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 * Dropped interactive mode from the 'openssl' program. From now on,
27 the `openssl` command without arguments is equivalent to `openssl
32 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
33 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
34 While the old function names have been retained for backward compatibility
35 they should not be used in new developments
36 because their return values are confusing: Unlike other `_cmp()` functions
37 they do not return 0 in case their arguments are equal.
41 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
42 EC_GROUP_get_field_type().
46 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
47 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
48 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
49 Applications should rely on the library automatically assigning a suitable
50 EC_METHOD internally upon EC_GROUP construction.
54 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
55 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
56 assigned internally without application intervention.
57 Users of EC_GROUP_new() should switch to a different suitable constructor.
61 * Add CAdES-BES signature verification support, mostly derived
62 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
64 *Filipe Raimundo da Silva*
66 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
70 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
71 functions are not widely used and now OpenSSL automatically perform this
72 conversion when needed.
76 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
77 EC_KEY_precompute_mult(). These functions are not widely used and
78 applications should instead switch to named curves which OpenSSL has
79 hardcoded lookup tables for.
83 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
84 should instead use the L<EC_POINT_mul(3)> function.
88 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
89 that are not applicable to the new provider model. Applications should
90 instead use EVP_default_properties_is_fips_enabled() and
91 EVP_default_properties_enable_fips().
95 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
96 is set, an unexpected EOF is ignored, it pretends a close notify was received
97 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
101 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
102 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
103 used and applications should instead use the
104 L<EC_POINT_set_affine_coordinates(3)> and
105 L<EC_POINT_get_affine_coordinates(3)> functions.
109 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
110 arrays to be more easily constructed via a series of utility functions.
111 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
112 the various push functions and finally convert to a passable OSSL_PARAM
113 array using OSSL_PARAM_BLD_to_param().
117 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
118 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
119 internal keys, if they correspond to one of those built in types.
123 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
124 contain a provider side internal key.
128 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
129 They are old functions that we don't use, and that you could disable with
130 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
134 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
135 have been converted to Markdown with the goal to produce documents
136 which not only look pretty when viewed online in the browser, but
137 remain well readable inside a plain text editor.
139 To achieve this goal, a 'minimalistic' Markdown style has been applied
140 which avoids formatting elements that interfere too much with the
141 reading flow in the text file. For example, it
143 * avoids [ATX headings][] and uses [setext headings][] instead
144 (which works for `<h1>` and `<h2>` headings only).
145 * avoids [inline links][] and uses [reference links][] instead.
146 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
148 [ATX headings]: https://github.github.com/gfm/#atx-headings
149 [setext headings]: https://github.github.com/gfm/#setext-headings
150 [inline links]: https://github.github.com/gfm/#inline-link
151 [reference links]: https://github.github.com/gfm/#reference-link
152 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
153 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
155 *Matthias St. Pierre*
157 * The test suite is changed to preserve results of each test recipe.
158 A new directory test-runs/ with subdirectories named like the
159 test recipes are created in the build tree for this purpose.
163 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
164 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
165 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
167 *David von Oheimb, Martin Peylo*
169 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
170 The legacy OCSP-focused and only partly documented API is retained.
171 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
175 * All of the low level RSA functions have been deprecated including:
177 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
178 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
179 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
180 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
181 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
182 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
183 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
184 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
185 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
186 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
187 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
188 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
189 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
190 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
191 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
192 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
193 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
194 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
195 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
196 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
197 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
198 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
199 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
200 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
201 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
202 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
203 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
204 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
206 Use of these low level functions has been informally discouraged for a long
207 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
208 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
209 L<EVP_PKEY_decrypt(3)>.
213 * X509 certificates signed using SHA1 are no longer allowed at security
215 In TLS/SSL the default security level is 1. It can be set either
216 using the cipher string with @SECLEVEL, or calling
217 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
218 a call to SSL_CTX_use_certificate() will fail if the security level is not
220 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
221 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
226 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
227 modified to use PKEY APIs. These commands are now in maintenance mode
228 and no new features will be added to them.
232 * The command line utility rsautl has been deprecated.
233 Instead use the pkeyutl program.
237 * The command line utilities genrsa and rsa have been modified to use PKEY
238 APIs These commands are now in maintenance mode and no new features will
243 * All of the low level DH functions have been deprecated including:
245 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
246 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
247 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
248 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
249 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
250 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
251 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
252 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
253 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
254 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
255 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
256 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
257 DH_meth_get_generate_params and DH_meth_set_generate_params.
259 Use of these low level functions has been informally discouraged for a long
260 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
261 and L<EVP_PKEY_derive(3)>.
265 * All of the low level DSA functions have been deprecated including:
267 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
268 DSA_get_default_method, DSA_set_method, DSA_get_method,
269 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
270 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
271 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
272 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
273 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
274 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
275 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
276 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
277 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
278 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
279 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
282 Use of these low level functions has been informally discouraged for a long
283 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
284 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
288 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
289 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
290 This means that applications don't have to look at the curve NID and
291 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
292 However, they still can, that EVP_PKEY_set_alias_type() call acts as
293 a no-op when the EVP_PKEY is already of the given type.
295 Parameter and key generation is also reworked to make it possible
296 to generate EVP_PKEY_SM2 parameters and keys without having to go
297 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
298 However, code that does the latter will still work as before.
302 * Deprecated low level ECDH and ECDSA functions. These include:
304 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
305 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
308 Use of these low level functions has been informally discouraged for a long
309 time. Instead applications should use the EVP_PKEY_derive(3),
310 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
314 * Deprecated the EC_KEY_METHOD functions. These include:
316 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
317 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
318 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
319 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
320 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
321 EC_KEY_METHOD_get_verify.
323 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
327 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
328 and EVP_PKEY_decrypt() instead.
329 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
330 and EVP_PKEY_encrypt() instead.
334 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
335 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
336 a new formulation to include all the things it can be used for,
337 as well as words of caution.
341 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
342 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
346 * All of the low level HMAC functions have been deprecated including:
348 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
349 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
352 Use of these low level functions has been informally discouraged for a long
353 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
354 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
355 and L<EVP_MAC_final(3)>.
359 * Over two thousand fixes were made to the documentation, including:
360 - Common options (such as -rand/-writerand, TLS version control, etc)
361 were refactored and point to newly-enhanced descriptions in openssl.pod.
362 - Added style conformance for all options (with help from Richard Levitte),
363 documented all reported missing options, added a CI build to check
364 that all options are documented and that no unimplemented options
366 - Documented some internals, such as all use of environment variables.
367 - Addressed all internal broken L<> references.
371 * All of the low level CMAC functions have been deprecated including:
373 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
374 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
376 Use of these low level functions has been informally discouraged for a long
377 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
378 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
379 and L<EVP_MAC_final(3)>.
383 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
384 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
387 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
388 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
389 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
390 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
391 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
392 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
393 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
394 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
395 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
396 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
398 Use of these low level functions has been informally discouraged
399 for a long time. Applications should use the EVP_DigestInit_ex(3),
400 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
404 * Corrected the documentation of the return values from the `EVP_DigestSign*`
405 set of functions. The documentation mentioned negative values for some
406 errors, but this was never the case, so the mention of negative values
409 Code that followed the documentation and thereby check with something
410 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
414 * All of the low level cipher functions have been deprecated including:
416 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
417 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
418 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
419 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
420 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
421 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
422 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
423 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
424 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
425 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
426 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
427 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
428 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
429 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
430 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
431 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
432 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
433 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
434 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
435 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
436 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
437 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
438 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
439 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
440 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
441 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
442 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
443 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
444 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
446 Use of these low level functions has been informally discouraged for
447 a long time. Applications should use the high level EVP APIs, e.g.
448 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
449 equivalently named decrypt functions instead.
451 *Matt Caswell and Paul Dale*
453 * Removed include/openssl/opensslconf.h.in and replaced it with
454 include/openssl/configuration.h.in, which differs in not including
455 <openssl/macros.h>. A short header include/openssl/opensslconf.h
456 was added to include both.
458 This allows internal hacks where one might need to modify the set
459 of configured macros, for example this if deprecated symbols are
460 still supposed to be available internally:
462 #include <openssl/configuration.h>
464 #undef OPENSSL_NO_DEPRECATED
465 #define OPENSSL_SUPPRESS_DEPRECATED
467 #include <openssl/macros.h>
469 This should not be used by applications that use the exported
470 symbols, as that will lead to linking errors.
474 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
475 used in exponentiation with 512-bit moduli. No EC algorithms are
476 affected. Analysis suggests that attacks against 2-prime RSA1024,
477 3-prime RSA1536, and DSA1024 as a result of this defect would be very
478 difficult to perform and are not believed likely. Attacks against DH512
479 are considered just feasible. However, for an attack the target would
480 have to re-use the DH512 private key, which is not recommended anyway.
481 Also applications directly using the low level API BN_mod_exp may be
482 affected if they use BN_FLG_CONSTTIME.
487 * Most memory-debug features have been deprecated, and the functionality
488 replaced with no-ops.
492 * Added documentation for the STACK API. OpenSSL only defines the STACK
493 functions where they are used.
497 * Introduced a new method type and API, OSSL_SERIALIZER, to
498 represent generic serializers. An implementation is expected to
499 be able to serialize an object associated with a given name (such
500 as an algorithm name for an asymmetric key) into forms given by
501 implementation properties.
503 Serializers are primarily used from inside libcrypto, through
504 calls to functions like EVP_PKEY_print_private(),
505 PEM_write_bio_PrivateKey() and similar.
507 Serializers are specified in such a way that they can be made to
508 directly handle the provider side portion of an object, if this
509 provider side part comes from the same provider as the serializer
510 itself, but can also be made to handle objects in parametrized
511 form (as an OSSL_PARAM array of data). This allows a provider to
512 offer generic serializers as a service for any other provider.
516 * Added a .pragma directive to the syntax of configuration files, to
517 allow varying behavior in a supported and predictable manner.
518 Currently added pragma:
522 This allows dollar signs to be a keyword character unless it's
523 followed by a opening brace or parenthesis. This is useful for
524 platforms where dollar signs are commonly used in names, such as
525 volume names and system directory names on VMS.
529 * Added functionality to create an EVP_PKEY from user data. This
530 is effectively the same as creating a RSA, DH or DSA object and
531 then assigning them to an EVP_PKEY, but directly using algorithm
532 agnostic EVP functions. A benefit is that this should be future
533 proof for public key algorithms to come.
537 * Change the interpretation of the '--api' configuration option to
538 mean that this is a desired API compatibility level with no
539 further meaning. The previous interpretation, that this would
540 also mean to remove all deprecated symbols up to and including
541 the given version, no requires that 'no-deprecated' is also used
542 in the configuration.
544 When building applications, the desired API compatibility level
545 can be set with the OPENSSL_API_COMPAT macro like before. For
546 API compatibility version below 3.0, the old style numerical
547 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
548 For version 3.0 and on, the value is expected to be the decimal
549 value calculated from the major and minor version like this:
551 MAJOR * 10000 + MINOR * 100
555 -DOPENSSL_API_COMPAT=30000 For 3.0
556 -DOPENSSL_API_COMPAT=30200 For 3.2
558 To hide declarations that are deprecated up to and including the
559 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
560 given when building the application as well.
564 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
565 access to certificate and CRL stores via URIs and OSSL_STORE
568 This adds the following functions:
570 - X509_LOOKUP_store()
571 - X509_STORE_load_file()
572 - X509_STORE_load_path()
573 - X509_STORE_load_store()
574 - SSL_add_store_cert_subjects_to_stack()
575 - SSL_CTX_set_default_verify_store()
576 - SSL_CTX_load_verify_file()
577 - SSL_CTX_load_verify_dir()
578 - SSL_CTX_load_verify_store()
582 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
583 The presence of this system service is determined at run-time.
587 * Added functionality to create an EVP_PKEY context based on data
588 for methods from providers. This takes an algorithm name and a
589 property query string and simply stores them, with the intent
590 that any operation that uses this context will use those strings
591 to fetch the needed methods implicitly, thereby making the port
592 of application written for pre-3.0 OpenSSL easier.
596 * The undocumented function NCONF_WIN32() has been deprecated; for
597 conversion details see the HISTORY section of doc/man5/config.pod
601 * Introduced the new functions EVP_DigestSignInit_ex() and
602 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
603 EVP_DigestVerifyUpdate() have been converted to functions. See the man
604 pages for further details.
608 * Over two thousand fixes were made to the documentation, including:
609 adding missing command flags, better style conformance, documentation
612 *Rich Salz, Richard Levitte*
614 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
615 X25519, X448, Ed25519 and Ed448.
619 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
624 * Deprecated the public definition of ERR_STATE as well as the function
625 ERR_get_state(). This is done in preparation of making ERR_STATE an
630 * Added ERR functionality to give callers access to the stored function
631 names that have replaced the older function code based functions.
633 New functions are ERR_get_error_func(), ERR_peek_error_func(),
634 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
635 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
636 and ERR_peek_last_error_all().
638 These functions have become deprecated: ERR_get_error_line_data(),
639 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
640 ERR_func_error_string().
644 * Extended testing to be verbose for failing tests only. The make variables
645 VERBOSE_FAILURE or VF can be used to enable this:
647 $ make VF=1 test # Unix
648 $ mms /macro=(VF=1) test ! OpenVMS
649 $ nmake VF=1 test # Windows
653 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
654 used even when parsing explicit parameters, when loading a serialized key
655 or calling `EC_GROUP_new_from_ecpkparameters()`/
656 `EC_GROUP_new_from_ecparameters()`.
657 This prevents bypass of security hardening and performance gains,
658 especially for curves with specialized EC_METHODs.
659 By default, if a key encoded with explicit parameters is loaded and later
660 serialized, the output is still encoded with explicit parameters, even if
661 internally a "named" EC_GROUP is used for computation.
665 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
666 this change, EC_GROUP_set_generator would accept order and/or cofactor as
667 NULL. After this change, only the cofactor parameter can be NULL. It also
668 does some minimal sanity checks on the passed order.
673 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
674 An attack is simple, if the first CMS_recipientInfo is valid but the
675 second CMS_recipientInfo is chosen ciphertext. If the second
676 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
677 encryption key will be replaced by garbage, and the message cannot be
678 decoded, but if the RSA decryption fails, the correct encryption key is
679 used and the recipient will not notice the attack.
