970 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14974 * Fix for SSL server purpose checking. Server checking was
14975 rejecting certificates which had extended key usage present
14976 but no ssl client purpose.
14978 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14980 * Make PKCS#12 code work with no password. The PKCS#12 spec
14981 is a little unclear about how a blank password is handled.
14982 Since the password in encoded as a BMPString with terminating
14983 double NULL a zero length password would end up as just the
14984 double NULL. However no password at all is different and is
14985 handled differently in the PKCS#12 key generation code. NS
14986 treats a blank password as zero length. MSIE treats it as no
14987 password on export: but it will try both on import. We now do
14988 the same: PKCS12_parse() tries zero length and no password if
14989 the password is set to "" or NULL (NULL is now a valid password:
14990 it wasn't before) as does the pkcs12 application.
14994 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14995 perror when PEM_read_bio_X509_REQ fails, the error message must
14996 be obtained from the error queue.
15000 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15001 it in ERR_remove_state if appropriate, and change ERR_get_state
15002 accordingly to avoid race conditions (this is necessary because
15003 thread_hash is no longer constant once set).
15007 * Bugfix for linux-elf makefile.one.
15011 * RSA_get_default_method() will now cause a default
15012 RSA_METHOD to be chosen if one doesn't exist already.
15013 Previously this was only set during a call to RSA_new()
15014 or RSA_new_method(NULL) meaning it was possible for
15015 RSA_get_default_method() to return NULL.
15019 * Added native name translation to the existing DSO code
15020 that will convert (if the flag to do so is set) filenames
15021 that are sufficiently small and have no path information
15022 into a canonical native form. Eg. "blah" converted to
15023 "libblah.so" or "blah.dll" etc.
15027 * New function ERR_error_string_n(e, buf, len) which is like
15028 ERR_error_string(e, buf), but writes at most 'len' bytes
15029 including the 0 terminator. For ERR_error_string_n, 'buf'
15032 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15034 * CONF library reworked to become more general. A new CONF
15035 configuration file reader "class" is implemented as well as a
15036 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15037 old `CONF_*` functions are still there, but are reimplemented to
15038 work in terms of the new functions. Also, a set of functions
15039 to handle the internal storage of the configuration data is
15040 provided to make it easier to write new configuration file
15041 reader "classes" (I can definitely see something reading a
15042 configuration file in XML format, for example), called `_CONF_*`,
15043 or "the configuration storage API"...
15045 The new configuration file reading functions are:
15047 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15048 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15050 NCONF_default, NCONF_WIN32
15052 NCONF_dump_fp, NCONF_dump_bio
15054 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15055 NCONF_new creates a new CONF object. This works in the same way
15056 as other interfaces in OpenSSL, like the BIO interface.
15057 `NCONF_dump_*` dump the internal storage of the configuration file,
15058 which is useful for debugging. All other functions take the same
15059 arguments as the old `CONF_*` functions with the exception of the
15060 first that must be a `CONF *` instead of a `LHASH *`.
15062 To make it easier to use the new classes with the old `CONF_*` functions,
15063 the function CONF_set_default_method is provided.
15067 * Add '-tls1' option to 'openssl ciphers', which was already
15068 mentioned in the documentation but had not been implemented.
15069 (This option is not yet really useful because even the additional
15070 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15074 * Initial DSO code added into libcrypto for letting OpenSSL (and
15075 OpenSSL-based applications) load shared libraries and bind to
15076 them in a portable way.
15078 *Geoff Thorpe, with contributions from Richard Levitte*
15080 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15082 * Make sure _lrotl and _lrotr are only used with MSVC.
15084 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15085 (the default implementation of RAND_status).
15087 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15088 to '-clrext' (= clear extensions), as intended and documented.
15089 *Bodo Moeller; inconsistency pointed out by Michael Attili
15090 <attili@amaxo.com>*
15092 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15093 was larger than the MD block size.
15095 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15097 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15098 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15099 using the passed key: if the passed key was a private key the result
15100 of X509_print(), for example, would be to print out all the private key
15105 * des_quad_cksum() byte order bug fix.
15106 *Ulf Möller, using the problem description in krb4-0.9.7, where
15107 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15109 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15112 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15114 * For easily testing in shell scripts whether some command
15115 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15116 returns with exit code 0 iff no command of the given name is available.
15117 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15118 the output goes to stdout and nothing is printed to stderr.
15119 Additional arguments are always ignored.
15121 Since for each cipher there is a command of the same name,
15122 the 'no-cipher' compilation switches can be tested this way.
15124 ('openssl no-XXX' is not able to detect pseudo-commands such
15125 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15129 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15133 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15134 is set; it will be thrown away anyway because each handshake creates
15136 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15137 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15138 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15139 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15143 * New s_client option -ign_eof: EOF at stdin is ignored, and
15144 'Q' and 'R' lose their special meanings (quit/renegotiate).
15145 This is part of what -quiet does; unlike -quiet, -ign_eof
15146 does not suppress any output.
15150 * Add compatibility options to the purpose and trust code. The
15151 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15152 accepts a certificate or CA, this was the previous behaviour,
15153 with all the associated security issues.
15155 X509_TRUST_COMPAT is the old trust behaviour: only and
15156 automatically trust self signed roots in certificate store. A
15157 new trust setting X509_TRUST_DEFAULT is used to specify that
15158 a purpose has no associated trust setting and it should instead
15159 use the value in the default purpose.
15163 * Fix the PKCS#8 DSA private key code so it decodes keys again
15164 and fix a memory leak.
15168 * In util/mkerr.pl (which implements 'make errors'), preserve
15169 reason strings from the previous version of the .c file, as
15170 the default to have only downcase letters (and digits) in
15171 automatically generated reasons codes is not always appropriate.
15175 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15176 using strerror. Previously, ERR_reason_error_string() returned
15177 library names as reason strings for SYSerr; but SYSerr is a special
15178 case where small numbers are errno values, not library numbers.
15182 * Add '-dsaparam' option to 'openssl dhparam' application. This
15183 converts DSA parameters into DH parameters. (When creating parameters,
15184 DSA_generate_parameters is used.)
15188 * Include 'length' (recommended exponent length) in C code generated
15189 by 'openssl dhparam -C'.
15193 * The second argument to set_label in perlasm was already being used
15194 so couldn't be used as a "file scope" flag. Moved to third argument
15199 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15200 instead of RAND_bytes for encryption IVs and salts.
15204 * Include RAND_status() into RAND_METHOD instead of implementing
15205 it only for md_rand.c Otherwise replacing the PRNG by calling
15206 RAND_set_rand_method would be impossible.
15210 * Don't let DSA_generate_key() enter an infinite loop if the random
15211 number generation fails.
15215 * New 'rand' application for creating pseudo-random output.
15219 * Added configuration support for Linux/IA64
15221 *Rolf Haberrecker <rolf@suse.de>*
15223 * Assembler module support for Mingw32.
15227 * Shared library support for HPUX (in shlib/).
15229 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15231 * Shared library support for Solaris gcc.
15233 *Lutz Behnke <behnke@trustcenter.de>*
15235 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15237 * PKCS7_encrypt() was adding text MIME headers twice because they
15238 were added manually and by SMIME_crlf_copy().
15242 * In bntest.c don't call BN_rand with zero bits argument.
15244 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15246 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15247 case was implemented. This caused BN_div_recp() to fail occasionally.
15251 * Add an optional second argument to the set_label() in the perl
15252 assembly language builder. If this argument exists and is set
15253 to 1 it signals that the assembler should use a symbol whose
15254 scope is the entire file, not just the current function. This
15255 is needed with MASM which uses the format label:: for this scope.
15257 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15259 * Change the ASN1 types so they are typedefs by default. Before
15260 almost all types were #define'd to ASN1_STRING which was causing
15261 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15266 * Change names of new functions to the new get1/get0 naming
15267 convention: After 'get1', the caller owns a reference count
15268 and has to call `..._free`; 'get0' returns a pointer to some
15269 data structure without incrementing reference counters.
15270 (Some of the existing 'get' functions increment a reference
15271 counter, some don't.)
15272 Similarly, 'set1' and 'add1' functions increase reference
15273 counters or duplicate objects.
15277 * Allow for the possibility of temp RSA key generation failure:
15278 the code used to assume it always worked and crashed on failure.
15282 * Fix potential buffer overrun problem in BIO_printf().
15283 *Ulf Möller, using public domain code by Patrick Powell; problem
15284 pointed out by David Sacerdote <das33@cornell.edu>*
15286 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15287 RAND_egd() and RAND_status(). In the command line application,
15288 the EGD socket can be specified like a seed file using RANDFILE
15293 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15294 Some CAs (e.g. Verisign) distribute certificates in this form.
15298 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15299 list to exclude them. This means that no special compilation option
15300 is needed to use anonymous DH: it just needs to be included in the
15305 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15306 EVP_MD_type. The old functionality is available in a new macro called
15307 EVP_MD_md(). Change code that uses it and update docs.
15311 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15312 where the `void *` argument is replaced by a function pointer argument.
15313 Previously `void *` was abused to point to functions, which works on
15314 many platforms, but is not correct. As these functions are usually
15315 called by macros defined in OpenSSL header files, most source code
15316 should work without changes.
15320 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15321 sections with information on -D... compiler switches used for
15322 compiling the library so that applications can see them. To enable
15323 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15324 must be defined. E.g.,
15325 #define OPENSSL_ALGORITHM_DEFINES
15326 #include <openssl/opensslconf.h>
15327 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15329 *Richard Levitte, Ulf and Bodo Möller*
15331 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15336 * Change the 'other' type in certificate aux info to a STACK_OF
15337 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15338 the required ASN1 format: arbitrary types determined by an OID.
15342 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15343 argument to 'req'. This is not because the function is newer or
15344 better than others it just uses the work 'NEW' in the certificate
15345 request header lines. Some software needs this.
15349 * Reorganise password command line arguments: now passwords can be
15350 obtained from various sources. Delete the PEM_cb function and make
15351 it the default behaviour: i.e. if the callback is NULL and the
15352 usrdata argument is not NULL interpret it as a null terminated pass
15353 phrase. If usrdata and the callback are NULL then the pass phrase
15354 is prompted for as usual.
15358 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15359 the support is automatically enabled. The resulting binaries will
15360 autodetect the card and use it if present.
15362 *Ben Laurie and Compaq Inc.*
15364 * Work around for Netscape hang bug. This sends certificate request
15365 and server done in one record. Since this is perfectly legal in the
15366 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15367 the bugs/SSLv3 entry for more info.
15371 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15375 * Add -rand argument to smime and pkcs12 applications and read/write
15380 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15384 * Add command line password options to the remaining applications.