680 As a work around for this potential attack the length of the decrypted
681 key must be equal to the cipher default key length, in case the
682 certifiate is not given and all recipientInfo are tried out.
683 The old behaviour can be re-enabled in the CMS code by setting the
684 CMS_DEBUG_DECRYPT flag.
688 * Early start up entropy quality from the DEVRANDOM seed source has been
689 improved for older Linux systems. The RAND subsystem will wait for
690 /dev/random to be producing output before seeding from /dev/urandom.
691 The seeded state is stored for future library initialisations using
692 a system global shared memory segment. The shared memory identifier
693 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
694 the desired value. The default identifier is 114.
698 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
699 when primes for RSA keys are computed.
700 Since we previously always generated primes == 2 (mod 3) for RSA keys,
701 the 2-prime and 3-prime RSA modules were easy to distinguish, since
702 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
703 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
704 This avoids possible fingerprinting of newly generated RSA modules.
708 * Correct the extended master secret constant on EBCDIC systems. Without this
709 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
710 negotiate EMS will fail. Unfortunately this also means that TLS connections
711 between EBCDIC systems with this fix, and EBCDIC systems without this
712 fix will fail if they negotiate EMS.
716 * Changed the library initialisation so that the config file is now loaded
717 by default. This was already the case for libssl. It now occurs for both
718 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
719 OPENSSL_init_crypto() to suppress automatic loading of a config file.
723 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
724 where the former acts as a replacement for ERR_put_error(), and the
725 latter replaces the combination ERR_put_error()+ERR_add_error_data().
726 ERR_raise_data() adds more flexibility by taking a format string and
727 an arbitrary number of arguments following it, to be processed with
732 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
733 to check if a named provider is loaded and available. When called, it
734 will also activate all fallback providers if such are still present.
738 * Enforce a minimum DH modulus size of 512 bits.
742 * Changed DH parameters to generate the order q subgroup instead of 2q.
743 Previously generated DH parameters are still accepted by DH_check
744 but DH_generate_key works around that by clearing bit 0 of the
745 private key for those. This avoids leaking bit 0 of the private key.
749 * Significantly reduce secure memory usage by the randomness pools.
753 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
758 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
759 algorithms. An implementation of a key exchange algorithm can be obtained
760 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
761 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
762 the older EVP_PKEY_derive_init() function. See the man pages for the new
763 functions for further details.
767 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
771 * Removed the function names from error messages and deprecated the
774 * Removed NextStep support and the macro OPENSSL_UNISTD
778 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
779 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
780 Also removed "export var as function" capability; we do not export
781 variables, only functions.
785 * RC5_32_set_key has been changed to return an int type, with 0 indicating
786 an error and 1 indicating success. In previous versions of OpenSSL this
787 was a void type. If a key was set longer than the maximum possible this
792 * Support SM2 signing and verification schemes with X509 certificate.
796 * Use SHA256 as the default digest for TS query in the ts app.
800 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
801 This checks that the salt length is at least 128 bits, the derived key
802 length is at least 112 bits, and that the iteration count is at least 1000.
803 For backwards compatibility these checks are disabled by default in the
804 default provider, but are enabled by default in the fips provider.
805 To enable or disable these checks use the control
806 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
810 * Default cipher lists/suites are now available via a function, the
811 #defines are deprecated.
815 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
816 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
817 for Windows Store apps easier. Also, the "no-uplink" option has been added.
821 * Join the directories crypto/x509 and crypto/x509v3
825 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
826 This changes the size when using the genpkey app when no size is given. It
827 fixes an omission in earlier changes that changed all RSA, DSA and DH
828 generation apps to use 2048 bits by default.
832 * Added command 'openssl kdf' that uses the EVP_KDF API.
836 * Added command 'openssl mac' that uses the EVP_MAC API.
840 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
841 as default directories. Also added the command 'openssl info'
842 for scripting purposes.
846 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
847 deprecated. These undocumented functions were never integrated into the EVP
848 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
849 Bi-directional IGE mode. These modes were never formally standardised and
850 usage of these functions is believed to be very small. In particular
851 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
852 is ever used. The security implications are believed to be minimal, but
853 this issue was never fixed for backwards compatibility reasons. New code
854 should not use these modes.
858 * Add prediction resistance to the DRBG reseeding process.
862 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
863 mandated by IEEE Std 1619-2018.
867 * Added newline escaping functionality to a filename when using openssl dgst.
868 This output format is to replicate the output format found in the `*sum`
869 checksum programs. This aims to preserve backward compatibility.
871 *Matt Eaton, Richard Levitte, and Paul Dale*
873 * Removed the heartbeat message in DTLS feature, as it has very
874 little usage and doesn't seem to fulfill a valuable purpose.
875 The configuration option is now deprecated.
879 * Changed the output of 'openssl {digestname} < file' to display the
880 digest name in its output.
884 * Added a new generic trace API which provides support for enabling
885 instrumentation through trace output. This feature is mainly intended
886 as an aid for developers and is disabled by default. To utilize it,
887 OpenSSL needs to be configured with the `enable-trace` option.
889 If the tracing API is enabled, the application can activate trace output
890 by registering BIOs as trace channels for a number of tracing and debugging
893 The 'openssl' application has been expanded to enable any of the types
894 available via environment variables defined by the user, and serves as
895 one possible example on how to use this functionality.
897 *Richard Levitte & Matthias St. Pierre*
899 * Added build tests for C++. These are generated files that only do one
900 thing, to include one public OpenSSL head file each. This tests that
901 the public header files can be usefully included in a C++ application.
903 This test isn't enabled by default. It can be enabled with the option
904 'enable-buildtest-c++'.
908 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
912 * Add KMAC to EVP_MAC.
916 * Added property based algorithm implementation selection framework to
921 * Added SCA hardening for modular field inversion in EC_GROUP through
922 a new dedicated field_inv() pointer in EC_METHOD.
923 This also addresses a leakage affecting conversions from projective
924 to affine coordinates.
926 *Billy Bob Brumley, Nicola Tuveri*
928 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
929 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
930 those algorithms that were already supported through the EVP_PKEY API
931 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
932 and scrypt are now wrappers that call EVP_KDF.
936 * Build devcrypto engine as a dynamic engine.
940 * Add keyed BLAKE2 to EVP_MAC.
944 * Fix a bug in the computation of the endpoint-pair shared secret used
945 by DTLS over SCTP. This breaks interoperability with older versions
946 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
947 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
948 interoperability with such broken implementations. However, enabling
949 this switch breaks interoperability with correct implementations.
951 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
952 re-used X509_PUBKEY object if the second PUBKEY is malformed.
956 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
960 * Change the license to the Apache License v2.0.
964 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
966 - Major releases (indicated by incrementing the MAJOR release number)
967 may introduce incompatible API/ABI changes.
968 - Minor releases (indicated by incrementing the MINOR release number)
969 may introduce new features but retain API/ABI compatibility.
970 - Patch releases (indicated by incrementing the PATCH number)
971 are intended for bug fixes and other improvements of existing
972 features only (like improving performance or adding documentation)
973 and retain API/ABI compatibility.
977 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
981 * Remove the 'dist' target and add a tarball building script. The
982 'dist' target has fallen out of use, and it shouldn't be
983 necessary to configure just to create a source distribution.
987 * Recreate the OS390-Unix config target. It no longer relies on a
988 special script like it did for OpenSSL pre-1.1.0.
992 * Instead of having the source directories listed in Configure, add
993 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
998 * Add GMAC to EVP_MAC.
1002 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1006 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1007 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1008 to facilitate the continued use of MACs through raw private keys in
1009 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
1013 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1014 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1018 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1019 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1020 are retained for backwards compatibility.
1024 * AES-XTS mode now enforces that its two keys are different to mitigate
1025 the attacked described in "Efficient Instantiations of Tweakable
1026 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1027 Details of this attack can be obtained from:
1028 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1032 * Rename the object files, i.e. give them other names than in previous
1033 versions. Their names now include the name of the final product, as
1034 well as its type mnemonic (bin, lib, shlib).
1038 * Added new option for 'openssl list', '-objects', which will display the
1039 list of built in objects, i.e. OIDs with names.
1043 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1044 improves application performance by removing data copies and providing
1045 applications with zero-copy system calls such as sendfile and splice.
1052 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1054 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1056 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1057 while reading in libssl then we would report an error back to the
1058 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1059 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1060 therefore give a hint as to what went wrong.
1064 * Check that ed25519 and ed448 are allowed by the security level. Previously
1065 signature algorithms not using an MD were not being checked that they were
1066 allowed by the security level.
1070 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1071 was not quite right. The behaviour was not consistent between resumption
1072 and normal handshakes, and also not quite consistent with historical
1073 behaviour. The behaviour in various scenarios has been clarified and
1074 it has been updated to make it match historical behaviour as closely as
1079 * *[VMS only]* The header files that the VMS compilers include automatically,
1080 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1081 that the C++ compiler doesn't understand. This is a shortcoming in the
1082 compiler, but can be worked around with `__cplusplus` guards.
1084 C++ applications that use OpenSSL libraries must be compiled using the
1085 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1086 functions. Otherwise, only functions with symbols of less than 31
1087 characters can be used, as the linker will not be able to successfully
1088 resolve symbols with longer names.
1092 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1093 The presence of this system service is determined at run-time.
1097 * Added newline escaping functionality to a filename when using openssl dgst.
1098 This output format is to replicate the output format found in the `*sum`
1099 checksum programs. This aims to preserve backward compatibility.
1101 *Matt Eaton, Richard Levitte, and Paul Dale*
1103 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1108 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1110 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1111 number generator (RNG). This was intended to include protection in the
1112 event of a fork() system call in order to ensure that the parent and child
1113 processes did not share the same RNG state. However this protection was not
1114 being used in the default case.
1116 A partial mitigation for this issue is that the output from a high
1117 precision timer is mixed into the RNG state so the likelihood of a parent
1118 and child process sharing state is significantly reduced.
1120 If an application already calls OPENSSL_init_crypto() explicitly using
1121 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1124 *Matthias St. Pierre*
1126 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1127 used even when parsing explicit parameters, when loading a serialized key
1128 or calling `EC_GROUP_new_from_ecpkparameters()`/
1129 `EC_GROUP_new_from_ecparameters()`.
1130 This prevents bypass of security hardening and performance gains,
1131 especially for curves with specialized EC_METHODs.
1132 By default, if a key encoded with explicit parameters is loaded and later
1133 serialized, the output is still encoded with explicit parameters, even if
1134 internally a "named" EC_GROUP is used for computation.
1138 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1139 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1140 NULL. After this change, only the cofactor parameter can be NULL. It also
1141 does some minimal sanity checks on the passed order.
1146 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1147 An attack is simple, if the first CMS_recipientInfo is valid but the
1148 second CMS_recipientInfo is chosen ciphertext. If the second
1149 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1150 encryption key will be replaced by garbage, and the message cannot be
1151 decoded, but if the RSA decryption fails, the correct encryption key is
1152 used and the recipient will not notice the attack.
1153 As a work around for this potential attack the length of the decrypted
1154 key must be equal to the cipher default key length, in case the
1155 certifiate is not given and all recipientInfo are tried out.
1156 The old behaviour can be re-enabled in the CMS code by setting the
1157 CMS_DEBUG_DECRYPT flag.
1162 * Early start up entropy quality from the DEVRANDOM seed source has been
1163 improved for older Linux systems. The RAND subsystem will wait for
1164 /dev/random to be producing output before seeding from /dev/urandom.
1165 The seeded state is stored for future library initialisations using
1166 a system global shared memory segment. The shared memory identifier
1167 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1168 the desired value. The default identifier is 114.
1172 * Correct the extended master secret constant on EBCDIC systems. Without this
1173 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1174 negotiate EMS will fail. Unfortunately this also means that TLS connections
1175 between EBCDIC systems with this fix, and EBCDIC systems without this
1176 fix will fail if they negotiate EMS.
1180 * Use Windows installation paths in the mingw builds
1182 Mingw isn't a POSIX environment per se, which means that Windows
1183 paths should be used for installation.
1188 * Changed DH_check to accept parameters with order q and 2q subgroups.
1189 With order 2q subgroups the bit 0 of the private key is not secret
1190 but DH_generate_key works around that by clearing bit 0 of the
1191 private key for those. This avoids leaking bit 0 of the private key.
1195 * Significantly reduce secure memory usage by the randomness pools.
1199 * Revert the DEVRANDOM_WAIT feature for Linux systems
1201 The DEVRANDOM_WAIT feature added a select() call to wait for the
1202 /dev/random device to become readable before reading from the
1203 /dev/urandom device.
1205 It turned out that this change had negative side effects on
1206 performance which were not acceptable. After some discussion it
1207 was decided to revert this feature and leave it up to the OS
1208 resp. the platform maintainer to ensure a proper initialization
1209 during early boot time.
1211 *Matthias St. Pierre*
1213 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1215 * Add build tests for C++. These are generated files that only do one
1216 thing, to include one public OpenSSL head file each. This tests that
1217 the public header files can be usefully included in a C++ application.
1219 This test isn't enabled by default. It can be enabled with the option
1220 'enable-buildtest-c++'.
1224 * Enable SHA3 pre-hashing for ECDSA and DSA.
1228 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1229 This changes the size when using the genpkey app when no size is given. It
1230 fixes an omission in earlier changes that changed all RSA, DSA and DH
1231 generation apps to use 2048 bits by default.
1235 * Reorganize the manual pages to consistently have RETURN VALUES,
1236 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1237 util/fix-doc-nits accordingly.
1239 *Paul Yang, Joshua Lock*
1241 * Add the missing accessor EVP_PKEY_get0_engine()
1245 * Have apps like 's_client' and 's_server' output the signature scheme
1246 along with other cipher suite parameters when debugging.
1250 * Make OPENSSL_config() error agnostic again.
1254 * Do the error handling in RSA decryption constant time.
1258 * Prevent over long nonces in ChaCha20-Poly1305.
1260 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1261 for every encryption operation. RFC 7539 specifies that the nonce value
1262 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1263 and front pads the nonce with 0 bytes if it is less than 12
1264 bytes. However it also incorrectly allows a nonce to be set of up to 16
1265 bytes. In this case only the last 12 bytes are significant and any
1266 additional leading bytes are ignored.
1268 It is a requirement of using this cipher that nonce values are
1269 unique. Messages encrypted using a reused nonce value are susceptible to
1270 serious confidentiality and integrity attacks. If an application changes
1271 the default nonce length to be longer than 12 bytes and then makes a
1272 change to the leading bytes of the nonce expecting the new value to be a
1273 new unique nonce then such an application could inadvertently encrypt
1274 messages with a reused nonce.