15388 * Bug fix for BN_div_recp() for numerators with an even number of
15393 * More tests in bntest.c, and changed test_bn output.
15397 * ./config recognizes MacOS X now.
15401 * Bug fix for BN_div() when the first words of num and divisor are
15402 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15406 * Add support for various broken PKCS#8 formats, and command line
15407 options to produce them.
15411 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15412 get temporary BIGNUMs from a BN_CTX.
15416 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15421 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15422 include a #define from the old name to the new. The original intent
15423 was that statically linked binaries could for example just call
15424 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15425 link with digests. This never worked because SSLeay_add_all_digests()
15426 and SSLeay_add_all_ciphers() were in the same source file so calling
15427 one would link with the other. They are now in separate source files.
15431 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15435 * Use a less unusual form of the Miller-Rabin primality test (it used
15436 a binary algorithm for exponentiation integrated into the Miller-Rabin
15437 loop, our standard modexp algorithms are faster).
15441 * Support for the EBCDIC character set completed.
15443 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15445 * Source code cleanups: use const where appropriate, eliminate casts,
15446 use `void *` instead of `char *` in lhash.
15450 * Bugfix: ssl3_send_server_key_exchange was not restartable
15451 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15452 this the server could overwrite ephemeral keys that the client
15457 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15458 using 50 iterations of the Rabin-Miller test.
15460 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15461 iterations of the Rabin-Miller test as required by the appendix
15462 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15463 As BN_is_prime_fasttest includes trial division, DSA parameter
15464 generation becomes much faster.
15466 This implies a change for the callback functions in DSA_is_prime
15467 and DSA_generate_parameters: The callback function is called once
15468 for each positive witness in the Rabin-Miller test, not just
15469 occasionally in the inner loop; and the parameters to the
15470 callback function now provide an iteration count for the outer
15471 loop rather than for the current invocation of the inner loop.
15472 DSA_generate_parameters additionally can call the callback
15473 function with an 'iteration count' of -1, meaning that a
15474 candidate has passed the trial division test (when q is generated
15475 from an application-provided seed, trial division is skipped).
15479 * New function BN_is_prime_fasttest that optionally does trial
15480 division before starting the Rabin-Miller test and has
15481 an additional BN_CTX * argument (whereas BN_is_prime always
15482 has to allocate at least one BN_CTX).
15483 'callback(1, -1, cb_arg)' is called when a number has passed the
15484 trial division stage.
15488 * Fix for bug in CRL encoding. The validity dates weren't being handled
15493 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15497 * New function BN_pseudo_rand().
15501 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15502 bignum version of BN_from_montgomery() with the working code from
15503 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15508 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15509 made it impossible to use the same SSL_SESSION data structure in
15510 SSL2 clients in multiple threads.
15514 * The return value of RAND_load_file() no longer counts bytes obtained
15515 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15516 to seed the PRNG (previously an explicit byte count was required).
15518 *Ulf Möller, Bodo Möller*
15520 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15521 used `char *` instead of `void *` and had casts all over the place.
15525 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15529 * Retain source code compatibility for BN_prime_checks macro:
15530 BN_is_prime(..., BN_prime_checks, ...) now uses
15531 BN_prime_checks_for_size to determine the appropriate number of
15532 Rabin-Miller iterations.
15536 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15537 DH_CHECK_P_NOT_SAFE_PRIME.
15538 (Check if this is true? OpenPGP calls them "strong".)
15542 * Merge the functionality of "dh" and "gendh" programs into a new program
15543 "dhparam". The old programs are retained for now but will handle DH keys
15544 (instead of parameters) in future.
15548 * Make the ciphers, s_server and s_client programs check the return values
15549 when a new cipher list is set.
15553 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15554 ciphers. Before when the 56bit ciphers were enabled the sorting was
15557 The syntax for the cipher sorting has been extended to support sorting by
15558 cipher-strength (using the strength_bits hard coded in the tables).
15559 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15561 Fix a bug in the cipher-command parser: when supplying a cipher command
15562 string with an "undefined" symbol (neither command nor alphanumeric
15563 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15564 an error is flagged.
15566 Due to the strength-sorting extension, the code of the
15567 ssl_create_cipher_list() function was completely rearranged. I hope that
15568 the readability was also increased :-)
15570 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15572 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15573 for the first serial number and places 2 in the serial number file. This
15574 avoids problems when the root CA is created with serial number zero and
15575 the first user certificate has the same issuer name and serial number
15580 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15581 the new code. Add documentation for this stuff.
15585 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15586 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15587 structures and behave in an analogous way to the X509v3 functions:
15588 they shouldn't be called directly but wrapper functions should be used
15591 So we also now have some wrapper functions that call the X509at functions
15592 when passed certificate requests. (TO DO: similar things can be done with
15593 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15594 things. Some of these need some d2i or i2d and print functionality
15595 because they handle more complex structures.)
15599 * Add missing #ifndefs that caused missing symbols when building libssl
15600 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15601 NO_RSA in `ssl/s2*.c`.
15603 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15605 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15606 has a return value which indicates the quality of the random data
15607 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15608 error queue. New function RAND_pseudo_bytes() generates output that is
15609 guaranteed to be unique but not unpredictable. RAND_add is like
15610 RAND_seed, but takes an extra argument for an entropy estimate
15611 (RAND_seed always assumes full entropy).
15615 * Do more iterations of Rabin-Miller probable prime test (specifically,
15616 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15617 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15618 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15619 false-positive rate of at most 2^-80 for random input.
15623 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15627 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15628 in the 0.9.5 release), this returns the chain
15629 from an X509_CTX structure with a dup of the stack and all
15630 the X509 reference counts upped: so the stack will exist
15631 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15634 Also make SSL_SESSION_print() print out the verify return
15639 * Add manpage for the pkcs12 command. Also change the default
15640 behaviour so MAC iteration counts are used unless the new
15641 -nomaciter option is used. This improves file security and
15642 only older versions of MSIE (4.0 for example) need it.
15646 * Honor the no-xxx Configure options when creating .DEF files.
15650 * Add PKCS#10 attributes to field table: challengePassword,
15651 unstructuredName and unstructuredAddress. These are taken from
15652 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15653 international characters are used.
15655 More changes to X509_ATTRIBUTE code: allow the setting of types
15656 based on strings. Remove the 'loc' parameter when adding
15657 attributes because these will be a SET OF encoding which is sorted
15662 * Initial changes to the 'req' utility to allow request generation
15663 automation. This will allow an application to just generate a template
15664 file containing all the field values and have req construct the
15667 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15668 used all over the place including certificate requests and PKCS#7
15669 structures. They are currently handled manually where necessary with
15670 some primitive wrappers for PKCS#7. The new functions behave in a
15671 manner analogous to the X509 extension functions: they allow
15672 attributes to be looked up by NID and added.
15674 Later something similar to the X509V3 code would be desirable to
15675 automatically handle the encoding, decoding and printing of the
15676 more complex types. The string types like challengePassword can
15677 be handled by the string table functions.
15679 Also modified the multi byte string table handling. Now there is
15680 a 'global mask' which masks out certain types. The table itself
15681 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15682 is useful when for example there is only one permissible type
15683 (as in countryName) and using the mask might result in no valid
15688 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15689 SSL_get_peer_finished to allow applications to obtain the latest
15690 Finished messages sent to the peer or expected from the peer,
15691 respectively. (SSL_get_peer_finished is usually the Finished message
15692 actually received from the peer, otherwise the protocol will be aborted.)
15694 As the Finished message are message digests of the complete handshake
15695 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15696 be used for external authentication procedures when the authentication
15697 provided by SSL/TLS is not desired or is not enough.
15701 * Enhanced support for Alpha Linux is added. Now ./config checks if
15702 the host supports BWX extension and if Compaq C is present on the
15703 $PATH. Just exploiting of the BWX extension results in 20-30%
15704 performance kick for some algorithms, e.g. DES and RC4 to mention
15705 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15710 * Add support for MS "fast SGC". This is arguably a violation of the
15711 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15712 weak crypto and after checking the certificate is SGC a second one
15713 with strong crypto. MS SGC stops the first handshake after receiving
15714 the server certificate message and sends a second client hello. Since
15715 a server will typically do all the time consuming operations before
15716 expecting any further messages from the client (server key exchange
15717 is the most expensive) there is little difference between the two.
15719 To get OpenSSL to support MS SGC we have to permit a second client
15720 hello message after we have sent server done. In addition we have to
15721 reset the MAC if we do get this second client hello.
15725 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15726 if a DER encoded private key is RSA or DSA traditional format. Changed
15727 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15728 format DER encoded private key. Newer code should use PKCS#8 format which
15729 has the key type encoded in the ASN1 structure. Added DER private key
15730 support to pkcs8 application.
15734 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15735 ciphersuites has been selected (as required by the SSL 3/TLS 1
15736 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15737 is set, we interpret this as a request to violate the specification
15738 (the worst that can happen is a handshake failure, and 'correct'
15739 behaviour would result in a handshake failure anyway).
15743 * In SSL_CTX_add_session, take into account that there might be multiple
15744 SSL_SESSION structures with the same session ID (e.g. when two threads
15745 concurrently obtain them from an external cache).
15746 The internal cache can handle only one SSL_SESSION with a given ID,
15747 so if there's a conflict, we now throw out the old one to achieve
15752 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15753 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15754 some routines that use cipher OIDs: some ciphers do not have OIDs
15755 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15760 * Simplify the trust setting structure and code. Now we just have
15761 two sequences of OIDs for trusted and rejected settings. These will
15762 typically have values the same as the extended key usage extension
15763 and any application specific purposes.
15765 The trust checking code now has a default behaviour: it will just
15766 check for an object with the same NID as the passed id. Functions can
15767 be provided to override either the default behaviour or the behaviour
15768 for a given id. SSL client, server and email already have functions
15769 in place for compatibility: they check the NID and also return "trusted"
15770 if the certificate is self signed.
15774 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15775 traditional format into an EVP_PKEY structure.
15779 * Add a password callback function PEM_cb() which either prompts for
15780 a password if usr_data is NULL or otherwise assumes it is a null
15781 terminated password. Allow passwords to be passed on command line
15782 environment or config files in a few more utilities.
15786 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15787 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15788 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15789 Update documentation.
15793 * Support for ASN1 "NULL" type. This could be handled before by using
15794 ASN1_TYPE but there wasn't any function that would try to read a NULL
15795 and produce an error if it couldn't. For compatibility we also have
15796 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15797 don't allocate anything because they don't need to.