1276 Additionally the ignored bytes in a long nonce are not covered by the
1277 integrity guarantee of this cipher. Any application that relies on the
1278 integrity of these ignored leading bytes of a long nonce may be further
1279 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1280 is safe because no such use sets such a long nonce value. However user
1281 applications that use this cipher directly and set a non-default nonce
1282 length to be longer than 12 bytes may be vulnerable.
1284 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1290 * Add DEVRANDOM_WAIT feature for Linux systems
1292 On older Linux systems where the getrandom() system call is not available,
1293 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1294 Contrary to getrandom(), the /dev/urandom device will not block during
1295 early boot when the kernel CSPRNG has not been seeded yet.
1297 To mitigate this known weakness, use select() to wait for /dev/random to
1298 become readable before reading from /dev/urandom.
1300 * Ensure that SM2 only uses SM3 as digest algorithm
1304 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1306 * Change the info callback signals for the start and end of a post-handshake
1307 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1308 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1309 confused by this and assume that a TLSv1.2 renegotiation has started. This
1310 can break KeyUpdate handling. Instead we no longer signal the start and end
1311 of a post handshake message exchange (although the messages themselves are
1312 still signalled). This could break some applications that were expecting
1313 the old signals. However without this KeyUpdate is not usable for many
1318 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1320 * Timing vulnerability in DSA signature generation
1322 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1323 timing side channel attack. An attacker could use variations in the signing
1324 algorithm to recover the private key.
1326 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1331 * Timing vulnerability in ECDSA signature generation
1333 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1334 timing side channel attack. An attacker could use variations in the signing
1335 algorithm to recover the private key.
1337 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1342 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1343 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1344 of two gigabytes and the error handling improved.
1346 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1347 categorized as a normal bug, not a security issue, because the DRBG reseeds
1348 automatically and is fully functional even without additional randomness
1349 provided by the application.
1351 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1353 * Add a new ClientHello callback. Provides a callback interface that gives
1354 the application the ability to adjust the nascent SSL object at the
1355 earliest stage of ClientHello processing, immediately after extensions have
1356 been collected but before they have been processed. In particular, this
1357 callback can adjust the supported TLS versions in response to the contents
1362 * Add SM2 base algorithm support.
1366 * s390x assembly pack: add (improved) hardware-support for the following
1367 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1368 aes-cfb/cfb8, aes-ecb.
1372 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1373 parameter is no longer accepted, as it leads to a corrupt table. NULL
1374 pem_str is reserved for alias entries only.
1378 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1379 step for prime curves. The new implementation is based on formulae from
1380 differential addition-and-doubling in homogeneous projective coordinates
1381 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1382 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1383 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1384 to work in projective coordinates.
1386 *Billy Bob Brumley, Nicola Tuveri*
1388 * Change generating and checking of primes so that the error rate of not
1389 being prime depends on the intended use based on the size of the input.
1390 For larger primes this will result in more rounds of Miller-Rabin.
1391 The maximal error rate for primes with more than 1080 bits is lowered
1394 *Kurt Roeckx, Annie Yousar*
1396 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1400 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1401 moving between systems, and to avoid confusion when a Windows build is
1402 done with mingw vs with MSVC. For POSIX installs, there's still a
1403 symlink or copy named 'tsget' to avoid that confusion as well.
1407 * Revert blinding in ECDSA sign and instead make problematic addition
1408 length-invariant. Switch even to fixed-length Montgomery multiplication.
1412 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1413 step for binary curves. The new implementation is based on formulae from
1414 differential addition-and-doubling in mixed Lopez-Dahab projective
1415 coordinates, modified to independently blind the operands.
1417 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1419 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1420 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1421 EC_METHODs to implement their own specialized "ladder step", to take
1422 advantage of more favorable coordinate systems or more efficient
1423 differential addition-and-doubling algorithms.
1425 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1427 * Modified the random device based seed sources to keep the relevant
1428 file descriptors open rather than reopening them on each access.
1429 This allows such sources to operate in a chroot() jail without
1430 the associated device nodes being available. This behaviour can be
1431 controlled using RAND_keep_random_devices_open().
1435 * Numerous side-channel attack mitigations have been applied. This may have
1436 performance impacts for some algorithms for the benefit of improved
1437 security. Specific changes are noted in this change log by their respective
1442 * AIX shared library support overhaul. Switch to AIX "natural" way of
1443 handling shared libraries, which means collecting shared objects of
1444 different versions and bitnesses in one common archive. This allows to
1445 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1446 doesn't affect the way 3rd party applications are linked, only how
1447 multi-version installation is managed.
1451 * Make ec_group_do_inverse_ord() more robust and available to other
1452 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1453 mitigations are applied to the fallback BN_mod_inverse().
1454 When using this function rather than BN_mod_inverse() directly, new
1455 EC cryptosystem implementations are then safer-by-default.
1459 * Add coordinate blinding for EC_POINT and implement projective
1460 coordinate blinding for generic prime curves as a countermeasure to
1461 chosen point SCA attacks.
1463 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1465 * Add blinding to ECDSA and DSA signatures to protect against side channel
1466 attacks discovered by Keegan Ryan (NCC Group).
1470 * Enforce checking in the pkeyutl command line app to ensure that the input
1471 length does not exceed the maximum supported digest length when performing
1472 a sign, verify or verifyrecover operation.
1476 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1477 I/O in combination with something like select() or poll() will hang. This
1478 can be turned off again using SSL_CTX_clear_mode().
1479 Many applications do not properly handle non-application data records, and
1480 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1481 around the problems in those applications, but can also break some.
1482 It's recommended to read the manpages about SSL_read(), SSL_write(),
1483 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1484 SSL_CTX_set_read_ahead() again.
1488 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1489 now allow empty (zero character) pass phrases.
1493 * Apply blinding to binary field modular inversion and remove patent
1494 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1498 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1499 binary and prime elliptic curves.
1503 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1504 constant time fixed point multiplication.
1508 * Revise elliptic curve scalar multiplication with timing attack
1509 defenses: ec_wNAF_mul redirects to a constant time implementation
1510 when computing fixed point and variable point multiplication (which
1511 in OpenSSL are mostly used with secret scalars in keygen, sign,
1512 ECDH derive operations).
1513 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1516 * Updated CONTRIBUTING
1520 * Updated DRBG / RAND to request nonce and additional low entropy
1521 randomness from the system.
1523 *Matthias St. Pierre*
1525 * Updated 'openssl rehash' to use OpenSSL consistent default.
1529 * Moved the load of the ssl_conf module to libcrypto, which helps
1530 loading engines that libssl uses before libssl is initialised.
1534 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1538 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1540 *Ingo Schwarze, Rich Salz*
1542 * Added output of accepting IP address and port for 'openssl s_server'
1546 * Added a new API for TLSv1.3 ciphersuites:
1547 SSL_CTX_set_ciphersuites()
1548 SSL_set_ciphersuites()
1552 * Memory allocation failures consistently add an error to the error
1557 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1558 in libcrypto when run as setuid/setgid.
1562 * Load any config file by default when libssl is used.
1566 * Added new public header file <openssl/rand_drbg.h> and documentation
1567 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1569 *Matthias St. Pierre*
1571 * QNX support removed (cannot find contributors to get their approval
1572 for the license change).
1576 * TLSv1.3 replay protection for early data has been implemented. See the
1577 SSL_read_early_data() man page for further details.
1581 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1582 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1583 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1584 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1585 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1586 configuration has been separated out. See the ciphers man page or the
1587 SSL_CTX_set_ciphersuites() man page for more information.
1591 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1592 in responder mode now supports the new "-multi" option, which
1593 spawns the specified number of child processes to handle OCSP
1594 requests. The "-timeout" option now also limits the OCSP
1595 responder's patience to wait to receive the full client request
1596 on a newly accepted connection. Child processes are respawned
1597 as needed, and the CA index file is automatically reloaded
1598 when changed. This makes it possible to run the "ocsp" responder
1599 as a long-running service, making the OpenSSL CA somewhat more
1600 feature-complete. In this mode, most diagnostic messages logged
1601 after entering the event loop are logged via syslog(3) rather than
1606 * Added support for X448 and Ed448. Heavily based on original work by
1611 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1612 objects loaded. This adds the functions OSSL_STORE_expect() and
1613 OSSL_STORE_find() as well as needed tools to construct searches and
1614 get the search data out of them.
1618 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1619 version of OpenSSL should review their configuration settings to ensure
1620 that they are still appropriate for TLSv1.3. For further information see:
1621 <https://wiki.openssl.org/index.php/TLS1.3>
1625 * Grand redesign of the OpenSSL random generator
1627 The default RAND method now utilizes an AES-CTR DRBG according to
1628 NIST standard SP 800-90Ar1. The new random generator is essentially
1629 a port of the default random generator from the OpenSSL FIPS 2.0
1630 object module. It is a hybrid deterministic random bit generator
1631 using an AES-CTR bit stream and which seeds and reseeds itself
1632 automatically using trusted system entropy sources.
1634 Some of its new features are:
1635 - Support for multiple DRBG instances with seed chaining.
1636 - The default RAND method makes use of a DRBG.
1637 - There is a public and private DRBG instance.
1638 - The DRBG instances are fork-safe.
1639 - Keep all global DRBG instances on the secure heap if it is enabled.
1640 - The public and private DRBG instance are per thread for lock free
1643 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1645 * Changed Configure so it only says what it does and doesn't dump
1646 so much data. Instead, ./configdata.pm should be used as a script
1647 to display all sorts of configuration data.
1651 * Added processing of "make variables" to Configure.
1655 * Added SHA512/224 and SHA512/256 algorithm support.
1659 * The last traces of Netware support, first removed in 1.1.0, have
1664 * Get rid of Makefile.shared, and in the process, make the processing
1665 of certain files (rc.obj, or the .def/.map/.opt files produced from
1666 the ordinal files) more visible and hopefully easier to trace and
1667 debug (or make silent).
1671 * Make it possible to have environment variable assignments as
1672 arguments to config / Configure.
1676 * Add multi-prime RSA (RFC 8017) support.
1680 * Add SM3 implemented according to GB/T 32905-2016
1681 * Jack Lloyd <jack.lloyd@ribose.com>,
1682 Ronald Tse <ronald.tse@ribose.com>,
1683 Erick Borsboom <erick.borsboom@ribose.com> *
1685 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1686 as documented in RFC6066.
1687 Based on a patch from Tomasz Moń
1689 *Filipe Raimundo da Silva*
1691 * Add SM4 implemented according to GB/T 32907-2016.
1692 * Jack Lloyd <jack.lloyd@ribose.com>,
1693 Ronald Tse <ronald.tse@ribose.com>,
1694 Erick Borsboom <erick.borsboom@ribose.com> *
1696 * Reimplement -newreq-nodes and ERR_error_string_n; the
1697 original author does not agree with the license change.
1701 * Add ARIA AEAD TLS support.
1705 * Some macro definitions to support VS6 have been removed. Visual
1706 Studio 6 has not worked since 1.1.0
1710 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1711 without clearing the errors.
1715 * Add "atfork" functions. If building on a system that without
1716 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1717 requirements. The RAND facility now uses/requires this.
1725 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1726 not possible to disable entirely. However, it's still possible to
1727 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1730 To disable, configure with 'no-ui-console'. 'no-ui' is still
1731 possible to use as an alias. Check at compile time with the
1732 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1733 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1737 * Add a STORE module, which implements a uniform and URI based reader of
1738 stores that can contain keys, certificates, CRLs and numerous other
1739 objects. The main API is loosely based on a few stdio functions,
1740 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1741 OSSL_STORE_error and OSSL_STORE_close.
1742 The implementation uses backends called "loaders" to implement arbitrary
1743 URI schemes. There is one built in "loader" for the 'file' scheme.
1747 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1748 then adjusted to work on FreeBSD 8.4 as well.
1749 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1750 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1754 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1755 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1756 error code calls like this:
1758 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1760 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1761 that can be encoded in C. For the foreseeable future, this will only
1764 *Richard Levitte and Tim Hudson*
1766 * Removed BSD cryptodev engine.
1770 * Add a build target 'build_all_generated', to build all generated files
1771 and only that. This can be used to prepare everything that requires
1772 things like perl for a system that lacks perl and then move everything
1773 to that system and do the rest of the build there.
1777 * In the UI interface, make it possible to duplicate the user data. This
1778 can be used by engines that need to retain the data for a longer time
1779 than just the call where this user data is passed.
1783 * Ignore the '-named_curve auto' value for compatibility of applications
1786 *Tomas Mraz <tmraz@fedoraproject.org>*
1788 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1789 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1790 alerts across multiple records (some of which could be empty). In practice
1791 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1792 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1793 support this at all. Supporting it adds significant complexity to the
1794 record layer, and its removal is unlikely to cause interoperability
1799 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1800 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1801 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1806 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1807 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1809 *Richard Levitte, Andy Polyakov*
1811 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1816 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1817 platform rather than 'mingw'.
1821 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1822 success if they are asked to add an object which already exists
1823 in the store. This change cascades to other functions which load
1824 certificates and CRLs.
1828 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1829 facilitate stack unwinding even from assembly subroutines.
1833 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1834 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1838 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1839 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1840 which is the minimum version we support.
1844 * Certificate time validation (X509_cmp_time) enforces stricter
1845 compliance with RFC 5280. Fractional seconds and timezone offsets
1846 are no longer allowed.
1850 * Add support for ARIA
1854 * s_client will now send the Server Name Indication (SNI) extension by
1855 default unless the new "-noservername" option is used. The server name is
1856 based on the host provided to the "-connect" option unless overridden by
1857 using "-servername".
1861 * Add support for SipHash
1865 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1866 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1867 prevent issues where no progress is being made and the peer continually
1868 sends unrecognised record types, using up resources processing them.
1872 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1873 using the algorithm defined in
1874 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1878 * Heartbeat support has been removed; the ABI is changed for now.
1880 *Richard Levitte, Rich Salz*
1882 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1886 * The RSA "null" method, which was partially supported to avoid patent
1887 issues, has been replaced to always returns NULL.
1894 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
1896 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1897 used even when parsing explicit parameters, when loading a serialized key
1898 or calling `EC_GROUP_new_from_ecpkparameters()`/
1899 `EC_GROUP_new_from_ecparameters()`.
1900 This prevents bypass of security hardening and performance gains,
1901 especially for curves with specialized EC_METHODs.
1902 By default, if a key encoded with explicit parameters is loaded and later
1903 serialized, the output is still encoded with explicit parameters, even if
1904 internally a "named" EC_GROUP is used for computation.
1908 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1909 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1910 NULL. After this change, only the cofactor parameter can be NULL. It also
1911 does some minimal sanity checks on the passed order.
1916 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1917 An attack is simple, if the first CMS_recipientInfo is valid but the
1918 second CMS_recipientInfo is chosen ciphertext. If the second
1919 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1920 encryption key will be replaced by garbage, and the message cannot be
1921 decoded, but if the RSA decryption fails, the correct encryption key is
1922 used and the recipient will not notice the attack.
1923 As a work around for this potential attack the length of the decrypted
1924 key must be equal to the cipher default key length, in case the
1925 certifiate is not given and all recipientInfo are tried out.
1926 The old behaviour can be re-enabled in the CMS code by setting the
1927 CMS_DEBUG_DECRYPT flag.
1932 * Use Windows installation paths in the mingw builds
1934 Mingw isn't a POSIX environment per se, which means that Windows
1935 paths should be used for installation.
1940 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
1942 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1943 This changes the size when using the genpkey app when no size is given. It
1944 fixes an omission in earlier changes that changed all RSA, DSA and DH
1945 generation apps to use 2048 bits by default.
1949 * Prevent over long nonces in ChaCha20-Poly1305.
1951 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1952 for every encryption operation. RFC 7539 specifies that the nonce value
1953 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1954 and front pads the nonce with 0 bytes if it is less than 12
1955 bytes. However it also incorrectly allows a nonce to be set of up to 16
1956 bytes. In this case only the last 12 bytes are significant and any
1957 additional leading bytes are ignored.
1959 It is a requirement of using this cipher that nonce values are
1960 unique. Messages encrypted using a reused nonce value are susceptible to
1961 serious confidentiality and integrity attacks. If an application changes
1962 the default nonce length to be longer than 12 bytes and then makes a
1963 change to the leading bytes of the nonce expecting the new value to be a
1964 new unique nonce then such an application could inadvertently encrypt
1965 messages with a reused nonce.
1967 Additionally the ignored bytes in a long nonce are not covered by the
1968 integrity guarantee of this cipher. Any application that relies on the
1969 integrity of these ignored leading bytes of a long nonce may be further
1970 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1971 is safe because no such use sets such a long nonce value. However user
1972 applications that use this cipher directly and set a non-default nonce
1973 length to be longer than 12 bytes may be vulnerable.
1975 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1981 * Added SCA hardening for modular field inversion in EC_GROUP through
1982 a new dedicated field_inv() pointer in EC_METHOD.
1983 This also addresses a leakage affecting conversions from projective
1984 to affine coordinates.
1986 *Billy Bob Brumley, Nicola Tuveri*
1988 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1989 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1993 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1997 * Remove the 'dist' target and add a tarball building script. The
1998 'dist' target has fallen out of use, and it shouldn't be
1999 necessary to configure just to create a source distribution.
2003 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2005 * Timing vulnerability in DSA signature generation
2007 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2008 timing side channel attack. An attacker could use variations in the signing
2009 algorithm to recover the private key.
2011 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2016 * Timing vulnerability in ECDSA signature generation
2018 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2019 timing side channel attack. An attacker could use variations in the signing
2020 algorithm to recover the private key.
2022 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2027 * Add coordinate blinding for EC_POINT and implement projective
2028 coordinate blinding for generic prime curves as a countermeasure to
2029 chosen point SCA attacks.
2031 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2033 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2035 * Client DoS due to large DH parameter
2037 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2038 malicious server can send a very large prime value to the client. This will
2039 cause the client to spend an unreasonably long period of time generating a
2040 key for this prime resulting in a hang until the client has finished. This
2041 could be exploited in a Denial Of Service attack.
2043 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2048 * Cache timing vulnerability in RSA Key Generation
2050 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2051 a cache timing side channel attack. An attacker with sufficient access to
2052 mount cache timing attacks during the RSA key generation process could
2053 recover the private key.
2055 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2056 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2061 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2062 parameter is no longer accepted, as it leads to a corrupt table. NULL
2063 pem_str is reserved for alias entries only.
2067 * Revert blinding in ECDSA sign and instead make problematic addition
2068 length-invariant. Switch even to fixed-length Montgomery multiplication.
2072 * Change generating and checking of primes so that the error rate of not
2073 being prime depends on the intended use based on the size of the input.
2074 For larger primes this will result in more rounds of Miller-Rabin.
2075 The maximal error rate for primes with more than 1080 bits is lowered
2078 *Kurt Roeckx, Annie Yousar*
2080 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2084 * Add blinding to ECDSA and DSA signatures to protect against side channel
2085 attacks discovered by Keegan Ryan (NCC Group).
2089 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2090 now allow empty (zero character) pass phrases.
2094 * Certificate time validation (X509_cmp_time) enforces stricter
2095 compliance with RFC 5280. Fractional seconds and timezone offsets
2096 are no longer allowed.
2100 * Fixed a text canonicalisation bug in CMS
2102 Where a CMS detached signature is used with text content the text goes
2103 through a canonicalisation process first prior to signing or verifying a
2104 signature. This process strips trailing space at the end of lines, converts
2105 line terminators to CRLF and removes additional trailing line terminators
2106 at the end of a file. A bug in the canonicalisation process meant that
2107 some characters, such as form-feed, were incorrectly treated as whitespace
2108 and removed. This is contrary to the specification (RFC5485). This fix
2109 could mean that detached text data signed with an earlier version of
2110 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2111 signed with a fixed OpenSSL may fail to verify with an earlier version of
2112 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2113 and use the "-binary" flag (for the "cms" command line application) or set
2114 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2118 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2120 * Constructed ASN.1 types with a recursive definition could exceed the stack
2122 Constructed ASN.1 types with a recursive definition (such as can be found
2123 in PKCS7) could eventually exceed the stack given malicious input with
2124 excessive recursion. This could result in a Denial Of Service attack. There
2125 are no such structures used within SSL/TLS that come from untrusted sources
2126 so this is considered safe.
2128 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2134 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2136 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2137 effectively reduced to only comparing the least significant bit of each
2138 byte. This allows an attacker to forge messages that would be considered as
2139 authenticated in an amount of tries lower than that guaranteed by the
2140 security claims of the scheme. The module can only be compiled by the
2141 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2143 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2149 * Add a build target 'build_all_generated', to build all generated files
2150 and only that. This can be used to prepare everything that requires
2151 things like perl for a system that lacks perl and then move everything
2152 to that system and do the rest of the build there.
2156 * Backport SSL_OP_NO_RENGOTIATION
2158 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2159 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2160 changes this is no longer possible in 1.1.0. Therefore the new
2161 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2162 1.1.0 to provide equivalent functionality.
2164 Note that if an application built against 1.1.0h headers (or above) is run
2165 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2166 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2170 * Removed the OS390-Unix config target. It relied on a script that doesn't
2175 * rsaz_1024_mul_avx2 overflow bug on x86_64
2177 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2178 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2179 Analysis suggests that attacks against RSA and DSA as a result of this
2180 defect would be very difficult to perform and are not believed likely.
2181 Attacks against DH1024 are considered just feasible, because most of the
2182 work necessary to deduce information about a private key may be performed
2183 offline. The amount of resources required for such an attack would be
2184 significant. However, for an attack on TLS to be meaningful, the server
2185 would have to share the DH1024 private key among multiple clients, which is
2186 no longer an option since CVE-2016-0701.
2188 This only affects processors that support the AVX2 but not ADX extensions
2189 like Intel Haswell (4th generation).
2191 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2192 was originally found via the OSS-Fuzz project.
2197 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2199 * bn_sqrx8x_internal carry bug on x86_64
2201 There is a carry propagating bug in the x86_64 Montgomery squaring
2202 procedure. No EC algorithms are affected. Analysis suggests that attacks
2203 against RSA and DSA as a result of this defect would be very difficult to
2204 perform and are not believed likely. Attacks against DH are considered just
2205 feasible (although very difficult) because most of the work necessary to
2206 deduce information about a private key may be performed offline. The amount
2207 of resources required for such an attack would be very significant and
2208 likely only accessible to a limited number of attackers. An attacker would
2209 additionally need online access to an unpatched system using the target
2210 private key in a scenario with persistent DH parameters and a private
2211 key that is shared between multiple clients.
2213 This only affects processors that support the BMI1, BMI2 and ADX extensions
2214 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2216 This issue was reported to OpenSSL by the OSS-Fuzz project.
2221 * Malformed X.509 IPAddressFamily could cause OOB read
2223 If an X.509 certificate has a malformed IPAddressFamily extension,
2224 OpenSSL could do a one-byte buffer overread. The most likely result
2225 would be an erroneous display of the certificate in text format.
2227 This issue was reported to OpenSSL by the OSS-Fuzz project.
2232 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2234 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2235 platform rather than 'mingw'.
2239 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2240 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2241 which is the minimum version we support.
2245 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2247 * Encrypt-Then-Mac renegotiation crash
2249 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2250 negotiated where it was not in the original handshake (or vice-versa) then
2251 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2252 and servers are affected.
2254 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2259 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2261 * Truncated packet could crash via OOB read
2263 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2264 cipher is being used, then a truncated packet can cause that host to
2265 perform an out-of-bounds read, usually resulting in a crash.
2267 This issue was reported to OpenSSL by Robert Święcki of Google.
2272 * Bad (EC)DHE parameters cause a client crash
2274 If a malicious server supplies bad parameters for a DHE or ECDHE key
2275 exchange then this can result in the client attempting to dereference a
2276 NULL pointer leading to a client crash. This could be exploited in a Denial
2279 This issue was reported to OpenSSL by Guido Vranken.
2284 * BN_mod_exp may produce incorrect results on x86_64
2286 There is a carry propagating bug in the x86_64 Montgomery squaring
2287 procedure. No EC algorithms are affected. Analysis suggests that attacks
2288 against RSA and DSA as a result of this defect would be very difficult to
2289 perform and are not believed likely. Attacks against DH are considered just
2290 feasible (although very difficult) because most of the work necessary to
2291 deduce information about a private key may be performed offline. The amount
2292 of resources required for such an attack would be very significant and
2293 likely only accessible to a limited number of attackers. An attacker would
2294 additionally need online access to an unpatched system using the target
2295 private key in a scenario with persistent DH parameters and a private
2296 key that is shared between multiple clients. For example this can occur by
2297 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2298 similar to CVE-2015-3193 but must be treated as a separate problem.
2300 This issue was reported to OpenSSL by the OSS-Fuzz project.
2305 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2307 * ChaCha20/Poly1305 heap-buffer-overflow
2309 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2310 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2311 crash. This issue is not considered to be exploitable beyond a DoS.
2313 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2318 * CMS Null dereference
2320 Applications parsing invalid CMS structures can crash with a NULL pointer
2321 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2322 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2323 structure callback if an attempt is made to free certain invalid encodings.
2324 Only CHOICE structures using a callback which do not handle NULL value are
2327 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2332 * Montgomery multiplication may produce incorrect results
2334 There is a carry propagating bug in the Broadwell-specific Montgomery
2335 multiplication procedure that handles input lengths divisible by, but
2336 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2337 and DH private keys are impossible. This is because the subroutine in
2338 question is not used in operations with the private key itself and an input
2339 of the attacker's direct choice. Otherwise the bug can manifest itself as
2340 transient authentication and key negotiation failures or reproducible
2341 erroneous outcome of public-key operations with specially crafted input.
2342 Among EC algorithms only Brainpool P-512 curves are affected and one
2343 presumably can attack ECDH key negotiation. Impact was not analyzed in
2344 detail, because pre-requisites for attack are considered unlikely. Namely
2345 multiple clients have to choose the curve in question and the server has to
2346 share the private key among them, neither of which is default behaviour.
2347 Even then only clients that chose the curve will be affected.
2349 This issue was publicly reported as transient failures and was not
2350 initially recognized as a security issue. Thanks to Richard Morgan for
2351 providing reproducible case.
2356 * Removed automatic addition of RPATH in shared libraries and executables,
2357 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2361 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2363 * Fix Use After Free for large message sizes
2365 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2366 message larger than approx 16k is received then the underlying buffer to
2367 store the incoming message is reallocated and moved. Unfortunately a
2368 dangling pointer to the old location is left which results in an attempt to
2369 write to the previously freed location. This is likely to result in a
2370 crash, however it could potentially lead to execution of arbitrary code.
2372 This issue only affects OpenSSL 1.1.0a.
2374 This issue was reported to OpenSSL by Robert Święcki.
2379 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2381 * OCSP Status Request extension unbounded memory growth
2383 A malicious client can send an excessively large OCSP Status Request
2384 extension. If that client continually requests renegotiation, sending a
2385 large OCSP Status Request extension each time, then there will be unbounded
2386 memory growth on the server. This will eventually lead to a Denial Of
2387 Service attack through memory exhaustion. Servers with a default
2388 configuration are vulnerable even if they do not support OCSP. Builds using
2389 the "no-ocsp" build time option are not affected.
2391 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2396 * SSL_peek() hang on empty record
2398 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2399 sends an empty record. This could be exploited by a malicious peer in a
2400 Denial Of Service attack.
2402 This issue was reported to OpenSSL by Alex Gaynor.
2407 * Excessive allocation of memory in tls_get_message_header() and
2408 dtls1_preprocess_fragment()
2410 A (D)TLS message includes 3 bytes for its length in the header for the
2411 message. This would allow for messages up to 16Mb in length. Messages of
2412 this length are excessive and OpenSSL includes a check to ensure that a
2413 peer is sending reasonably sized messages in order to avoid too much memory
2414 being consumed to service a connection. A flaw in the logic of version
2415 1.1.0 means that memory for the message is allocated too early, prior to
2416 the excessive message length check. Due to way memory is allocated in
2417 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2418 to service a connection. This could lead to a Denial of Service through
2419 memory exhaustion. However, the excessive message length check still takes
2420 place, and this would cause the connection to immediately fail. Assuming
2421 that the application calls SSL_free() on the failed connection in a timely
2422 manner then the 21Mb of allocated memory will then be immediately freed
2423 again. Therefore the excessive memory allocation will be transitory in
2424 nature. This then means that there is only a security impact if:
2426 1) The application does not call SSL_free() in a timely manner in the event
2427 that the connection fails
2429 2) The application is working in a constrained environment where there is
2430 very little free memory
2432 3) The attacker initiates multiple connection attempts such that there are
2433 multiple connections in a state where memory has been allocated for the
2434 connection; SSL_free() has not yet been called; and there is insufficient
2435 memory to service the multiple requests.