15801 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15804 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15806 * Rebuild of the memory allocation routines used by OpenSSL code and
15807 possibly others as well. The purpose is to make an interface that
15808 provide hooks so anyone can build a separate set of allocation and
15809 deallocation routines to be used by OpenSSL, for example memory
15810 pool implementations, or something else, which was previously hard
15811 since Malloc(), Realloc() and Free() were defined as macros having
15812 the values malloc, realloc and free, respectively (except for Win32
15813 compilations). The same is provided for memory debugging code.
15814 OpenSSL already comes with functionality to find memory leaks, but
15815 this gives people a chance to debug other memory problems.
15817 With these changes, a new set of functions and macros have appeared:
15819 CRYPTO_set_mem_debug_functions() [F]
15820 CRYPTO_get_mem_debug_functions() [F]
15821 CRYPTO_dbg_set_options() [F]
15822 CRYPTO_dbg_get_options() [F]
15823 CRYPTO_malloc_debug_init() [M]
15825 The memory debug functions are NULL by default, unless the library
15826 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15827 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15828 gives the standard debugging functions that come with OpenSSL) or
15829 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15830 provided by the library user) must be used. When the standard
15831 debugging functions are used, CRYPTO_dbg_set_options can be used to
15832 request additional information:
15833 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15834 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15836 Also, things like CRYPTO_set_mem_functions will always give the
15837 expected result (the new set of functions is used for allocation
15838 and deallocation) at all times, regardless of platform and compiler
15841 To finish it up, some functions that were never use in any other
15842 way than through macros have a new API and new semantic:
15844 CRYPTO_dbg_malloc()
15845 CRYPTO_dbg_realloc()
15848 All macros of value have retained their old syntax.
15850 *Richard Levitte and Bodo Moeller*
15852 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15853 ordering of SMIMECapabilities wasn't in "strength order" and there
15854 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15859 * Some ASN1 types with illegal zero length encoding (INTEGER,
15860 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15862 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15864 * Merge in my S/MIME library for OpenSSL. This provides a simple
15865 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15866 functionality to handle multipart/signed properly) and a utility
15867 called 'smime' to call all this stuff. This is based on code I
15868 originally wrote for Celo who have kindly allowed it to be
15869 included in OpenSSL.
15873 * Add variants des_set_key_checked and des_set_key_unchecked of
15874 des_set_key (aka des_key_sched). Global variable des_check_key
15875 decides which of these is called by des_set_key; this way
15876 des_check_key behaves as it always did, but applications and
15877 the library itself, which was buggy for des_check_key == 1,
15878 have a cleaner way to pick the version they need.
15882 * New function PKCS12_newpass() which changes the password of a
15887 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15888 dynamic mix. In both cases the ids can be used as an index into the
15889 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15890 functions so they accept a list of the field values and the
15891 application doesn't need to directly manipulate the X509_TRUST
15896 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15901 * Modify the way the V3 extension code looks up extensions. This now
15902 works in a similar way to the object code: we have some "standard"
15903 extensions in a static table which is searched with OBJ_bsearch()
15904 and the application can add dynamic ones if needed. The file
15905 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15906 updated whenever a new extension is added to the core code and kept
15907 in ext_nid order. There is a simple program 'tabtest.c' which checks
15908 this. New extensions are not added too often so this file can readily
15909 be maintained manually.
15911 There are two big advantages in doing things this way. The extensions
15912 can be looked up immediately and no longer need to be "added" using
15913 X509V3_add_standard_extensions(): this function now does nothing.
15914 Side note: I get *lots* of email saying the extension code doesn't
15915 work because people forget to call this function.
15916 Also no dynamic allocation is done unless new extensions are added:
15917 so if we don't add custom extensions there is no need to call
15918 X509V3_EXT_cleanup().
15922 * Modify enc utility's salting as follows: make salting the default. Add a
15923 magic header, so unsalted files fail gracefully instead of just decrypting
15924 to garbage. This is because not salting is a big security hole, so people
15925 should be discouraged from doing it.
15929 * Fixes and enhancements to the 'x509' utility. It allowed a message
15930 digest to be passed on the command line but it only used this
15931 parameter when signing a certificate. Modified so all relevant
15932 operations are affected by the digest parameter including the
15933 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15934 DSA key was used because it didn't fix the digest.
15938 * Initial certificate chain verify code. Currently tests the untrusted
15939 certificates for consistency with the verify purpose (which is set
15940 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15942 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15943 this is because it will reject chains with invalid extensions whereas
15944 every previous version of OpenSSL and SSLeay made no checks at all.
15946 Trust code: checks the root CA for the relevant trust settings. Trust
15947 settings have an initial value consistent with the verify purpose: e.g.
15948 if the verify purpose is for SSL client use it expects the CA to be
15949 trusted for SSL client use. However the default value can be changed to
15950 permit custom trust settings: one example of this would be to only trust
15951 certificates from a specific "secure" set of CAs.
15953 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15954 which should be used for version portability: especially since the
15955 verify structure is likely to change more often now.
15957 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15958 to set them. If not set then assume SSL clients will verify SSL servers
15961 Two new options to the verify program: -untrusted allows a set of
15962 untrusted certificates to be passed in and -purpose which sets the
15963 intended purpose of the certificate. If a purpose is set then the
15964 new chain verify code is used to check extension consistency.
15968 * Support for the authority information access extension.
15972 * Modify RSA and DSA PEM read routines to transparently handle
15973 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15974 public keys in a format compatible with certificate
15975 SubjectPublicKeyInfo structures. Unfortunately there were already
15976 functions called *_PublicKey_* which used various odd formats so
15977 these are retained for compatibility: however the DSA variants were
15978 never in a public release so they have been deleted. Changed dsa/rsa
15979 utilities to handle the new format: note no releases ever handled public
15980 keys so we should be OK.
15982 The primary motivation for this change is to avoid the same fiasco
15983 that dogs private keys: there are several incompatible private key
15984 formats some of which are standard and some OpenSSL specific and
15985 require various evil hacks to allow partial transparent handling and
15986 even then it doesn't work with DER formats. Given the option anything
15987 other than PKCS#8 should be dumped: but the other formats have to
15988 stay in the name of compatibility.
15990 With public keys and the benefit of hindsight one standard format
15991 is used which works with EVP_PKEY, RSA or DSA structures: though
15992 it clearly returns an error if you try to read the wrong kind of key.
15994 Added a -pubkey option to the 'x509' utility to output the public key.
15995 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15996 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15997 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15998 that do the same as the `EVP_PKEY_assign_*()` except they up the
15999 reference count of the added key (they don't "swallow" the
16004 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16005 CRLs would fail if the file contained no certificates or no CRLs:
16006 added a new function to read in both types and return the number
16007 read: this means that if none are read it will be an error. The
16008 DER versions of the certificate and CRL reader would always fail
16009 because it isn't possible to mix certificates and CRLs in DER format
16010 without choking one or the other routine. Changed this to just read
16011 a certificate: this is the best we can do. Also modified the code
16012 in apps/verify.c to take notice of return codes: it was previously
16013 attempting to read in certificates from NULL pointers and ignoring
16014 any errors: this is one reason why the cert and CRL reader seemed
16015 to work. It doesn't check return codes from the default certificate
16016 routines: these may well fail if the certificates aren't installed.
16020 * Code to support otherName option in GeneralName.
16024 * First update to verify code. Change the verify utility
16025 so it warns if it is passed a self signed certificate:
16026 for consistency with the normal behaviour. X509_verify
16027 has been modified to it will now verify a self signed
16028 certificate if *exactly* the same certificate appears
16029 in the store: it was previously impossible to trust a
16030 single self signed certificate. This means that:
16031 openssl verify ss.pem
16032 now gives a warning about a self signed certificate but
16033 openssl verify -CAfile ss.pem ss.pem
16038 * For servers, store verify_result in SSL_SESSION data structure
16039 (and add it to external session representation).
16040 This is needed when client certificate verifications fails,
16041 but an application-provided verification callback (set by
16042 SSL_CTX_set_cert_verify_callback) allows accepting the session
16043 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16044 but returns 1): When the session is reused, we have to set
16045 ssl->verify_result to the appropriate error code to avoid
16048 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16050 * Fix a bug in the new PKCS#7 code: it didn't consider the
16051 case in PKCS7_dataInit() where the signed PKCS7 structure
16052 didn't contain any existing data because it was being created.
16054 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16056 * Add a salt to the key derivation routines in enc.c. This
16057 forms the first 8 bytes of the encrypted file. Also add a
16058 -S option to allow a salt to be input on the command line.
16062 * New function X509_cmp(). Oddly enough there wasn't a function
16063 to compare two certificates. We do this by working out the SHA1
16064 hash and comparing that. X509_cmp() will be needed by the trust
16069 * SSL_get1_session() is like SSL_get_session(), but increments
16070 the reference count in the SSL_SESSION returned.
16072 *Geoff Thorpe <geoff@eu.c2.net>*
16074 * Fix for 'req': it was adding a null to request attributes.
16075 Also change the X509_LOOKUP and X509_INFO code to handle
16076 certificate auxiliary information.
16080 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16085 * Add the possibility to add extra information to the memory leak
16086 detecting output, to form tracebacks, showing from where each
16087 allocation was originated: CRYPTO_push_info("constant string") adds
16088 the string plus current file name and line number to a per-thread
16089 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16090 is like calling CYRPTO_pop_info() until the stack is empty.
16091 Also updated memory leak detection code to be multi-thread-safe.
16095 * Add options -text and -noout to pkcs7 utility and delete the
16096 encryption options which never did anything. Update docs.
16100 * Add options to some of the utilities to allow the pass phrase
16101 to be included on either the command line (not recommended on
16102 OSes like Unix) or read from the environment. Update the
16103 manpages and fix a few bugs.
16107 * Add a few manpages for some of the openssl commands.
16111 * Fix the -revoke option in ca. It was freeing up memory twice,
16112 leaking and not finding already revoked certificates.
16116 * Extensive changes to support certificate auxiliary information.
16117 This involves the use of X509_CERT_AUX structure and X509_AUX
16118 functions. An X509_AUX function such as PEM_read_X509_AUX()
16119 can still read in a certificate file in the usual way but it
16120 will also read in any additional "auxiliary information". By
16121 doing things this way a fair degree of compatibility can be
16122 retained: existing certificates can have this information added
16123 using the new 'x509' options.
16125 Current auxiliary information includes an "alias" and some trust
16126 settings. The trust settings will ultimately be used in enhanced
16127 certificate chain verification routines: currently a certificate
16128 can only be trusted if it is self signed and then it is trusted
16133 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16134 The problem was that one of the replacement routines had not been working
16135 since SSLeay releases. For now the offending routine has been replaced
16136 with non-optimised assembler. Even so, this now gives around 95%
16137 performance improvement for 1024 bit RSA signs.