2437 Except in the instance of (1) above any Denial Of Service is likely to be
2438 transitory because as soon as the connection fails the memory is
2439 subsequently freed again in the SSL_free() call. However there is an
2440 increased risk during this period of application crashes due to the lack of
2441 memory - which would then mean a more serious Denial of Service.
2443 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2444 (CVE-2016-6307 and CVE-2016-6308)
2448 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2449 had to be removed. Primary reason is that vendor assembler can't
2450 assemble our modules with -KPIC flag. As result it, assembly
2451 support, was not even available as option. But its lack means
2452 lack of side-channel resistant code, which is incompatible with
2453 security by todays standards. Fortunately gcc is readily available
2454 prepackaged option, which we firmly point at...
2458 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2460 * Windows command-line tool supports UTF-8 opt-in option for arguments
2461 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2462 (to any value) allows Windows user to access PKCS#12 file generated
2463 with Windows CryptoAPI and protected with non-ASCII password, as well
2464 as files generated under UTF-8 locale on Linux also protected with
2469 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2470 have been disabled by default and removed from DEFAULT, just like RC4.
2471 See the RC4 item below to re-enable both.
2475 * The method for finding the storage location for the Windows RAND seed file
2476 has changed. First we check %RANDFILE%. If that is not set then we check
2477 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2478 all else fails we fall back to C:\.
2482 * The EVP_EncryptUpdate() function has had its return type changed from void
2483 to int. A return of 0 indicates and error while a return of 1 indicates
2488 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2489 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2490 off the constant time implementation for RSA, DSA and DH have been made
2491 no-ops and deprecated.
2495 * Windows RAND implementation was simplified to only get entropy by
2496 calling CryptGenRandom(). Various other RAND-related tickets
2499 *Joseph Wylie Yandle, Rich Salz*
2501 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2502 and `OPENSSL_LH_`, respectively. The old names are available
2503 with API compatibility. They new names are now completely documented.
2507 * Unify TYPE_up_ref(obj) methods signature.
2508 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2509 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2510 int (instead of void) like all others TYPE_up_ref() methods.
2511 So now these methods also check the return value of CRYPTO_atomic_add(),
2512 and the validity of object reference counter.
2514 *fdasilvayy@gmail.com*
2516 * With Windows Visual Studio builds, the .pdb files are installed
2517 alongside the installed libraries and executables. For a static
2518 library installation, ossl_static.pdb is the associate compiler
2519 generated .pdb file to be used when linking programs.
2523 * Remove openssl.spec. Packaging files belong with the packagers.
2527 * Automatic Darwin/OSX configuration has had a refresh, it will now
2528 recognise x86_64 architectures automatically. You can still decide
2529 to build for a different bitness with the environment variable
2530 KERNEL_BITS (can be 32 or 64), for example:
2532 KERNEL_BITS=32 ./config
2536 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2537 256 bit AES and HMAC with SHA256.
2541 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2545 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2549 * To enable users to have their own config files and build file templates,
2550 Configure looks in the directory indicated by the environment variable
2551 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2552 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2553 name and is used as is.
2557 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2558 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2559 X509_CERT_FILE_CTX was removed.
2563 * "shared" builds are now the default. To create only static libraries use
2564 the "no-shared" Configure option.
2568 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2569 All of these option have not worked for some while and are fundamental
2574 * Make various cleanup routines no-ops and mark them as deprecated. Most
2575 global cleanup functions are no longer required because they are handled
2576 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2577 Explicitly de-initing can cause problems (e.g. where a library that uses
2578 OpenSSL de-inits, but an application is still using it). The affected
2579 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2580 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2581 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2582 COMP_zlib_cleanup().
2586 * --strict-warnings no longer enables runtime debugging options
2587 such as REF_DEBUG. Instead, debug options are automatically
2588 enabled with '--debug' builds.
2590 *Andy Polyakov, Emilia Käsper*
2592 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2593 have been moved out of the public header files. New functions for managing
2594 these have been added.
2598 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2599 objects have been moved out of the public header files. New
2600 functions for managing these have been added.
2604 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2605 have been moved out of the public header files. New functions for managing
2606 these have been added.
2610 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2611 moved out of the public header files. New functions for managing these
2616 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2620 * Removed the mk1mf build scripts.
2624 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2625 it is always safe to #include a header now.
2629 * Removed the aged BC-32 config and all its supporting scripts
2633 * Removed support for Ultrix, Netware, and OS/2.
2637 * Add support for HKDF.
2639 *Alessandro Ghedini*
2641 * Add support for blake2b and blake2s
2645 * Added support for "pipelining". Ciphers that have the
2646 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2647 encryptions/decryptions simultaneously. There are currently no built-in
2648 ciphers with this property but the expectation is that engines will be able
2649 to offer it to significantly improve throughput. Support has been extended
2650 into libssl so that multiple records for a single connection can be
2651 processed in one go (for >=TLS 1.1).
2655 * Added the AFALG engine. This is an async capable engine which is able to
2656 offload work to the Linux kernel. In this initial version it only supports
2657 AES128-CBC. The kernel must be version 4.1.0 or greater.
2661 * OpenSSL now uses a new threading API. It is no longer necessary to
2662 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2663 are two supported threading models: pthreads and windows threads. It is
2664 also possible to configure OpenSSL at compile time for "no-threads". The
2665 old threading API should no longer be used. The functions have been
2666 replaced with "no-op" compatibility macros.
2668 *Alessandro Ghedini, Matt Caswell*
2670 * Modify behavior of ALPN to invoke callback after SNI/servername
2671 callback, such that updates to the SSL_CTX affect ALPN.
2675 * Add SSL_CIPHER queries for authentication and key-exchange.
2679 * Changes to the DEFAULT cipherlist:
2680 - Prefer (EC)DHE handshakes over plain RSA.
2681 - Prefer AEAD ciphers over legacy ciphers.
2682 - Prefer ECDSA over RSA when both certificates are available.
2683 - Prefer TLSv1.2 ciphers/PRF.
2684 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2689 * Change the ECC default curve list to be this, in order: x25519,
2690 secp256r1, secp521r1, secp384r1.
2694 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2695 disabled by default. They can be re-enabled using the
2696 enable-weak-ssl-ciphers option to Configure.
2700 * If the server has ALPN configured, but supports no protocols that the
2701 client advertises, send a fatal "no_application_protocol" alert.
2702 This behaviour is SHALL in RFC 7301, though it isn't universally
2703 implemented by other servers.
2707 * Add X25519 support.
2708 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2709 for public and private key encoding using the format documented in
2710 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2711 key generation and key derivation.
2713 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2718 * Deprecate SRP_VBASE_get_by_user.
2719 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2720 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2721 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2722 seed, even if the seed is configured.
2724 Users should use SRP_VBASE_get1_by_user instead. Note that in
2725 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2726 also that even though configuring the SRP seed attempts to hide
2727 invalid usernames by continuing the handshake with fake
2728 credentials, this behaviour is not constant time and no strong
2729 guarantees are made that the handshake is indistinguishable from
2730 that of a valid user.
2734 * Configuration change; it's now possible to build dynamic engines
2735 without having to build shared libraries and vice versa. This
2736 only applies to the engines in engines/, those in crypto/engine/
2737 will always be built into libcrypto (i.e. "static").
2739 Building dynamic engines is enabled by default; to disable, use
2740 the configuration option "disable-dynamic-engine".
2742 The only requirements for building dynamic engines are the
2743 presence of the DSO module and building with position independent
2744 code, so they will also automatically be disabled if configuring
2745 with "disable-dso" or "disable-pic".
2747 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2748 are also taken away from openssl/opensslconf.h, as they are
2753 * Configuration change; if there is a known flag to compile
2754 position independent code, it will always be applied on the
2755 libcrypto and libssl object files, and never on the application
2756 object files. This means other libraries that use routines from
2757 libcrypto / libssl can be made into shared libraries regardless
2758 of how OpenSSL was configured.
2760 If this isn't desirable, the configuration options "disable-pic"
2761 or "no-pic" can be used to disable the use of PIC. This will
2762 also disable building shared libraries and dynamic engines.
2766 * Removed JPAKE code. It was experimental and has no wide use.
2770 * The INSTALL_PREFIX Makefile variable has been renamed to
2771 DESTDIR. That makes for less confusion on what this variable
2772 is for. Also, the configuration option --install_prefix is
2777 * Heartbeat for TLS has been removed and is disabled by default
2778 for DTLS; configure with enable-heartbeats. Code that uses the
2779 old #define's might need to be updated.
2781 *Emilia Käsper, Rich Salz*
2783 * Rename REF_CHECK to REF_DEBUG.
2787 * New "unified" build system
2789 The "unified" build system is aimed to be a common system for all
2790 platforms we support. With it comes new support for VMS.
2792 This system builds supports building in a different directory tree
2793 than the source tree. It produces one Makefile (for unix family
2794 or lookalikes), or one descrip.mms (for VMS).
2796 The source of information to make the Makefile / descrip.mms is
2797 small files called 'build.info', holding the necessary
2798 information for each directory with source to compile, and a
2799 template in Configurations, like unix-Makefile.tmpl or
2802 With this change, the library names were also renamed on Windows
2803 and on VMS. They now have names that are closer to the standard
2804 on Unix, and include the major version number, and in certain
2805 cases, the architecture they are built for. See "Notes on shared
2806 libraries" in INSTALL.
2808 We rely heavily on the perl module Text::Template.
2812 * Added support for auto-initialisation and de-initialisation of the library.
2813 OpenSSL no longer requires explicit init or deinit routines to be called,
2814 except in certain circumstances. See the OPENSSL_init_crypto() and
2815 OPENSSL_init_ssl() man pages for further information.
2819 * The arguments to the DTLSv1_listen function have changed. Specifically the
2820 "peer" argument is now expected to be a BIO_ADDR object.
2822 * Rewrite of BIO networking library. The BIO library lacked consistent
2823 support of IPv6, and adding it required some more extensive
2824 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2825 which hold all types of addresses and chains of address information.
2826 It also introduces a new API, with functions like BIO_socket,
2827 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2828 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2829 have been adapted accordingly.
2833 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2838 * CRIME protection: disable compression by default, even if OpenSSL is
2839 compiled with zlib enabled. Applications can still enable compression
2840 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2841 using the SSL_CONF library to configure compression.
2845 * The signature of the session callback configured with
2846 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2847 was explicitly marked as `const unsigned char*` instead of
2852 * Always DPURIFY. Remove the use of uninitialized memory in the
2853 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2857 * Removed many obsolete configuration items, including
2858 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2859 MD2_CHAR, MD2_INT, MD2_LONG
2861 IDEA_SHORT, IDEA_LONG
2862 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2864 *Rich Salz, with advice from Andy Polyakov*
2866 * Many BN internals have been moved to an internal header file.
2868 *Rich Salz with help from Andy Polyakov*
2870 * Configuration and writing out the results from it has changed.
2871 Files such as Makefile include/openssl/opensslconf.h and are now
2872 produced through general templates, such as Makefile.in and
2873 crypto/opensslconf.h.in and some help from the perl module
2876 Also, the center of configuration information is no longer
2877 Makefile. Instead, Configure produces a perl module in
2878 configdata.pm which holds most of the config data (in the hash
2879 table %config), the target data that comes from the target
2880 configuration in one of the `Configurations/*.conf~ files (in
2885 * To clarify their intended purposes, the Configure options
2886 --prefix and --openssldir change their semantics, and become more
2887 straightforward and less interdependent.
2889 --prefix shall be used exclusively to give the location INSTALLTOP
2890 where programs, scripts, libraries, include files and manuals are
2891 going to be installed. The default is now /usr/local.
2893 --openssldir shall be used exclusively to give the default
2894 location OPENSSLDIR where certificates, private keys, CRLs are
2895 managed. This is also where the default openssl.cnf gets
2897 If the directory given with this option is a relative path, the
2898 values of both the --prefix value and the --openssldir value will
2899 be combined to become OPENSSLDIR.
2900 The default for --openssldir is INSTALLTOP/ssl.
2902 Anyone who uses --openssldir to specify where OpenSSL is to be
2903 installed MUST change to use --prefix instead.
2907 * The GOST engine was out of date and therefore it has been removed. An up
2908 to date GOST engine is now being maintained in an external repository.
2909 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
2910 support for GOST ciphersuites (these are only activated if a GOST engine
2915 * EGD is no longer supported by default; use enable-egd when
2918 *Ben Kaduk and Rich Salz*
2920 * The distribution now has Makefile.in files, which are used to
2921 create Makefile's when Configure is run. *Configure must be run
2922 before trying to build now.*
2926 * The return value for SSL_CIPHER_description() for error conditions
2931 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2933 Obtaining and performing DNSSEC validation of TLSA records is
2934 the application's responsibility. The application provides
2935 the TLSA records of its choice to OpenSSL, and these are then
2936 used to authenticate the peer.
2938 The TLSA records need not even come from DNS. They can, for
2939 example, be used to implement local end-entity certificate or
2940 trust-anchor "pinning", where the "pin" data takes the form
2941 of TLSA records, which can augment or replace verification
2942 based on the usual WebPKI public certification authorities.
2946 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2947 continues to support deprecated interfaces in default builds.
2948 However, applications are strongly advised to compile their
2949 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2950 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2951 or the 1.1.0 releases.
2953 In environments in which all applications have been ported to
2954 not use any deprecated interfaces OpenSSL's Configure script
2955 should be used with the --api=1.1.0 option to entirely remove
2956 support for the deprecated features from the library and
2957 unconditionally disable them in the installed headers.
2958 Essentially the same effect can be achieved with the "no-deprecated"
2959 argument to Configure, except that this will always restrict
2960 the build to just the latest API, rather than a fixed API
2963 As applications are ported to future revisions of the API,
2964 they should update their compile-time OPENSSL_API_COMPAT define
2965 accordingly, but in most cases should be able to continue to
2966 compile with later releases.
2968 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2969 0x10000000L and 0x00908000L, respectively. However those
2970 versions did not support the OPENSSL_API_COMPAT feature, and
2971 so applications are not typically tested for explicit support
2972 of just the undeprecated features of either release.
2976 * Add support for setting the minimum and maximum supported protocol.
2977 It can bet set via the SSL_set_min_proto_version() and
2978 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2979 MaxProtocol. It's recommended to use the new APIs to disable
2980 protocols instead of disabling individual protocols using
2981 SSL_set_options() or SSL_CONF's Protocol. This change also
2982 removes support for disabling TLS 1.2 in the OpenSSL TLS
2983 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2987 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2991 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2992 and integrates ECDSA and ECDH functionality into EC. Implementations can
2993 now redirect key generation and no longer need to convert to or from
2996 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2997 include the ec.h header file instead.
3001 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3002 ciphers who are no longer supported and drops support the ephemeral RSA key
3003 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3007 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3008 opaque. For HMAC_CTX, the following constructors and destructors
3011 HMAC_CTX *HMAC_CTX_new(void);
3012 void HMAC_CTX_free(HMAC_CTX *ctx);
3014 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3015 destroy such methods has been added. See EVP_MD_meth_new(3) and
3016 EVP_CIPHER_meth_new(3) for documentation.
3019 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
3020 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
3021 EVP_MD_CTX_reset() should be called instead to reinitialise
3022 an already created structure.
3023 2) For consistency with the majority of our object creators and
3024 destructors, EVP_MD_CTX_(create|destroy) were renamed to
3025 EVP_MD_CTX_(new|free). The old names are retained as macros
3026 for deprecated builds.
3030 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3031 cryptographic operations to be performed asynchronously as long as an
3032 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3033 further details. Libssl has also had this capability integrated with the
3034 introduction of the new mode SSL_MODE_ASYNC and associated error
3035 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3036 pages. This work was developed in partnership with Intel Corp.
3040 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3041 always enabled now. If you want to disable the support you should
3042 exclude it using the list of supported ciphers. This also means that the
3043 "-no_ecdhe" option has been removed from s_server.
3047 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3048 SSL_{CTX_}set1_curves() which can set a list.
3052 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3053 curve you want to support using SSL_{CTX_}set1_curves().
3057 * State machine rewrite. The state machine code has been significantly
3058 refactored in order to remove much duplication of code and solve issues
3059 with the old code (see ssl/statem/README for further details). This change
3060 does have some associated API changes. Notably the SSL_state() function
3061 has been removed and replaced by SSL_get_state which now returns an
3062 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
3063 altogether. The previous handshake states defined in ssl.h and ssl3.h have
3068 * All instances of the string "ssleay" in the public API were replaced
3069 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3070 Some error codes related to internal RSA_eay API's were renamed.
3074 * The demo files in crypto/threads were moved to demo/threads.
3078 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3081 *Matt Caswell, Rich Salz*
3083 * New ASN.1 embed macro.
3085 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3086 structure is not allocated: it is part of the parent. That is instead of
3094 This reduces memory fragmentation and make it impossible to accidentally
3095 set a mandatory field to NULL.
3097 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3098 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3099 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3104 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3108 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3109 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3110 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3111 DES and RC4 ciphersuites.
3115 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3116 This changes the decoding behaviour for some invalid messages,
3117 though the change is mostly in the more lenient direction, and
3118 legacy behaviour is preserved as much as possible.
3122 * Fix no-stdio build.
3123 * David Woodhouse <David.Woodhouse@intel.com> and also
3124 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3126 * New testing framework
3127 The testing framework has been largely rewritten and is now using
3128 perl and the perl modules Test::Harness and an extended variant of
3129 Test::More called OpenSSL::Test to do its work. All test scripts in
3130 test/ have been rewritten into test recipes, and all direct calls to
3131 executables in test/Makefile have become individual recipes using the
3132 simplified testing OpenSSL::Test::Simple.
3134 For documentation on our testing modules, do:
3136 perldoc test/testlib/OpenSSL/Test/Simple.pm
3137 perldoc test/testlib/OpenSSL/Test.pm
3141 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3142 are used; the latter aborts on memory leaks (usually checked on exit).
3143 Some undocumented "set malloc, etc., hooks" functions were removed
3144 and others were changed. All are now documented.
3148 * In DSA_generate_parameters_ex, if the provided seed is too short,
3151 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3153 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3154 from RFC4279, RFC4785, RFC5487, RFC5489.
3156 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3157 original RSA_PSK patch.
3161 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3162 era flag was never set throughout the codebase (only read). Also removed
3163 SSL3_FLAGS_POP_BUFFER which was only used if
3164 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3168 * Changed the default name options in the "ca", "crl", "req" and "x509"
3169 to be "oneline" instead of "compat".
3173 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3174 not aware of clients that still exhibit this bug, and the workaround
3175 hasn't been working properly for a while.
3179 * The return type of BIO_number_read() and BIO_number_written() as well as
3180 the corresponding num_read and num_write members in the BIO structure has
3181 changed from unsigned long to uint64_t. On platforms where an unsigned
3182 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3187 * Given the pervasive nature of TLS extensions it is inadvisable to run
3188 OpenSSL without support for them. It also means that maintaining
3189 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3190 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3194 * Removed support for the two export grade static DH ciphersuites
3195 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3196 were newly added (along with a number of other static DH ciphersuites) to
3197 1.0.2. However the two export ones have *never* worked since they were
3198 introduced. It seems strange in any case to be adding new export
3199 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3203 * Version negotiation has been rewritten. In particular SSLv23_method(),
3204 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3205 and turned into macros which simply call the new preferred function names
3206 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3207 should use the new names instead. Also as part of this change the ssl23.h
3208 header file has been removed.
3212 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3213 code and the associated standard is no longer considered fit-for-purpose.
3217 * RT2547 was closed. When generating a private key, try to make the
3218 output file readable only by the owner. This behavior change might
3219 be noticeable when interacting with other software.
3221 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3226 * Added HTTP GET support to the ocsp command.
3230 * Changed default digest for the dgst and enc commands from MD5 to
3235 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3239 * Added support for TLS extended master secret from
3240 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3241 initial patch which was a great help during development.
3245 * All libssl internal structures have been removed from the public header
3246 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3247 now redundant). Users should not attempt to access internal structures
3248 directly. Instead they should use the provided API functions.
3252 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3253 Access to deprecated functions can be re-enabled by running config with
3254 "enable-deprecated". In addition applications wishing to use deprecated
3255 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3256 will, by default, disable some transitive includes that previously existed
3257 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3261 * Added support for OCB mode. OpenSSL has been granted a patent license
3262 compatible with the OpenSSL license for use of OCB. Details are available
3263 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3264 for OCB can be removed by calling config with no-ocb.
3268 * SSLv2 support has been removed. It still supports receiving a SSLv2
3269 compatible client hello.
3273 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3274 done while fixing the error code for the key-too-small case.
3276 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3278 * CA.sh has been removed; use CA.pl instead.
3282 * Removed old DES API.
3286 * Remove various unsupported platforms:
3292 Sinix/ReliantUNIX RM400
3297 16-bit platforms such as WIN16
3301 * Clean up OPENSSL_NO_xxx #define's
3302 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3303 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3304 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3305 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3306 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3307 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3308 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3309 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3310 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3311 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3315 * Cleaned up dead code
3316 Remove all but one '#ifdef undef' which is to be looked at.
3320 * Clean up calling of xxx_free routines.
3321 Just like free(), fix most of the xxx_free routines to accept
3322 NULL. Remove the non-null checks from callers. Save much code.
3326 * Add secure heap for storage of private keys (when possible).
3327 Add BIO_s_secmem(), CBIGNUM, etc.
3328 Contributed by Akamai Technologies under our Corporate CLA.
3332 * Experimental support for a new, fast, unbiased prime candidate generator,
3333 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3335 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3337 * New output format NSS in the sess_id command line tool. This allows
3338 exporting the session id and the master key in NSS keylog format.
3340 *Martin Kaiser <martin@kaiser.cx>*
3342 * Harmonize version and its documentation. -f flag is used to display
3345 *mancha <mancha1@zoho.com>*
3347 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3348 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3350 *mancha <mancha1@zoho.com>*
3352 * Fix some double frees. These are not thought to be exploitable.
3354 *mancha <mancha1@zoho.com>*
3356 * A missing bounds check in the handling of the TLS heartbeat extension
3357 can be used to reveal up to 64k of memory to a connected client or
3360 Thanks for Neel Mehta of Google Security for discovering this bug and to
3361 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3362 preparing the fix [CVE-2014-0160][]
3364 *Adam Langley, Bodo Moeller*
3366 * Fix for the attack described in the paper "Recovering OpenSSL
3367 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3368 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3369 <http://eprint.iacr.org/2014/140>
3371 Thanks to Yuval Yarom and Naomi Benger for discovering this
3372 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3374 *Yuval Yarom and Naomi Benger*
3376 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3377 this fixes a limitation in previous versions of OpenSSL.
3381 * Experimental encrypt-then-mac support.
3383 Experimental support for encrypt then mac from
3384 draft-gutmann-tls-encrypt-then-mac-02.txt
3386 To enable it set the appropriate extension number (0x42 for the test
3387 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3389 For non-compliant peers (i.e. just about everything) this should have no
3392 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3396 * Add EVP support for key wrapping algorithms, to avoid problems with
3397 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3398 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3399 algorithms and include tests cases.
3403 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3408 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3409 MGF1 digest and OAEP label.
3413 * Make openssl verify return errors.
3415 *Chris Palmer <palmer@google.com> and Ben Laurie*
3417 * New function ASN1_TIME_diff to calculate the difference between two
3418 ASN1_TIME structures or one structure and the current time.
3422 * Update fips_test_suite to support multiple command line options. New
3423 test to induce all self test errors in sequence and check expected
3428 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3429 sign or verify all in one operation.
3433 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3434 test programs and fips_test_suite. Includes functionality to parse
3435 the minimal script output of fipsalgest.pl directly.
3439 * Add authorisation parameter to FIPS_module_mode_set().
3443 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3447 * Use separate DRBG fields for internal and external flags. New function
3448 FIPS_drbg_health_check() to perform on demand health checking. Add
3449 generation tests to fips_test_suite with reduced health check interval to
3450 demonstrate periodic health checking. Add "nodh" option to
3451 fips_test_suite to skip very slow DH test.
3455 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3460 * More extensive health check for DRBG checking many more failure modes.
3461 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3462 combination: call this in fips_test_suite.
3466 * Add support for canonical generation of DSA parameter 'g'. See
3469 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3470 POST to handle HMAC cases.
3474 * Add functions FIPS_module_version() and FIPS_module_version_text()
3475 to return numerical and string versions of the FIPS module number.
3479 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3480 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3481 outside the validated module in the FIPS capable OpenSSL.
3485 * Minor change to DRBG entropy callback semantics. In some cases
3486 there is no multiple of the block length between min_len and
3487 max_len. Allow the callback to return more than max_len bytes
3488 of entropy but discard any extra: it is the callback's responsibility
3489 to ensure that the extra data discarded does not impact the
3490 requested amount of entropy.
3494 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3495 information in FIPS186-3, SP800-57 and SP800-131A.
3499 * CCM support via EVP. Interface is very similar to GCM case except we
3500 must supply all data in one chunk (i.e. no update, final) and the
3501 message length must be supplied if AAD is used. Add algorithm test
3506 * Initial version of POST overhaul. Add POST callback to allow the status
3507 of POST to be monitored and/or failures induced. Modify fips_test_suite
3508 to use callback. Always run all selftests even if one fails.
3512 * XTS support including algorithm test driver in the fips_gcmtest program.
3513 Note: this does increase the maximum key length from 32 to 64 bytes but
3514 there should be no binary compatibility issues as existing applications
3515 will never use XTS mode.
3519 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3520 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3521 performs algorithm blocking for unapproved PRNG types. Also do not
3522 set PRNG type in FIPS_mode_set(): leave this to the application.
3523 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3524 the standard OpenSSL PRNG: set additional data to a date time vector.
3528 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3529 This shouldn't present any incompatibility problems because applications
3530 shouldn't be using these directly and any that are will need to rethink
3531 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3535 * Extensive self tests and health checking required by SP800-90 DRBG.
3536 Remove strength parameter from FIPS_drbg_instantiate and always
3537 instantiate at maximum supported strength.
3541 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3545 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3549 * New function DH_compute_key_padded() to compute a DH key and pad with
3550 leading zeroes if needed: this complies with SP800-56A et al.
3554 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3555 anything, incomplete, subject to change and largely untested at present.
3559 * Modify fipscanisteronly build option to only build the necessary object
3560 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3564 * Add experimental option FIPSSYMS to give all symbols in
3565 fipscanister.o and FIPS or fips prefix. This will avoid
3566 conflicts with future versions of OpenSSL. Add perl script
3567 util/fipsas.pl to preprocess assembly language source files
3568 and rename any affected symbols.
3572 * Add selftest checks and algorithm block of non-fips algorithms in
3573 FIPS mode. Remove DES2 from selftests.
3577 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3578 return internal method without any ENGINE dependencies. Add new
3579 tiny fips sign and verify functions.
3583 * New build option no-ec2m to disable characteristic 2 code.
3587 * New build option "fipscanisteronly". This only builds fipscanister.o
3588 and (currently) associated fips utilities. Uses the file Makefile.fips
3589 instead of Makefile.org as the prototype.
3593 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3594 Update fips_gcmtest to use IV generator.
3598 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3599 setting output buffer to NULL. The `*Final` function must be
3600 called although it will not retrieve any additional data. The tag
3601 can be set or retrieved with a ctrl. The IV length is by default 12
3602 bytes (96 bits) but can be set to an alternative value. If the IV
3603 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3608 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3609 underlying do_cipher function handles all cipher semantics itself
3610 including padding and finalisation. This is useful if (for example)
3611 an ENGINE cipher handles block padding itself. The behaviour of
3612 do_cipher is subtly changed if this flag is set: the return value
3613 is the number of characters written to the output buffer (zero is
3614 no longer an error code) or a negative error code. Also if the
3615 input buffer is NULL and length 0 finalisation should be performed.