16141 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16142 handling. Most clients have the effective key size in bits equal to
16143 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16144 A few however don't do this and instead use the size of the decrypted key
16145 to determine the RC2 key length and the AlgorithmIdentifier to determine
16146 the effective key length. In this case the effective key length can still
16147 be 40 bits but the key length can be 168 bits for example. This is fixed
16148 by manually forcing an RC2 key into the EVP_PKEY structure because the
16149 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16150 the key length and effective key length are equal.
16154 * Add a bunch of functions that should simplify the creation of
16155 X509_NAME structures. Now you should be able to do:
16156 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16157 and have it automatically work out the correct field type and fill in
16158 the structures. The more adventurous can try:
16159 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16160 and it will (hopefully) work out the correct multibyte encoding.
16164 * Change the 'req' utility to use the new field handling and multibyte
16165 copy routines. Before the DN field creation was handled in an ad hoc
16166 way in req, ca, and x509 which was rather broken and didn't support
16167 BMPStrings or UTF8Strings. Since some software doesn't implement
16168 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16169 using the dirstring_type option. See the new comment in the default
16170 openssl.cnf for more info.
16174 * Make crypto/rand/md_rand.c more robust:
16175 - Assure unique random numbers after fork().
16176 - Make sure that concurrent threads access the global counter and
16177 md serializably so that we never lose entropy in them
16178 or use exactly the same state in multiple threads.
16179 Access to the large state is not always serializable because
16180 the additional locking could be a performance killer, and
16181 md should be large enough anyway.
16185 * New file apps/app_rand.c with commonly needed functionality
16186 for handling the random seed file.
16188 Use the random seed file in some applications that previously did not:
16190 dsaparam -genkey (which also ignored its '-rand' option),
16193 x509 (when signing).
16194 Except on systems with /dev/urandom, it is crucial to have a random
16195 seed file at least for key creation, DSA signing, and for DH exchanges;
16196 for RSA signatures we could do without one.
16198 gendh and gendsa (unlike genrsa) used to read only the first byte
16199 of each file listed in the '-rand' option. The function as previously
16200 found in genrsa is now in app_rand.c and is used by all programs
16201 that support '-rand'.
16205 * In RAND_write_file, use mode 0600 for creating files;
16206 don't just chmod when it may be too late.
16210 * Report an error from X509_STORE_load_locations
16211 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16215 * New function ASN1_mbstring_copy() this copies a string in either
16216 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16217 into an ASN1_STRING type. A mask of permissible types is passed
16218 and it chooses the "minimal" type to use or an error if not type
16223 * Add function equivalents to the various macros in asn1.h. The old
16224 macros are retained with an `M_` prefix. Code inside the library can
16225 use the `M_` macros. External code (including the openssl utility)
16226 should *NOT* in order to be "shared library friendly".
16230 * Add various functions that can check a certificate's extensions
16231 to see if it usable for various purposes such as SSL client,
16232 server or S/MIME and CAs of these types. This is currently
16233 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16234 verification. Also added a -purpose flag to x509 utility to
16235 print out all the purposes.
16239 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16244 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16245 for, obtain and decode and extension and obtain its critical flag.
16246 This allows all the necessary extension code to be handled in a
16247 single function call.
16251 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16252 platforms. See crypto/rc4/rc4_enc.c for further details.
16256 * New -noout option to asn1parse. This causes no output to be produced
16257 its main use is when combined with -strparse and -out to extract data
16258 from a file (which may not be in ASN.1 format).
16262 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16263 when producing the local key id.
16265 *Richard Levitte <levitte@stacken.kth.se>*
16267 * New option -dhparam in s_server. This allows a DH parameter file to be
16268 stated explicitly. If it is not stated then it tries the first server
16269 certificate file. The previous behaviour hard coded the filename
16274 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16275 a public key to be input or output. For example:
16276 openssl rsa -in key.pem -pubout -out pubkey.pem
16277 Also added necessary DSA public key functions to handle this.
16281 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16282 in the message. This was handled by allowing
16283 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16285 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16287 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16288 to the end of the strings whereas this didn't. This would cause problems
16289 if strings read with d2i_ASN1_bytes() were later modified.
16291 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16293 * Fix for base64 decode bug. When a base64 bio reads only one line of
16294 data and it contains EOF it will end up returning an error. This is
16295 caused by input 46 bytes long. The cause is due to the way base64
16296 BIOs find the start of base64 encoded data. They do this by trying a
16297 trial decode on each line until they find one that works. When they
16298 do a flag is set and it starts again knowing it can pass all the
16299 data directly through the decoder. Unfortunately it doesn't reset
16300 the context it uses. This means that if EOF is reached an attempt
16301 is made to pass two EOFs through the context and this causes the
16302 resulting error. This can also cause other problems as well. As is
16303 usual with these problems it takes *ages* to find and the fix is
16304 trivial: move one line.
16306 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16308 * Ugly workaround to get s_client and s_server working under Windows. The
16309 old code wouldn't work because it needed to select() on sockets and the
16310 tty (for keypresses and to see if data could be written). Win32 only
16311 supports select() on sockets so we select() with a 1s timeout on the
16312 sockets and then see if any characters are waiting to be read, if none
16313 are present then we retry, we also assume we can always write data to
16314 the tty. This isn't nice because the code then blocks until we've
16315 received a complete line of data and it is effectively polling the
16316 keyboard at 1s intervals: however it's quite a bit better than not
16317 working at all :-) A dedicated Windows application might handle this
16318 with an event loop for example.
16322 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16323 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16324 will be called when RSA_sign() and RSA_verify() are used. This is useful
16325 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16326 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16327 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16328 This necessitated the support of an extra signature type NID_md5_sha1
16329 for SSL signatures and modifications to the SSL library to use it instead
16330 of calling RSA_public_decrypt() and RSA_private_encrypt().
16334 * Add new -verify -CAfile and -CApath options to the crl program, these
16335 will lookup a CRL issuers certificate and verify the signature in a
16336 similar way to the verify program. Tidy up the crl program so it
16337 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16338 less strict. It will now permit CRL extensions even if it is not
16339 a V2 CRL: this will allow it to tolerate some broken CRLs.
16343 * Initialize all non-automatic variables each time one of the openssl
16344 sub-programs is started (this is necessary as they may be started
16345 multiple times from the "OpenSSL>" prompt).
16347 *Lennart Bang, Bodo Moeller*
16349 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16350 removing all other RSA functionality (this is what NO_RSA does). This
16351 is so (for example) those in the US can disable those operations covered
16352 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16357 * Non-copying interface to BIO pairs.
16358 (still largely untested)
16362 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16363 ASCII string. This was handled independently in various places before.
16367 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16368 UTF8 strings a character at a time.
16372 * Use client_version from client hello to select the protocol
16373 (s23_srvr.c) and for RSA client key exchange verification
16374 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16378 * Add various utility functions to handle SPKACs, these were previously
16379 handled by poking round in the structure internals. Added new function
16380 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16381 print, verify and generate SPKACs. Based on an original idea from
16382 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16386 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16390 * Allow the config file extension section to be overwritten on the
16391 command line. Based on an original idea from Massimiliano Pala
16392 <madwolf@comune.modena.it>. The new option is called -extensions
16393 and can be applied to ca, req and x509. Also -reqexts to override
16394 the request extensions in req and -crlexts to override the crl extensions
16399 * Add new feature to the SPKAC handling in ca. Now you can include
16400 the same field multiple times by preceding it by "XXXX." for example:
16403 this is the same syntax as used in the req config file.
16407 * Allow certificate extensions to be added to certificate requests. These
16408 are specified in a 'req_extensions' option of the req section of the
16409 config file. They can be printed out with the -text option to req but
16410 are otherwise ignored at present.
16414 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16415 data read consists of only the final block it would not decrypted because
16416 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16417 A misplaced 'break' also meant the decrypted final block might not be
16418 copied until the next read.
16422 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16423 a few extra parameters to the DH structure: these will be useful if
16424 for example we want the value of 'q' or implement X9.42 DH.
16428 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16429 provides hooks that allow the default DSA functions or functions on a
16430 "per key" basis to be replaced. This allows hardware acceleration and
16431 hardware key storage to be handled without major modification to the
16432 library. Also added low level modexp hooks and CRYPTO_EX structure and
16433 associated functions.
16437 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16438 as "read only": it can't be written to and the buffer it points to will
16439 not be freed. Reading from a read only BIO is much more efficient than
16440 a normal memory BIO. This was added because there are several times when
16441 an area of memory needs to be read from a BIO. The previous method was
16442 to create a memory BIO and write the data to it, this results in two
16443 copies of the data and an O(n^2) reading algorithm. There is a new
16444 function BIO_new_mem_buf() which creates a read only memory BIO from
16445 an area of memory. Also modified the PKCS#7 routines to use read only
16450 * Bugfix: ssl23_get_client_hello did not work properly when called in
16451 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16452 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16453 but a retry condition occurred while trying to read the rest.
16457 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16458 NID_pkcs7_encrypted by default: this was wrong since this should almost
16459 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16460 the encrypted data type: this is a more sensible place to put it and it
16461 allows the PKCS#12 code to be tidied up that duplicated this
16466 * Changed obj_dat.pl script so it takes its input and output files on
16467 the command line. This should avoid shell escape redirection problems
16472 * Initial support for certificate extension requests, these are included
16473 in things like Xenroll certificate requests. Included functions to allow
16474 extensions to be obtained and added.
16478 * -crlf option to s_client and s_server for sending newlines as
16479 CRLF (as required by many protocols).
16483 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16485 * Install libRSAglue.a when OpenSSL is built with RSAref.
16487 *Ralf S. Engelschall*
16489 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16491 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16493 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16498 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16499 DH parameters/keys (q is lost during that conversion, but the resulting
16500 DH parameters contain its length).
16502 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16503 much faster than DH_generate_parameters (which creates parameters
16504 where `p = 2*q + 1`), and also the smaller q makes DH computations
16505 much more efficient (160-bit exponentiation instead of 1024-bit
16506 exponentiation); so this provides a convenient way to support DHE
16507 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16508 utter importance to use
16509 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16511 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16512 when such DH parameters are used, because otherwise small subgroup
16513 attacks may become possible!
16517 * Avoid memory leak in i2d_DHparams.
16521 * Allow the -k option to be used more than once in the enc program:
16522 this allows the same encrypted message to be read by multiple recipients.
16526 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16527 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16528 it will always use the numerical form of the OID, even if it has a short
16533 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16534 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16535 otherwise bn_mod_exp was called. In the case of hardware keys for example
16536 no private key components need be present and it might store extra data
16537 in the RSA structure, which cannot be accessed from bn_mod_exp.