3619 * If a candidate issuer certificate is already part of the constructed
3620 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3624 * Improve forward-security support: add functions
3626 void SSL_CTX_set_not_resumable_session_callback(
3627 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3628 void SSL_set_not_resumable_session_callback(
3629 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3631 for use by SSL/TLS servers; the callback function will be called whenever a
3632 new session is created, and gets to decide whether the session may be
3633 cached to make it resumable (return 0) or not (return 1). (As by the
3634 SSL/TLS protocol specifications, the session_id sent by the server will be
3635 empty to indicate that the session is not resumable; also, the server will
3636 not generate RFC 4507 (RFC 5077) session tickets.)
3638 A simple reasonable callback implementation is to return is_forward_secure.
3639 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3640 by the SSL/TLS server library, indicating whether it can provide forward
3643 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3645 * New -verify_name option in command line utilities to set verification
3650 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3651 Add CMAC pkey methods.
3655 * Experimental renegotiation in s_server -www mode. If the client
3656 browses /reneg connection is renegotiated. If /renegcert it is
3657 renegotiated requesting a certificate.
3661 * Add an "external" session cache for debugging purposes to s_server. This
3662 should help trace issues which normally are only apparent in deployed
3663 multi-process servers.
3667 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3668 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3669 BIO_set_cipher() and some obscure PEM functions were changed so they
3670 can now return an error. The RAND changes required a change to the
3671 RAND_METHOD structure.
3675 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3676 a gcc attribute to warn if the result of a function is ignored. This
3677 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3678 whose return value is often ignored.
3682 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3683 These allow SCTs (signed certificate timestamps) to be requested and
3684 validated when establishing a connection.
3686 *Rob Percival <robpercival@google.com>*
3691 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3693 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3694 used even when parsing explicit parameters, when loading a serialized key
3695 or calling `EC_GROUP_new_from_ecpkparameters()`/
3696 `EC_GROUP_new_from_ecparameters()`.
3697 This prevents bypass of security hardening and performance gains,
3698 especially for curves with specialized EC_METHODs.
3699 By default, if a key encoded with explicit parameters is loaded and later
3700 serialized, the output is still encoded with explicit parameters, even if
3701 internally a "named" EC_GROUP is used for computation.
3705 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3706 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3707 NULL. After this change, only the cofactor parameter can be NULL. It also
3708 does some minimal sanity checks on the passed order.
3713 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3714 An attack is simple, if the first CMS_recipientInfo is valid but the
3715 second CMS_recipientInfo is chosen ciphertext. If the second
3716 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3717 encryption key will be replaced by garbage, and the message cannot be
3718 decoded, but if the RSA decryption fails, the correct encryption key is
3719 used and the recipient will not notice the attack.
3720 As a work around for this potential attack the length of the decrypted
3721 key must be equal to the cipher default key length, in case the
3722 certifiate is not given and all recipientInfo are tried out.
3723 The old behaviour can be re-enabled in the CMS code by setting the
3724 CMS_DEBUG_DECRYPT flag.
3729 * Document issue with installation paths in diverse Windows builds
3731 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3732 binaries and run-time config file.
3737 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3739 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3740 This changes the size when using the genpkey app when no size is given. It
3741 fixes an omission in earlier changes that changed all RSA, DSA and DH
3742 generation apps to use 2048 bits by default.
3746 * Add FIPS support for Android Arm 64-bit
3748 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3749 Module in Version 2.0.10. For some reason, the corresponding target
3750 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3751 built with FIPS support on Android Arm 64-bit. This omission has been
3754 *Matthias St. Pierre*
3756 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3758 * 0-byte record padding oracle
3760 If an application encounters a fatal protocol error and then calls
3761 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3762 then OpenSSL can respond differently to the calling application if a 0 byte
3763 record is received with invalid padding compared to if a 0 byte record is
3764 received with an invalid MAC. If the application then behaves differently
3765 based on that in a way that is detectable to the remote peer, then this
3766 amounts to a padding oracle that could be used to decrypt data.
3768 In order for this to be exploitable "non-stitched" ciphersuites must be in
3769 use. Stitched ciphersuites are optimised implementations of certain
3770 commonly used ciphersuites. Also the application must call SSL_shutdown()
3771 twice even if a protocol error has occurred (applications should not do
3772 this but some do anyway).
3774 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3775 Aviram, with additional investigation by Steven Collison and Andrew
3776 Hourselt. It was reported to OpenSSL on 10th December 2018.
3781 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3785 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3787 * Microarchitecture timing vulnerability in ECC scalar multiplication
3789 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3790 shown to be vulnerable to a microarchitecture timing side channel attack.
3791 An attacker with sufficient access to mount local timing attacks during
3792 ECDSA signature generation could recover the private key.
3794 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3795 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3801 * Timing vulnerability in DSA signature generation
3803 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3804 timing side channel attack. An attacker could use variations in the signing
3805 algorithm to recover the private key.
3807 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3812 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3813 Module, accidentally introduced while backporting security fixes from the
3814 development branch and hindering the use of ECC in FIPS mode.
3818 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3820 * Client DoS due to large DH parameter
3822 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3823 malicious server can send a very large prime value to the client. This will
3824 cause the client to spend an unreasonably long period of time generating a
3825 key for this prime resulting in a hang until the client has finished. This
3826 could be exploited in a Denial Of Service attack.
3828 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3833 * Cache timing vulnerability in RSA Key Generation
3835 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3836 a cache timing side channel attack. An attacker with sufficient access to
3837 mount cache timing attacks during the RSA key generation process could
3838 recover the private key.
3840 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3841 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3846 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3847 parameter is no longer accepted, as it leads to a corrupt table. NULL
3848 pem_str is reserved for alias entries only.
3852 * Revert blinding in ECDSA sign and instead make problematic addition
3853 length-invariant. Switch even to fixed-length Montgomery multiplication.
3857 * Change generating and checking of primes so that the error rate of not
3858 being prime depends on the intended use based on the size of the input.
3859 For larger primes this will result in more rounds of Miller-Rabin.
3860 The maximal error rate for primes with more than 1080 bits is lowered
3863 *Kurt Roeckx, Annie Yousar*
3865 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3869 * Add blinding to ECDSA and DSA signatures to protect against side channel
3870 attacks discovered by Keegan Ryan (NCC Group).
3874 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3875 now allow empty (zero character) pass phrases.
3879 * Certificate time validation (X509_cmp_time) enforces stricter
3880 compliance with RFC 5280. Fractional seconds and timezone offsets
3881 are no longer allowed.
3885 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3887 * Constructed ASN.1 types with a recursive definition could exceed the stack
3889 Constructed ASN.1 types with a recursive definition (such as can be found
3890 in PKCS7) could eventually exceed the stack given malicious input with
3891 excessive recursion. This could result in a Denial Of Service attack. There
3892 are no such structures used within SSL/TLS that come from untrusted sources
3893 so this is considered safe.
3895 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3901 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
3903 * Read/write after SSL object in error state
3905 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3906 mechanism. The intent was that if a fatal error occurred during a handshake
3907 then OpenSSL would move into the error state and would immediately fail if
3908 you attempted to continue the handshake. This works as designed for the
3909 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3910 SSL_connect()), however due to a bug it does not work correctly if
3911 SSL_read() or SSL_write() is called directly. In that scenario, if the
3912 handshake fails then a fatal error will be returned in the initial function
3913 call. If SSL_read()/SSL_write() is subsequently called by the application
3914 for the same SSL object then it will succeed and the data is passed without
3915 being decrypted/encrypted directly from the SSL/TLS record layer.
3917 In order to exploit this issue an application bug would have to be present
3918 that resulted in a call to SSL_read()/SSL_write() being issued after having
3919 already received a fatal error.
3921 This issue was reported to OpenSSL by David Benjamin (Google).
3926 * rsaz_1024_mul_avx2 overflow bug on x86_64
3928 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3929 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3930 Analysis suggests that attacks against RSA and DSA as a result of this
3931 defect would be very difficult to perform and are not believed likely.
3932 Attacks against DH1024 are considered just feasible, because most of the
3933 work necessary to deduce information about a private key may be performed
3934 offline. The amount of resources required for such an attack would be
3935 significant. However, for an attack on TLS to be meaningful, the server
3936 would have to share the DH1024 private key among multiple clients, which is
3937 no longer an option since CVE-2016-0701.
3939 This only affects processors that support the AVX2 but not ADX extensions
3940 like Intel Haswell (4th generation).
3942 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3943 was originally found via the OSS-Fuzz project.
3948 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
3950 * bn_sqrx8x_internal carry bug on x86_64
3952 There is a carry propagating bug in the x86_64 Montgomery squaring
3953 procedure. No EC algorithms are affected. Analysis suggests that attacks
3954 against RSA and DSA as a result of this defect would be very difficult to
3955 perform and are not believed likely. Attacks against DH are considered just
3956 feasible (although very difficult) because most of the work necessary to
3957 deduce information about a private key may be performed offline. The amount
3958 of resources required for such an attack would be very significant and
3959 likely only accessible to a limited number of attackers. An attacker would
3960 additionally need online access to an unpatched system using the target
3961 private key in a scenario with persistent DH parameters and a private
3962 key that is shared between multiple clients.
3964 This only affects processors that support the BMI1, BMI2 and ADX extensions
3965 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3967 This issue was reported to OpenSSL by the OSS-Fuzz project.
3972 * Malformed X.509 IPAddressFamily could cause OOB read
3974 If an X.509 certificate has a malformed IPAddressFamily extension,
3975 OpenSSL could do a one-byte buffer overread. The most likely result
3976 would be an erroneous display of the certificate in text format.
3978 This issue was reported to OpenSSL by the OSS-Fuzz project.
3983 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
3985 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3986 platform rather than 'mingw'.
3990 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
3992 * Truncated packet could crash via OOB read
3994 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3995 cipher is being used, then a truncated packet can cause that host to
3996 perform an out-of-bounds read, usually resulting in a crash.
3998 This issue was reported to OpenSSL by Robert Święcki of Google.
4003 * BN_mod_exp may produce incorrect results on x86_64
4005 There is a carry propagating bug in the x86_64 Montgomery squaring
4006 procedure. No EC algorithms are affected. Analysis suggests that attacks
4007 against RSA and DSA as a result of this defect would be very difficult to
4008 perform and are not believed likely. Attacks against DH are considered just
4009 feasible (although very difficult) because most of the work necessary to
4010 deduce information about a private key may be performed offline. The amount
4011 of resources required for such an attack would be very significant and
4012 likely only accessible to a limited number of attackers. An attacker would
4013 additionally need online access to an unpatched system using the target
4014 private key in a scenario with persistent DH parameters and a private
4015 key that is shared between multiple clients. For example this can occur by
4016 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4017 similar to CVE-2015-3193 but must be treated as a separate problem.
4019 This issue was reported to OpenSSL by the OSS-Fuzz project.
4024 * Montgomery multiplication may produce incorrect results
4026 There is a carry propagating bug in the Broadwell-specific Montgomery
4027 multiplication procedure that handles input lengths divisible by, but
4028 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4029 and DH private keys are impossible. This is because the subroutine in
4030 question is not used in operations with the private key itself and an input
4031 of the attacker's direct choice. Otherwise the bug can manifest itself as
4032 transient authentication and key negotiation failures or reproducible
4033 erroneous outcome of public-key operations with specially crafted input.
4034 Among EC algorithms only Brainpool P-512 curves are affected and one
4035 presumably can attack ECDH key negotiation. Impact was not analyzed in
4036 detail, because pre-requisites for attack are considered unlikely. Namely
4037 multiple clients have to choose the curve in question and the server has to
4038 share the private key among them, neither of which is default behaviour.
4039 Even then only clients that chose the curve will be affected.
4041 This issue was publicly reported as transient failures and was not
4042 initially recognized as a security issue. Thanks to Richard Morgan for
4043 providing reproducible case.
4048 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4049 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4050 prevent issues where no progress is being made and the peer continually
4051 sends unrecognised record types, using up resources processing them.
4055 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4057 * Missing CRL sanity check
4059 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4060 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4061 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4063 This issue only affects the OpenSSL 1.0.2i
4068 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4070 * OCSP Status Request extension unbounded memory growth
4072 A malicious client can send an excessively large OCSP Status Request
4073 extension. If that client continually requests renegotiation, sending a
4074 large OCSP Status Request extension each time, then there will be unbounded
4075 memory growth on the server. This will eventually lead to a Denial Of
4076 Service attack through memory exhaustion. Servers with a default
4077 configuration are vulnerable even if they do not support OCSP. Builds using
4078 the "no-ocsp" build time option are not affected.
4080 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4085 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4088 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4094 * OOB write in MDC2_Update()
4096 An overflow can occur in MDC2_Update() either if called directly or
4097 through the EVP_DigestUpdate() function using MDC2. If an attacker
4098 is able to supply very large amounts of input data after a previous
4099 call to EVP_EncryptUpdate() with a partial block then a length check
4100 can overflow resulting in a heap corruption.
4102 The amount of data needed is comparable to SIZE_MAX which is impractical
4105 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4110 * Malformed SHA512 ticket DoS
4112 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4113 DoS attack where a malformed ticket will result in an OOB read which will
4116 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4117 a custom server callback and ticket lookup mechanism.
4119 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4124 * OOB write in BN_bn2dec()
4126 The function BN_bn2dec() does not check the return value of BN_div_word().
4127 This can cause an OOB write if an application uses this function with an
4128 overly large BIGNUM. This could be a problem if an overly large certificate
4129 or CRL is printed out from an untrusted source. TLS is not affected because
4130 record limits will reject an oversized certificate before it is parsed.
4132 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4137 * OOB read in TS_OBJ_print_bio()
4139 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4140 the total length the OID text representation would use and not the amount
4141 of data written. This will result in OOB reads when large OIDs are
4144 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4149 * Pointer arithmetic undefined behaviour
4151 Avoid some undefined pointer arithmetic
4153 A common idiom in the codebase is to check limits in the following manner:
4156 Where "p" points to some malloc'd data of SIZE bytes and
4159 "len" here could be from some externally supplied data (e.g. from a TLS
4162 The rules of C pointer arithmetic are such that "p + len" is only well
4163 defined where len <= SIZE. Therefore the above idiom is actually
4164 undefined behaviour.
4166 For example this could cause problems if some malloc implementation
4167 provides an address for "p" such that "p + len" actually overflows for
4168 values of len that are too big and therefore p + len < limit.
4170 This issue was reported to OpenSSL by Guido Vranken
4175 * Constant time flag not preserved in DSA signing
4177 Operations in the DSA signing algorithm should run in constant time in
4178 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4179 implementation means that a non-constant time codepath is followed for
4180 certain operations. This has been demonstrated through a cache-timing
4181 attack to be sufficient for an attacker to recover the private DSA key.