16538 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16539 private key operations.
16543 * Added support for SPARC Linux.
16547 * pem_password_cb function type incompatibly changed from
16548 typedef int pem_password_cb(char *buf, int size, int rwflag);
16550 ....(char *buf, int size, int rwflag, void *userdata);
16551 so that applications can pass data to their callbacks:
16552 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16553 additional void * argument, which is just handed through whenever
16554 the password callback is called.
16556 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16558 New function SSL_CTX_set_default_passwd_cb_userdata.
16560 Compatibility note: As many C implementations push function arguments
16561 onto the stack in reverse order, the new library version is likely to
16562 interoperate with programs that have been compiled with the old
16563 pem_password_cb definition (PEM_whatever takes some data that
16564 happens to be on the stack as its last argument, and the callback
16565 just ignores this garbage); but there is no guarantee whatsoever that
16568 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16569 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16570 problems not only on Windows, but also on some Unix platforms.
16571 To avoid problematic command lines, these definitions are now in an
16572 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16573 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16577 * MIPS III/IV assembler module is reimplemented.
16581 * More DES library cleanups: remove references to srand/rand and
16582 delete an unused file.
16586 * Add support for the free Netwide assembler (NASM) under Win32,
16587 since not many people have MASM (ml) and it can be hard to obtain.
16588 This is currently experimental but it seems to work OK and pass all
16589 the tests. Check out INSTALL.W32 for info.
16593 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16594 without temporary keys kept an extra copy of the server key,
16595 and connections with temporary keys did not free everything in case
16600 * New function RSA_check_key and new openssl rsa option -check
16601 for verifying the consistency of RSA keys.
16603 *Ulf Moeller, Bodo Moeller*
16605 * Various changes to make Win32 compile work:
16606 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16607 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16608 comparison" warnings.
16609 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16613 * Add a debugging option to PKCS#5 v2 key generation function: when
16614 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16615 derived keys are printed to stderr.
16619 * Copy the flags in ASN1_STRING_dup().
16621 *Roman E. Pavlov <pre@mo.msk.ru>*
16623 * The x509 application mishandled signing requests containing DSA
16624 keys when the signing key was also DSA and the parameters didn't match.
16626 It was supposed to omit the parameters when they matched the signing key:
16627 the verifying software was then supposed to automatically use the CA's
16628 parameters if they were absent from the end user certificate.
16630 Omitting parameters is no longer recommended. The test was also
16631 the wrong way round! This was probably due to unusual behaviour in
16632 EVP_cmp_parameters() which returns 1 if the parameters match.
16633 This meant that parameters were omitted when they *didn't* match and
16634 the certificate was useless. Certificates signed with 'ca' didn't have
16637 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16639 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16640 The interface is as follows:
16641 Applications can use
16642 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16643 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16644 "off" is now the default.
16645 The library internally uses
16646 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16647 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16648 to disable memory-checking temporarily.
16650 Some inconsistent states that previously were possible (and were
16651 even the default) are now avoided.
16653 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16654 with each memory chunk allocated; this is occasionally more helpful
16655 than just having a counter.
16657 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16659 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16664 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16665 which largely parallels "options", but is for changing API behaviour,
16666 whereas "options" are about protocol behaviour.
16667 Initial "mode" flags are:
16669 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16670 a single record has been written.
16671 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16672 retries use the same buffer location.
16673 (But all of the contents must be
16678 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16681 * Fix problems with no-hmac etc.
16683 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16685 * New functions RSA_get_default_method(), RSA_set_method() and
16686 RSA_get_method(). These allows replacement of RSA_METHODs without having
16687 to mess around with the internals of an RSA structure.
16691 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16692 Also really enable memory leak checks in openssl.c and in some
16695 *Chad C. Mulligan, Bodo Moeller*
16697 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16698 up the length of negative integers. This has now been simplified to just
16699 store the length when it is first determined and use it later, rather
16700 than trying to keep track of where data is copied and updating it to
16702 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16704 * Add a new function PKCS7_signatureVerify. This allows the verification
16705 of a PKCS#7 signature but with the signing certificate passed to the
16706 function itself. This contrasts with PKCS7_dataVerify which assumes the
16707 certificate is present in the PKCS#7 structure. This isn't always the
16708 case: certificates can be omitted from a PKCS#7 structure and be
16709 distributed by "out of band" means (such as a certificate database).
16713 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16714 function prototypes in pem.h, also change util/mkdef.pl to add the
16715 necessary function names.
16719 * mk1mf.pl (used by Windows builds) did not properly read the
16720 options set by Configure in the top level Makefile, and Configure
16721 was not even able to write more than one option correctly.
16722 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16726 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16727 file to be loaded from a BIO or FILE pointer. The BIO version will
16728 for example allow memory BIOs to contain config info.
16732 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16733 Whoever hopes to achieve shared-library compatibility across versions
16734 must use this, not the compile-time macro.
16735 (Exercise 0.9.4: Which is the minimum library version required by
16737 Note: All this applies only to multi-threaded programs, others don't
16742 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16743 through a BIO pair triggered the default case, i.e.
16744 SSLerr(...,SSL_R_UNKNOWN_STATE).
16748 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16749 can use the SSL library even if none of the specific BIOs is
16754 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16755 for the encoded length.
16757 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16759 * Add initial documentation of the X509V3 functions.
16763 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16764 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16765 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16766 secure PKCS#8 private key format with a high iteration count.
16770 * Fix determination of Perl interpreter: A perl or perl5
16771 *directory* in $PATH was also accepted as the interpreter.
16773 *Ralf S. Engelschall*
16775 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16776 wrong with it but it was very old and did things like calling
16777 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16778 unusual formatting.
16782 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16783 to use the new extension code.
16787 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16788 with macros. This should make it easier to change their form, add extra
16789 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16794 * Add to configuration table a new entry that can specify an alternative
16795 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16796 according to Mark Crispin <MRC@Panda.COM>.
16800 * DES CBC did not update the IV. Weird.
16804 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16805 Changing the behaviour of the former might break existing programs --
16806 where IV updating is needed, des_ncbc_encrypt can be used.
16809 * When bntest is run from "make test" it drives bc to check its
16810 calculations, as well as internally checking them. If an internal check
16811 fails, it needs to cause bc to give a non-zero result or make test carries
16812 on without noticing the failure. Fixed.
16816 * DES library cleanups.
16820 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16821 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16822 ciphers. NOTE: although the key derivation function has been verified
16823 against some published test vectors it has not been extensively tested
16824 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16829 * Instead of "mkdir -p", which is not fully portable, use new
16830 Perl script "util/mkdir-p.pl".
16834 * Rewrite the way password based encryption (PBE) is handled. It used to
16835 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16836 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16837 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16838 the 'parameter' field of the AlgorithmIdentifier is passed to the
16839 underlying key generation function so it must do its own ASN1 parsing.
16840 This has also changed the EVP_PBE_CipherInit() function which now has a
16841 'parameter' argument instead of literal salt and iteration count values
16842 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16846 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16847 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16848 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16849 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16850 value was just used as a "magic string" and not used directly its
16851 value doesn't matter.
16855 * Introduce some semblance of const correctness to BN. Shame C doesn't
16860 * "linux-sparc64" configuration (ultrapenguin).
16862 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16863 "linux-sparc" configuration.
16865 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16867 * config now generates no-xxx options for missing ciphers.
16871 * Support the EBCDIC character set (work in progress).
16872 File ebcdic.c not yet included because it has a different license.
16874 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16876 * Support BS2000/OSD-POSIX.
16878 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16880 * Make callbacks for key generation use `void *` instead of `char *`.
16884 * Make S/MIME samples compile (not yet tested).
16888 * Additional typesafe stacks.
16892 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16896 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16898 * New configuration variant "sco5-gcc".
16900 * Updated some demos.
16902 *Sean O Riordain, Wade Scholine*
16904 * Add missing BIO_free at exit of pkcs12 application.
16908 * Fix memory leak in conf.c.
16912 * Updates for Win32 to assembler version of MD5.
16916 * Set #! path to perl in apps/der_chop to where we found it
16917 instead of using a fixed path.
16921 * SHA library changes for irix64-mips4-cc.
16925 * Improvements for VMS support.
16929 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16931 * Bignum library bug fix. IRIX 6 passes "make test" now!
16932 This also avoids the problems with SC4.2 and unpatched SC5.
16934 *Andy Polyakov <appro@fy.chalmers.se>*
16936 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16937 These are required because of the typesafe stack would otherwise break
16938 existing code. If old code used a structure member which used to be STACK
16939 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16940 sk_num or sk_value it would produce an error because the num, data members
16941 are not present in STACK_OF. Now it just produces a warning. sk_set
16942 replaces the old method of assigning a value to sk_value
16943 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16944 that does this will no longer work (and should use sk_set instead) but
16945 this could be regarded as a "questionable" behaviour anyway.
16949 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16950 correctly handle encrypted S/MIME data.
16954 * Change type of various DES function arguments from des_cblock
16955 (which means, in function argument declarations, pointer to char)
16956 to des_cblock * (meaning pointer to array with 8 char elements),
16957 which allows the compiler to do more typechecking; it was like
16958 that back in SSLeay, but with lots of ugly casts.
16960 Introduce new type const_des_cblock.
16964 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16965 problems: find RecipientInfo structure that matches recipient certificate
16966 and initialise the ASN1 structures properly based on passed cipher.
16970 * Belatedly make the BN tests actually check the results.
16974 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16975 to and from BNs: it was completely broken. New compilation option
16976 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16977 key elements as negative integers.
16981 * Reorganize and speed up MD5.
16983 *Andy Polyakov <appro@fy.chalmers.se>*
16987 *Richard Levitte <richard@levitte.org>*
16989 * New option -out to asn1parse to allow the parsed structure to be
16990 output to a file. This is most useful when combined with the -strparse
16991 option to examine the output of things like OCTET STRINGS.
16995 * Make SSL library a little more fool-proof by not requiring any longer
16996 that `SSL_set_{accept,connect}_state` be called before
16997 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16998 in many applications because usually everything *appeared* to work as
16999 intended anyway -- now it really works as intended).
17003 * Move openssl.cnf out of lib/.
17007 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17008 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17009 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17011 *Ralf S. Engelschall*
17013 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17014 handle PKCS#7 enveloped data properly.
17016 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17018 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17019 copying pointers. The cert_st handling is changed by this in
17020 various ways (and thus what used to be known as ctx->default_cert
17021 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17022 any longer when s->cert does not give us what we need).
17023 ssl_cert_instantiate becomes obsolete by this change.
17024 As soon as we've got the new code right (possibly it already is?),
17025 we have solved a couple of bugs of the earlier code where s->cert
17026 was used as if it could not have been shared with other SSL structures.
17028 Note that using the SSL API in certain dirty ways now will result
17029 in different behaviour than observed with earlier library versions:
17030 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17031 does not influence s as it used to.
17033 In order to clean up things more thoroughly, inside SSL_SESSION
17034 we don't use CERT any longer, but a new structure SESS_CERT
17035 that holds per-session data (if available); currently, this is
17036 the peer's certificate chain and, for clients, the server's certificate
17037 and temporary key. CERT holds only those values that can have
17038 meaningful defaults in an SSL_CTX.
17042 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17043 from the internal representation. Various PKCS#7 fixes: remove some
17044 evil casts and set the enc_dig_alg field properly based on the signing
17049 * Allow PKCS#12 password to be set from the command line or the
17050 environment. Let 'ca' get its config file name from the environment
17051 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17056 * Allow certificate policies extension to use an IA5STRING for the
17057 organization field. This is contrary to the PKIX definition but
17058 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17063 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17064 without disallowing inline assembler and the like for non-pedantic builds.
17068 * Support Borland C++ builder.
17070 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17076 * SHA-1 cleanups and performance enhancements.
17078 *Andy Polyakov <appro@fy.chalmers.se>*
17080 * Sparc v8plus assembler for the bignum library.
17082 *Andy Polyakov <appro@fy.chalmers.se>*
17084 * Accept any -xxx and +xxx compiler options in Configure.
17088 * Update HPUX configuration.
17092 * Add missing `sk_<type>_unshift()` function to safestack.h
17094 *Ralf S. Engelschall*
17096 * New function SSL_CTX_use_certificate_chain_file that sets the
17097 "extra_cert"s in addition to the certificate. (This makes sense
17098 only for "PEM" format files, as chains as a whole are not
17103 * Support verify_depth from the SSL API.
17104 x509_vfy.c had what can be considered an off-by-one-error:
17105 Its depth (which was not part of the external interface)
17106 was actually counting the number of certificates in a chain;
17107 now it really counts the depth.
17111 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17112 instead of X509err, which often resulted in confusing error
17113 messages since the error codes are not globally unique
17114 (e.g. an alleged error in ssl3_accept when a certificate
17115 didn't match the private key).
17117 * New function SSL_CTX_set_session_id_context that allows to set a default
17118 value (so that you don't need SSL_set_session_id_context for each
17119 connection using the SSL_CTX).
17123 * OAEP decoding bug fix.
17127 * Support INSTALL_PREFIX for package builders, as proposed by
17132 * New Configure options "threads" and "no-threads". For systems
17133 where the proper compiler options are known (currently Solaris
17134 and Linux), "threads" is the default.
17138 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17142 * Install various scripts to $(OPENSSLDIR)/misc, not to
17143 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17144 such as /usr/local/bin.
17148 * "make linux-shared" to build shared libraries.
17150 *Niels Poppe <niels@netbox.org>*
17152 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17156 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17157 extension adding in x509 utility.
17161 * Remove NOPROTO sections and error code comments.
17165 * Partial rewrite of the DEF file generator to now parse the ANSI
17170 * New Configure options --prefix=DIR and --openssldir=DIR.
17174 * Complete rewrite of the error code script(s). It is all now handled
17175 by one script at the top level which handles error code gathering,
17176 header rewriting and C source file generation. It should be much better
17177 than the old method: it now uses a modified version of Ulf's parser to
17178 read the ANSI prototypes in all header files (thus the old K&R definitions
17179 aren't needed for error creation any more) and do a better job of
17180 translating function codes into names. The old 'ASN1 error code embedded
17181 in a comment' is no longer necessary and it doesn't use .err files which
17182 have now been deleted. Also the error code call doesn't have to appear all
17183 on one line (which resulted in some large lines...).
17187 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17191 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17192 0 (which usually indicates a closed connection), but continue reading.
17196 * Fix some race conditions.
17200 * Add support for CRL distribution points extension. Add Certificate
17201 Policies and CRL distribution points documentation.
17205 * Move the autogenerated header file parts to crypto/opensslconf.h.
17209 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17210 8 of keying material. Merlin has also confirmed interop with this fix
17211 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17213 *Merlin Hughes <merlin@baltimore.ie>*
17215 * Fix lots of warnings.
17217 *Richard Levitte <levitte@stacken.kth.se>*
17219 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17220 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17222 *Richard Levitte <levitte@stacken.kth.se>*
17224 * Fix problems with sizeof(long) == 8.
17226 *Andy Polyakov <appro@fy.chalmers.se>*
17228 * Change functions to ANSI C.
17232 * Fix typos in error codes.
17234 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17236 * Remove defunct assembler files from Configure.
17240 * SPARC v8 assembler BIGNUM implementation.
17242 *Andy Polyakov <appro@fy.chalmers.se>*
17244 * Support for Certificate Policies extension: both print and set.
17245 Various additions to support the r2i method this uses.
17249 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17250 return a const string when you are expecting an allocated buffer.
17254 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17255 types DirectoryString and DisplayText.
17259 * Add code to allow r2i extensions to access the configuration database,
17260 add an LHASH database driver and add several ctx helper functions.
17264 * Fix an evil bug in bn_expand2() which caused various BN functions to
17265 fail when they extended the size of a BIGNUM.
17269 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17270 support typesafe stack.
17274 * Fix typo in SSL_[gs]et_options().
17276 *Nils Frostberg <nils@medcom.se>*
17278 * Delete various functions and files that belonged to the (now obsolete)
17279 old X509V3 handling code.
17283 * New Configure option "rsaref".
17287 * Don't auto-generate pem.h.
17291 * Introduce type-safe ASN.1 SETs.
17295 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17297 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17299 * Introduce type-safe STACKs. This will almost certainly break lots of code
17300 that links with OpenSSL (well at least cause lots of warnings), but fear
17301 not: the conversion is trivial, and it eliminates loads of evil casts. A
17302 few STACKed things have been converted already. Feel free to convert more.
17303 In the fullness of time, I'll do away with the STACK type altogether.
17307 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17308 specified in `<certfile>` by updating the entry in the index.txt file.
17309 This way one no longer has to edit the index.txt file manually for
17310 revoking a certificate. The -revoke option does the gory details now.
17312 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17314 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17315 `-text` option at all and this way the `-noout -text` combination was
17316 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17318 *Ralf S. Engelschall*
17320 * Make sure a corresponding plain text error message exists for the
17321 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17322 verify callback function determined that a certificate was revoked.
17324 *Ralf S. Engelschall*
17326 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17327 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17328 all available ciphers including rc5, which was forgotten until now.
17329 In order to let the testing shell script know which algorithms
17330 are available, a new (up to now undocumented) command
17331 `openssl list-cipher-commands` is used.
17335 * Bugfix: s_client occasionally would sleep in select() when
17336 it should have checked SSL_pending() first.
17340 * New functions DSA_do_sign and DSA_do_verify to provide access to
17341 the raw DSA values prior to ASN.1 encoding.
17345 * Tweaks to Configure
17347 *Niels Poppe <niels@netbox.org>*
17349 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17354 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17358 * New config option to avoid instructions that are illegal on the 80386.
17359 The default code is faster, but requires at least a 486.
17363 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17364 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17365 same as SSL2_VERSION anyway.
17369 * New "-showcerts" option for s_client.
17373 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17374 application. Various cleanups and fixes.
17378 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17379 modify error routines to work internally. Add error codes and PBE init
17380 to library startup routines.
17384 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17385 packing functions to asn1 and evp. Changed function names and error
17386 codes along the way.
17390 * PKCS12 integration: and so it begins... First of several patches to
17391 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17392 objects to objects.h
17396 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17397 and display support for Thawte strong extranet extension.
17401 * Add LinuxPPC support.
17403 *Jeff Dubrule <igor@pobox.org>*
17405 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17406 bn_div_words in alpha.s.
17408 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17410 * Make sure the RSA OAEP test is skipped under -DRSAref because
17411 OAEP isn't supported when OpenSSL is built with RSAref.
17413 *Ulf Moeller <ulf@fitug.de>*
17415 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17416 so they no longer are missing under -DNOPROTO.
17418 *Soren S. Jorvang <soren@t.dk>*
17420 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17422 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17423 doesn't work when the session is reused. Coming soon!
17427 * Fix a security hole, that allows sessions to be reused in the wrong
17428 context thus bypassing client cert protection! All software that uses
17429 client certs and session caches in multiple contexts NEEDS PATCHING to
17430 allow session reuse! A fuller solution is in the works.
17432 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17434 * Some more source tree cleanups (removed obsolete files
17435 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17436 permission on "config" script to be executable) and a fix for the INSTALL
17439 *Ulf Moeller <ulf@fitug.de>*
17441 * Remove some legacy and erroneous uses of malloc, free instead of
17444 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17446 * Make rsa_oaep_test return non-zero on error.
17448 *Ulf Moeller <ulf@fitug.de>*
17450 * Add support for native Solaris shared libraries. Configure
17451 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17452 if someone would make that last step automatic.
17454 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17456 * ctx_size was not built with the right compiler during "make links". Fixed.
17460 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17461 except NULL ciphers". This means the default cipher list will no longer
17462 enable NULL ciphers. They need to be specifically enabled e.g. with
17463 the string "DEFAULT:eNULL".
17467 * Fix to RSA private encryption routines: if p < q then it would
17468 occasionally produce an invalid result. This will only happen with
17469 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17473 * Be less restrictive and allow also `perl util/perlpath.pl
17474 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17475 because this way one can also use an interpreter named `perl5` (which is
17476 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17477 installed as `perl`).
17479 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17481 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17483 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17485 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17486 advapi32.lib to Win32 build and change the pem test comparison
17487 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17488 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17489 and crypto/des/ede_cbcm_enc.c.
17493 * DES quad checksum was broken on big-endian architectures. Fixed.
17497 * Comment out two functions in bio.h that aren't implemented. Fix up the
17498 Win32 test batch file so it (might) work again. The Win32 test batch file
17499 is horrible: I feel ill....
17503 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17504 in e_os.h. Audit of header files to check ANSI and non ANSI
17505 sections: 10 functions were absent from non ANSI section and not exported
17506 from Windows DLLs. Fixed up libeay.num for new functions.
17510 * Make `openssl version` output lines consistent.
17512 *Ralf S. Engelschall*
17514 * Fix Win32 symbol export lists for BIO functions: Added
17515 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17516 to ms/libeay{16,32}.def.
17518 *Ralf S. Engelschall*
17520 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17521 fine under Unix and passes some trivial tests I've now added. But the
17522 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17523 added to make sure no one expects that this stuff really works in the
17524 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17525 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17528 *Ralf S. Engelschall*
17530 * Fix the generation of two part addresses in perl.
17532 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17534 * Add config entry for Linux on MIPS.
17536 *John Tobey <jtobey@channel1.com>*
17538 * Make links whenever Configure is run, unless we are on Windoze.
17542 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17543 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17548 * Add a useful kludge to allow package maintainers to specify compiler and
17549 other platforms details on the command line without having to patch the
17550 Configure script every time: One now can use
17551 `perl Configure <id>:<details>`,
17552 i.e. platform ids are allowed to have details appended
17553 to them (separated by colons). This is treated as there would be a static
17554 pre-configured entry in Configure's %table under key `<id>` with value
17555 `<details>` and `perl Configure <id>` is called. So, when you want to
17556 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17557 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17558 now, which overrides the FreeBSD-elf entry on-the-fly.
17560 *Ralf S. Engelschall*
17562 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17566 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17567 on the `perl Configure ...` command line. This way one can compile
17568 OpenSSL libraries with Position Independent Code (PIC) which is needed
17569 for linking it into DSOs.
17571 *Ralf S. Engelschall*
17573 * Remarkably, export ciphers were totally broken and no-one had noticed!
17578 * Cleaned up the LICENSE document: The official contact for any license
17579 questions now is the OpenSSL core team under openssl-core@openssl.org.
17580 And add a paragraph about the dual-license situation to make sure people
17581 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17582 to the OpenSSL toolkit.
17584 *Ralf S. Engelschall*
17586 * General source tree makefile cleanups: Made `making xxx in yyy...`
17587 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17588 Additionally cleaned up the `make links` target: Remove unnecessary
17589 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17590 to speed processing and no longer clutter the display with confusing
17591 stuff. Instead only the actually done links are displayed.
17593 *Ralf S. Engelschall*
17595 * Permit null encryption ciphersuites, used for authentication only. It used
17596 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17597 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17602 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17603 signed attributes when verifying signatures (this would break them),
17604 the detached data encoding was wrong and public keys obtained using
17605 X509_get_pubkey() weren't freed.
17609 * Add text documentation for the BUFFER functions. Also added a work around
17610 to a Win95 console bug. This was triggered by the password read stuff: the
17611 last character typed gets carried over to the next fread(). If you were
17612 generating a new cert request using 'req' for example then the last
17613 character of the passphrase would be CR which would then enter the first
17618 * Added the new 'Includes OpenSSL Cryptography Software' button as
17619 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17620 button and can be used by applications based on OpenSSL to show the
17621 relationship to the OpenSSL project.
17623 *Ralf S. Engelschall*
17625 * Remove confusing variables in function signatures in files
17626 ssl/ssl_lib.c and ssl/ssl.h.
17628 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17630 * Don't install bss_file.c under PREFIX/include/
17632 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17634 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17635 functions that return function pointers and has support for NT specific
17636 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17637 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17638 unsigned to signed types: this was killing the Win32 compile.
17642 * Add new certificate file to stack functions,
17643 SSL_add_dir_cert_subjects_to_stack() and
17644 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17645 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17646 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17647 This means that Apache-SSL and similar packages don't have to mess around
17648 to add as many CAs as they want to the preferred list.
17652 * Experiment with doxygen documentation. Currently only partially applied to
17654 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17655 openssl.doxy as the configuration file.
17659 * Get rid of remaining C++-style comments which strict C compilers hate.
17661 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17663 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17664 compiled in by default: it has problems with large keys.
17668 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17669 DH private keys and/or callback functions which directly correspond to
17670 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17671 is needed for applications which have to configure certificates on a
17672 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17674 For the RSA certificate situation is makes no difference, but
17675 for the DSA certificate situation this fixes the "no shared cipher"
17676 problem where the OpenSSL cipher selection procedure failed because the
17677 temporary keys were not overtaken from the context and the API provided
17678 no way to reconfigure them.
17679 The new functions now let applications reconfigure the stuff and they
17680 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17681 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17682 non-public-API function ssl_cert_instantiate() is used as a helper
17683 function and also to reduce code redundancy inside ssl_rsa.c.
17685 *Ralf S. Engelschall*
17687 * Move s_server -dcert and -dkey options out of the undocumented feature
17688 area because they are useful for the DSA situation and should be
17689 recognized by the users.
17691 *Ralf S. Engelschall*
17693 * Fix the cipher decision scheme for export ciphers: the export bits are
17694 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17695 SSL_EXP_MASK. So, the original variable has to be used instead of the
17696 already masked variable.
17698 *Richard Levitte <levitte@stacken.kth.se>*
17700 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17702 *Richard Levitte <levitte@stacken.kth.se>*
17704 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17705 from `int` to `unsigned int` because it is a length and initialized by
17706 EVP_DigestFinal() which expects an `unsigned int *`.
17708 *Richard Levitte <levitte@stacken.kth.se>*
17710 * Don't hard-code path to Perl interpreter on shebang line of Configure
17711 script. Instead use the usual Shell->Perl transition trick.
17713 *Ralf S. Engelschall*
17715 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17716 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17717 -noout -modulus` as it's already the case for `openssl rsa -noout
17718 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17719 currently the public key is printed (a decision which was already done by
17720 `openssl dsa -modulus` in the past) which serves a similar purpose.
17721 Additionally the NO_RSA no longer completely removes the whole -modulus
17722 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17725 *Ralf S. Engelschall*
17727 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17728 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17730 *Arne Ansper <arne@ats.cyber.ee>*
17732 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17733 to be added. Now both 'req' and 'ca' can use new objects defined in the
17738 * Add cool BIO that does syslog (or event log on NT).
17740 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17742 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17743 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17744 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17745 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17749 * Add preliminary config info for new extension code.
17753 * Make RSA_NO_PADDING really use no padding.
17755 *Ulf Moeller <ulf@fitug.de>*
17757 * Generate errors when private/public key check is done.
17761 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17762 for some CRL extensions and new objects added.
17766 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17767 key usage extension and fuller support for authority key id.
17771 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17772 padding method for RSA, which is recommended for new applications in PKCS
17773 #1 v2.0 (RFC 2437, October 1998).
17774 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17775 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17776 against Bleichbacher's attack on RSA.
17777 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17780 * Updates to the new SSL compression code
17782 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17784 * Fix so that the version number in the master secret, when passed
17785 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17786 (because the server will not accept higher), that the version number
17787 is 0x03,0x01, not 0x03,0x00
17789 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17791 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17792 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17793 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17797 * Support for RAW extensions where an arbitrary extension can be
17798 created by including its DER encoding. See apps/openssl.cnf for
17803 * Make sure latest Perl versions don't interpret some generated C array
17804 code as Perl array code in the crypto/err/err_genc.pl script.
17806 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17808 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17809 not many people have the assembler. Various Win32 compilation fixes and
17810 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17811 build instructions.
17815 * Modify configure script 'Configure' to automatically create crypto/date.h
17816 file under Win32 and also build pem.h from pem.org. New script
17817 util/mkfiles.pl to create the MINFO file on environments that can't do a
17818 'make files': perl util/mkfiles.pl >MINFO should work.
17822 * Major rework of DES function declarations, in the pursuit of correctness
17823 and purity. As a result, many evil casts evaporated, and some weirdness,
17824 too. You may find this causes warnings in your code. Zapping your evil
17825 casts will probably fix them. Mostly.
17829 * Fix for a typo in asn1.h. Bug fix to object creation script
17830 obj_dat.pl. It considered a zero in an object definition to mean
17831 "end of object": none of the objects in objects.h have any zeros
17832 so it wasn't spotted.
17834 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17836 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17837 Masking (CBCM). In the absence of test vectors, the best I have been able
17838 to do is check that the decrypt undoes the encrypt, so far. Send me test
17839 vectors if you have them.
17843 * Correct calculation of key length for export ciphers (too much space was
17844 allocated for null ciphers). This has not been tested!
17848 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17849 message is now correct (it understands "crypto" and "ssl" on its
17850 command line). There is also now an "update" option. This will update
17851 the util/ssleay.num and util/libeay.num files with any new functions.
17853 perl util/mkdef.pl crypto ssl update
17854 it will update them.
17858 * Overhauled the Perl interface:
17859 - ported BN stuff to OpenSSL's different BN library
17860 - made the perl/ source tree CVS-aware
17861 - renamed the package from SSLeay to OpenSSL (the files still contain
17862 their history because I've copied them in the repository)
17863 - removed obsolete files (the test scripts will be replaced
17864 by better Test::Harness variants in the future)
17866 *Ralf S. Engelschall*
17868 * First cut for a very conservative source tree cleanup:
17869 1. merge various obsolete readme texts into doc/ssleay.txt
17870 where we collect the old documents and readme texts.
17871 2. remove the first part of files where I'm already sure that we no
17872 longer need them because of three reasons: either they are just temporary
17873 files which were left by Eric or they are preserved original files where
17874 I've verified that the diff is also available in the CVS via "cvs diff
17875 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17876 the crypto/md/ stuff).
17878 *Ralf S. Engelschall*
17880 * More extension code. Incomplete support for subject and issuer alt
17881 name, issuer and authority key id. Change the i2v function parameters
17882 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17883 what that's for :-) Fix to ASN1 macro which messed up
17884 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17888 * Preliminary support for ENUMERATED type. This is largely copied from the
17893 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17895 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17897 * Make sure `make rehash` target really finds the `openssl` program.
17899 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17901 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17902 like to hear about it if this slows down other processors.
17906 * Add CygWin32 platform information to Configure script.
17908 *Alan Batie <batie@aahz.jf.intel.com>*
17910 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17912 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17914 * New program nseq to manipulate netscape certificate sequences
17918 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17923 * Fixes to BN code. Previously the default was to define BN_RECURSION
17924 but the BN code had some problems that would cause failures when
17925 doing certificate verification and some other functions.
17927 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17929 * Add ASN1 and PEM code to support netscape certificate sequences.
17933 * Add ASN1 and PEM code to support netscape certificate sequences.
17937 * Add several PKIX and private extended key usage OIDs.
17941 * Modify the 'ca' program to handle the new extension code. Modify
17942 openssl.cnf for new extension format, add comments.
17946 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17947 and add a sample to openssl.cnf so req -x509 now adds appropriate
17952 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17953 error code, add initial support to X509_print() and x509 application.
17957 * Takes a deep breath and start adding X509 V3 extension support code. Add
17958 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17959 stuff is currently isolated and isn't even compiled yet.
17963 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17964 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17965 Removed the versions check from X509 routines when loading extensions:
17966 this allows certain broken certificates that don't set the version
17967 properly to be processed.
17971 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17972 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17973 can still be regenerated with "make depend".
17977 * Spelling mistake in C version of CAST-128.
17979 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17981 * Changes to the error generation code. The perl script err-code.pl
17982 now reads in the old error codes and retains the old numbers, only
17983 adding new ones if necessary. It also only changes the .err files if new
17984 codes are added. The makefiles have been modified to only insert errors
17985 when needed (to avoid needlessly modifying header files). This is done
17986 by only inserting errors if the .err file is newer than the auto generated
17987 C file. To rebuild all the error codes from scratch (the old behaviour)
17988 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17989 or delete all the .err files.
17993 * CAST-128 was incorrectly implemented for short keys. The C version has
17994 been fixed, but is untested. The assembler versions are also fixed, but
17995 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17996 to regenerate it if needed.
17997 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17998 Hagino <itojun@kame.net>*
18000 * File was opened incorrectly in randfile.c.
18002 *Ulf Möller <ulf@fitug.de>*
18004 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18005 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18006 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18007 al: it's just almost always a UTCTime. Note this patch adds new error
18008 codes so do a "make errors" if there are problems.
18012 * Correct Linux 1 recognition in config.
18014 *Ulf Möller <ulf@fitug.de>*
18016 * Remove pointless MD5 hash when using DSA keys in ca.
18018 *Anonymous <nobody@replay.com>*
18020 * Generate an error if given an empty string as a cert directory. Also
18021 generate an error if handed NULL (previously returned 0 to indicate an
18022 error, but didn't set one).
18024 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18026 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18030 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18031 parameters. This was causing a warning which killed off the Win32 compile.
18035 * Remove C++ style comments from crypto/bn/bn_local.h.
18037 *Neil Costigan <neil.costigan@celocom.com>*
18039 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18040 based on a text string, looking up short and long names and finally
18041 "dot" format. The "dot" format stuff didn't work. Added new function
18042 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18043 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18044 OID is not part of the table.
18048 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18049 X509_LOOKUP_by_alias().
18053 * Sort openssl functions by name.
18057 * Get the gendsa program working (hopefully) and add it to app list. Remove
18058 encryption from sample DSA keys (in case anyone is interested the password
18063 * Make *all* `*_free` functions accept a NULL pointer.
18065 *Frans Heymans <fheymans@isaserver.be>*
18067 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18070 *Anonymous <nobody@replay.com>*
18072 * s_server should send the CAfile as acceptable CAs, not its own cert.
18074 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18076 * Don't blow it for numeric -newkey arguments to apps/req.
18078 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18080 * Temp key "for export" tests were wrong in s3_srvr.c.
18082 *Anonymous <nobody@replay.com>*
18084 * Add prototype for temp key callback functions
18085 SSL_CTX_set_tmp_{rsa,dh}_callback().
18089 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18090 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18094 * X509_name_add_entry() freed the wrong thing after an error.
18096 *Arne Ansper <arne@ats.cyber.ee>*
18098 * rsa_eay.c would attempt to free a NULL context.
18100 *Arne Ansper <arne@ats.cyber.ee>*
18102 * BIO_s_socket() had a broken should_retry() on Windoze.
18104 *Arne Ansper <arne@ats.cyber.ee>*
18106 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18108 *Arne Ansper <arne@ats.cyber.ee>*
18110 * Make sure the already existing X509_STORE->depth variable is initialized
18111 in X509_STORE_new(), but document the fact that this variable is still
18112 unused in the certificate verification process.
18114 *Ralf S. Engelschall*
18116 * Fix the various library and apps files to free up pkeys obtained from
18117 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18121 * Fix reference counting in X509_PUBKEY_get(). This makes
18122 demos/maurice/example2.c work, amongst others, probably.
18124 *Steve Henson and Ben Laurie*
18126 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18127 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18128 are no longer created. This way we have a single and consistent command
18129 line interface `openssl <command>`, similar to `cvs <command>`.
18131 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18133 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18134 BIT STRING wrapper always have zero unused bits.
18138 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18142 * Make the top-level INSTALL documentation easier to understand.
18146 * Makefiles updated to exit if an error occurs in a sub-directory
18147 make (including if user presses ^C) [Paul Sutton]
18149 * Make Montgomery context stuff explicit in RSA data structure.
18153 * Fix build order of pem and err to allow for generated pem.h.
18157 * Fix renumbering bug in X509_NAME_delete_entry().
18161 * Enhanced the err-ins.pl script so it makes the error library number
18162 global and can add a library name. This is needed for external ASN1 and
18163 other error libraries.
18167 * Fixed sk_insert which never worked properly.
18171 * Fix ASN1 macros so they can handle indefinite length constructed
18172 EXPLICIT tags. Some non standard certificates use these: they can now
18177 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18178 into a single doc/ssleay.txt bundle. This way the information is still
18179 preserved but no longer messes up this directory. Now it's new room for
18180 the new set of documentation files.
18182 *Ralf S. Engelschall*
18184 * SETs were incorrectly DER encoded. This was a major pain, because they
18185 shared code with SEQUENCEs, which aren't coded the same. This means that
18186 almost everything to do with SETs or SEQUENCEs has either changed name or
18187 number of arguments.
18189 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18191 * Fix test data to work with the above.
18195 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18196 was already fixed by Eric for 0.9.1 it seems.
18198 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18200 * Autodetect FreeBSD3.
18204 * Fix various bugs in Configure. This affects the following platforms:
18208 unixware-2.0-pentium
18213 * Eliminate generated files from CVS. Reorder tests to regenerate files
18214 before they are needed.
18218 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18222 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18224 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18225 changed SSLeay to OpenSSL in version strings.
18227 *Ralf S. Engelschall*
18229 * Some fixups to the top-level documents.
18233 * Fixed the nasty bug where rsaref.h was not found under compile-time
18234 because the symlink to include/ was missing.
18236 *Ralf S. Engelschall*
18238 * Incorporated the popular no-RSA/DSA-only patches
18239 which allow to compile a RSA-free SSLeay.
18241 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18243 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18244 when "ssleay" is still not found.
18246 *Ralf S. Engelschall*
18248 * Added more platforms to Configure: Cray T3E, HPUX 11,
18250 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18252 * Updated the README file.
18254 *Ralf S. Engelschall*
18256 * Added various .cvsignore files in the CVS repository subdirs
18257 to make a "cvs update" really silent.
18259 *Ralf S. Engelschall*
18261 * Recompiled the error-definition header files and added
18262 missing symbols to the Win32 linker tables.
18264 *Ralf S. Engelschall*
18266 * Cleaned up the top-level documents;
18267 o new files: CHANGES and LICENSE
18268 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18269 o merged COPYRIGHT into LICENSE
18270 o removed obsolete TODO file
18271 o renamed MICROSOFT to INSTALL.W32
18273 *Ralf S. Engelschall*
18275 * Removed dummy files from the 0.9.1b source tree:
18276 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18277 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18278 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18279 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18280 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18282 *Ralf S. Engelschall*
18284 * Added various platform portability fixes.
18288 * The Genesis of the OpenSSL rpject:
18289 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18290 Young and Tim J. Hudson created while they were working for C2Net until
18293 *The OpenSSL Project*
18295 ### Changes between 0.9.0b and 0.9.1b [not released]
18297 * Updated a few CA certificates under certs/
18301 * Changed some BIGNUM api stuff.
18305 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18306 DGUX x86, Linux Alpha, etc.
18310 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18311 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18316 * Add -strparse option to asn1pars program which parses nested
18319 *Dr Stephen Henson <shenson@bigfoot.com>*
18321 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18325 * DSA fix for "ca" program.
18329 * Added "-genkey" option to "dsaparam" program.
18333 * Added RIPE MD160 (rmd160) message digest.
18337 * Added -a (all) option to "ssleay version" command.
18341 * Added PLATFORM define which is the id given to Configure.
18345 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18349 * Extended the ASN.1 parser routines.
18353 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18357 * Added a BN_CTX to the BN library.
18361 * Fixed the weak key values in DES library
18365 * Changed API in EVP library for cipher aliases.
18369 * Added support for RC2/64bit cipher.
18373 * Converted the lhash library to the crypto/mem.c functions.
18377 * Added more recognized ASN.1 object ids.
18381 * Added more RSA padding checks for SSL/TLS.
18385 * Added BIO proxy/filter functionality.
18389 * Added extra_certs to SSL_CTX which can be used
18390 send extra CA certificates to the client in the CA cert chain sending
18391 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18395 * Now Fortezza is denied in the authentication phase because
18396 this is key exchange mechanism is not supported by SSLeay at all.
18400 * Additional PKCS1 checks.
18404 * Support the string "TLSv1" for all TLS v1 ciphers.
18408 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18409 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18413 * Fixed a few memory leaks.
18417 * Fixed various code and comment typos.
18421 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18422 bytes sent in the client random.
18424 *Edward Bishop <ebishop@spyglass.com>*
18428 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18429 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18430 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18431 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18432 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18433 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18434 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18435 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18436 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18437 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18438 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18439 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18440 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18441 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18442 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18443 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18444 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18445 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18446 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18447 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18448 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18449 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18450 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18451 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18452 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18453 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18454 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18455 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18456 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18457 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18458 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18459 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18460 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18461 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18462 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18463 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18464 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18465 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18466 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18467 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18468 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18469 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18470 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18471 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18472 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18473 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18474 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18475 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18476 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18477 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18478 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18479 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18480 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18481 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18482 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18483 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18484 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18485 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18486 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18487 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18488 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18489 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18490 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18491 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18492 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18493 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18494 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18495 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18496 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18497 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18498 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18499 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18500 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18501 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18502 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18503 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18504 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18505 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18506 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18507 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18508 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18509 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18510 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18511 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18512 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18513 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18514 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18515 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18516 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18517 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18518 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18519 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18520 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18521 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18522 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18523 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18524 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18525 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18526 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18527 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18528 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18529 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18530 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18531 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18532 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18533 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18534 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18535 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18536 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18537 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18538 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18539 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18540 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18541 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18542 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18543 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18544 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18545 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18546 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18547 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18548 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18549 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18550 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18551 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18552 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18553 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18554 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18555 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18556 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18557 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18558 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18559 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18560 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18561 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18562 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18563 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18564 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18565 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18566 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18567 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18568 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18569 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18570 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18571 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18572 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18573 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18574 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18575 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18576 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18577 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18578 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18579 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18580 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18581 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18582 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18583 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18584 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18585 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18586 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18587 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655