4183 This issue was reported by César Pereida (Aalto University), Billy Brumley
4184 (Tampere University of Technology), and Yuval Yarom (The University of
4185 Adelaide and NICTA).
4190 * DTLS buffered message DoS
4192 In a DTLS connection where handshake messages are delivered out-of-order
4193 those messages that OpenSSL is not yet ready to process will be buffered
4194 for later use. Under certain circumstances, a flaw in the logic means that
4195 those messages do not get removed from the buffer even though the handshake
4196 has been completed. An attacker could force up to approx. 15 messages to
4197 remain in the buffer when they are no longer required. These messages will
4198 be cleared when the DTLS connection is closed. The default maximum size for
4199 a message is 100k. Therefore the attacker could force an additional 1500k
4200 to be consumed per connection. By opening many simulataneous connections an
4201 attacker could cause a DoS attack through memory exhaustion.
4203 This issue was reported to OpenSSL by Quan Luo.
4208 * DTLS replay protection DoS
4210 A flaw in the DTLS replay attack protection mechanism means that records
4211 that arrive for future epochs update the replay protection "window" before
4212 the MAC for the record has been validated. This could be exploited by an
4213 attacker by sending a record for the next epoch (which does not have to
4214 decrypt or have a valid MAC), with a very large sequence number. This means
4215 that all subsequent legitimate packets are dropped causing a denial of
4216 service for a specific DTLS connection.
4218 This issue was reported to OpenSSL by the OCAP audit team.
4223 * Certificate message OOB reads
4225 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4226 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4227 theoretical DoS risk but this has not been observed in practice on common
4230 The messages affected are client certificate, client certificate request
4231 and server certificate. As a result the attack can only be performed
4232 against a client or a server which enables client authentication.
4234 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4239 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4241 * Prevent padding oracle in AES-NI CBC MAC check
4243 A MITM attacker can use a padding oracle attack to decrypt traffic
4244 when the connection uses an AES CBC cipher and the server support
4247 This issue was introduced as part of the fix for Lucky 13 padding
4248 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4249 constant time by making sure that always the same bytes are read and
4250 compared against either the MAC or padding bytes. But it no longer
4251 checked that there was enough data to have both the MAC and padding
4254 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4259 * Fix EVP_EncodeUpdate overflow
4261 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4262 Base64 encoding of binary data. If an attacker is able to supply very large
4263 amounts of input data then a length check can overflow resulting in a heap
4266 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4267 the `PEM_write_bio*` family of functions. These are mainly used within the
4268 OpenSSL command line applications, so any application which processes data
4269 from an untrusted source and outputs it as a PEM file should be considered
4270 vulnerable to this issue. User applications that call these APIs directly
4271 with large amounts of untrusted data may also be vulnerable.
4273 This issue was reported by Guido Vranken.
4278 * Fix EVP_EncryptUpdate overflow
4280 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4281 is able to supply very large amounts of input data after a previous call to
4282 EVP_EncryptUpdate() with a partial block then a length check can overflow
4283 resulting in a heap corruption. Following an analysis of all OpenSSL
4284 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4285 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4286 the first called function after an EVP_EncryptInit(), and therefore that
4287 specific call must be safe. The second form is where the length passed to
4288 EVP_EncryptUpdate() can be seen from the code to be some small value and
4289 therefore there is no possibility of an overflow. Since all instances are
4290 one of these two forms, it is believed that there can be no overflows in
4291 internal code due to this problem. It should be noted that
4292 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4293 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4294 of these calls have also been analysed too and it is believed there are no
4295 instances in internal usage where an overflow could occur.
4297 This issue was reported by Guido Vranken.
4302 * Prevent ASN.1 BIO excessive memory allocation
4304 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4305 a short invalid encoding can cause allocation of large amounts of memory
4306 potentially consuming excessive resources or exhausting memory.
4308 Any application parsing untrusted data through d2i BIO functions is
4309 affected. The memory based functions such as d2i_X509() are *not* affected.
4310 Since the memory based functions are used by the TLS library, TLS
4311 applications are not affected.
4313 This issue was reported by Brian Carpenter.
4320 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4321 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4322 in arbitrary stack data being returned in the buffer.
4324 This issue was reported by Guido Vranken.
4329 * Modify behavior of ALPN to invoke callback after SNI/servername
4330 callback, such that updates to the SSL_CTX affect ALPN.
4334 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4339 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4340 methods are enabled and ssl2 is disabled the methods return NULL.
4344 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4346 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4347 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4348 provide any "EXPORT" or "LOW" strength ciphers.
4352 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4353 is by default disabled at build-time. Builds that are not configured with
4354 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4355 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4356 will need to explicitly call either of:
4358 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4360 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4362 as appropriate. Even if either of those is used, or the application
4363 explicitly uses the version-specific SSLv2_method() or its client and
4364 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4365 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4366 ciphers, and SSLv2 56-bit DES are no longer available.
4371 * Fix a double-free in DSA code
4373 A double free bug was discovered when OpenSSL parses malformed DSA private
4374 keys and could lead to a DoS attack or memory corruption for applications
4375 that receive DSA private keys from untrusted sources. This scenario is
4378 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4384 * Disable SRP fake user seed to address a server memory leak.
4386 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4388 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4389 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4390 was changed to ignore the "fake user" SRP seed, even if the seed
4393 Users should use SRP_VBASE_get1_by_user instead. Note that in
4394 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4395 also that even though configuring the SRP seed attempts to hide
4396 invalid usernames by continuing the handshake with fake
4397 credentials, this behaviour is not constant time and no strong
4398 guarantees are made that the handshake is indistinguishable from
4399 that of a valid user.
4404 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4406 In the BN_hex2bn function the number of hex digits is calculated using an
4407 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4408 large values of |i| this can result in |bn_expand| not allocating any
4409 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4410 field as NULL leading to a subsequent NULL ptr deref. For very large values
4411 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4412 In this case memory is allocated to the internal BIGNUM data field, but it
4413 is insufficiently sized leading to heap corruption. A similar issue exists
4414 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4415 is ever called by user applications with very large untrusted hex/dec data.
4416 This is anticipated to be a rare occurrence.
4418 All OpenSSL internal usage of these functions use data that is not expected
4419 to be untrusted, e.g. config file data or application command line
4420 arguments. If user developed applications generate config file data based
4421 on untrusted data then it is possible that this could also lead to security
4422 consequences. This is also anticipated to be rare.
4424 This issue was reported to OpenSSL by Guido Vranken.
4429 * Fix memory issues in `BIO_*printf` functions
4431 The internal |fmtstr| function used in processing a "%s" format string in
4432 the `BIO_*printf` functions could overflow while calculating the length of a
4433 string and cause an OOB read when printing very long strings.
4435 Additionally the internal |doapr_outch| function can attempt to write to an
4436 OOB memory location (at an offset from the NULL pointer) in the event of a
4437 memory allocation failure. In 1.0.2 and below this could be caused where
4438 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4439 could be in processing a very long "%s" format string. Memory leaks can
4442 The first issue may mask the second issue dependent on compiler behaviour.
4443 These problems could enable attacks where large amounts of untrusted data
4444 is passed to the `BIO_*printf` functions. If applications use these functions
4445 in this way then they could be vulnerable. OpenSSL itself uses these
4446 functions when printing out human-readable dumps of ASN.1 data. Therefore
4447 applications that print this data could be vulnerable if the data is from
4448 untrusted sources. OpenSSL command line applications could also be
4449 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4450 as command line arguments.
4452 Libssl is not considered directly vulnerable. Additionally certificates etc
4453 received via remote connections via libssl are also unlikely to be able to
4454 trigger these issues because of message size limits enforced within libssl.
4456 This issue was reported to OpenSSL Guido Vranken.
4461 * Side channel attack on modular exponentiation
4463 A side-channel attack was found which makes use of cache-bank conflicts on
4464 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4465 of RSA keys. The ability to exploit this issue is limited as it relies on
4466 an attacker who has control of code in a thread running on the same
4467 hyper-threaded core as the victim thread which is performing decryptions.
4469 This issue was reported to OpenSSL by Yuval Yarom, The University of
4470 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4471 Nadia Heninger, University of Pennsylvania with more information at
4472 <http://cachebleed.info>.
4477 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4478 if no keysize is specified with default_bits. This fixes an
4479 omission in an earlier change that changed all RSA/DSA key generation
4480 apps to use 2048 bits by default.
4484 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4486 * DH small subgroups
4488 Historically OpenSSL only ever generated DH parameters based on "safe"
4489 primes. More recently (in version 1.0.2) support was provided for
4490 generating X9.42 style parameter files such as those required for RFC 5114
4491 support. The primes used in such files may not be "safe". Where an
4492 application is using DH configured with parameters based on primes that are
4493 not "safe" then an attacker could use this fact to find a peer's private
4494 DH exponent. This attack requires that the attacker complete multiple
4495 handshakes in which the peer uses the same private DH exponent. For example
4496 this could be used to discover a TLS server's private DH exponent if it's
4497 reusing the private DH exponent or it's using a static DH ciphersuite.
4499 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4500 TLS. It is not on by default. If the option is not set then the server
4501 reuses the same private DH exponent for the life of the server process and
4502 would be vulnerable to this attack. It is believed that many popular
4503 applications do set this option and would therefore not be at risk.
4505 The fix for this issue adds an additional check where a "q" parameter is
4506 available (as is the case in X9.42 based parameters). This detects the
4507 only known attack, and is the only possible defense for static DH
4508 ciphersuites. This could have some performance impact.
4510 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4511 default and cannot be disabled. This could have some performance impact.
4513 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4518 * SSLv2 doesn't block disabled ciphers
4520 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4521 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4522 been disabled, provided that the SSLv2 protocol was not also disabled via
4525 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4526 and Sebastian Schinzel.
4531 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4533 * BN_mod_exp may produce incorrect results on x86_64
4535 There is a carry propagating bug in the x86_64 Montgomery squaring
4536 procedure. No EC algorithms are affected. Analysis suggests that attacks
4537 against RSA and DSA as a result of this defect would be very difficult to
4538 perform and are not believed likely. Attacks against DH are considered just
4539 feasible (although very difficult) because most of the work necessary to
4540 deduce information about a private key may be performed offline. The amount
4541 of resources required for such an attack would be very significant and
4542 likely only accessible to a limited number of attackers. An attacker would
4543 additionally need online access to an unpatched system using the target
4544 private key in a scenario with persistent DH parameters and a private
4545 key that is shared between multiple clients. For example this can occur by
4546 default in OpenSSL DHE based SSL/TLS ciphersuites.
4548 This issue was reported to OpenSSL by Hanno Böck.
4553 * Certificate verify crash with missing PSS parameter
4555 The signature verification routines will crash with a NULL pointer
4556 dereference if presented with an ASN.1 signature using the RSA PSS
4557 algorithm and absent mask generation function parameter. Since these
4558 routines are used to verify certificate signature algorithms this can be
4559 used to crash any certificate verification operation and exploited in a
4560 DoS attack. Any application which performs certificate verification is
4561 vulnerable including OpenSSL clients and servers which enable client
4564 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4569 * X509_ATTRIBUTE memory leak
4571 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4572 memory. This structure is used by the PKCS#7 and CMS routines so any
4573 application which reads PKCS#7 or CMS data from untrusted sources is
4574 affected. SSL/TLS is not affected.
4576 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4582 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4583 This changes the decoding behaviour for some invalid messages,
4584 though the change is mostly in the more lenient direction, and
4585 legacy behaviour is preserved as much as possible.
4589 * In DSA_generate_parameters_ex, if the provided seed is too short,
4592 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4594 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4596 * Alternate chains certificate forgery
4598 During certificate verification, OpenSSL will attempt to find an
4599 alternative certificate chain if the first attempt to build such a chain
4600 fails. An error in the implementation of this logic can mean that an
4601 attacker could cause certain checks on untrusted certificates to be
4602 bypassed, such as the CA flag, enabling them to use a valid leaf
4603 certificate to act as a CA and "issue" an invalid certificate.
4605 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4610 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4612 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4613 incompatibility in the handling of HMAC. The previous ABI has now been
4618 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4620 * Malformed ECParameters causes infinite loop
4622 When processing an ECParameters structure OpenSSL enters an infinite loop
4623 if the curve specified is over a specially malformed binary polynomial
4626 This can be used to perform denial of service against any
4627 system which processes public keys, certificate requests or
4628 certificates. This includes TLS clients and TLS servers with
4629 client authentication enabled.
4631 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4636 * Exploitable out-of-bounds read in X509_cmp_time
4638 X509_cmp_time does not properly check the length of the ASN1_TIME
4639 string and can read a few bytes out of bounds. In addition,
4640 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4643 An attacker can use this to craft malformed certificates and CRLs of
4644 various sizes and potentially cause a segmentation fault, resulting in
4645 a DoS on applications that verify certificates or CRLs. TLS clients
4646 that verify CRLs are affected. TLS clients and servers with client
4647 authentication enabled may be affected if they use custom verification
4650 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4651 independently by Hanno Böck.
4656 * PKCS7 crash with missing EnvelopedContent
4658 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4659 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4660 with missing content and trigger a NULL pointer dereference on parsing.
4662 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4663 structures from untrusted sources are affected. OpenSSL clients and
4664 servers are not affected.
4666 This issue was reported to OpenSSL by Michal Zalewski (Google).
4671 * CMS verify infinite loop with unknown hash function
4673 When verifying a signedData message the CMS code can enter an infinite loop
4674 if presented with an unknown hash function OID. This can be used to perform
4675 denial of service against any system which verifies signedData messages using
4677 This issue was reported to OpenSSL by Johannes Bauer.
4682 * Race condition handling NewSessionTicket
4684 If a NewSessionTicket is received by a multi-threaded client when attempting to
4685 reuse a previous ticket then a race condition can occur potentially leading to
4686 a double free of the ticket data.
4691 * Only support 256-bit or stronger elliptic curves with the
4692 'ecdh_auto' setting (server) or by default (client). Of supported
4693 curves, prefer P-256 (both).
4697 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4699 * ClientHello sigalgs DoS fix
4701 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4702 invalid signature algorithms extension a NULL pointer dereference will
4703 occur. This can be exploited in a DoS attack against the server.
4705 This issue was was reported to OpenSSL by David Ramos of Stanford
4709 *Stephen Henson and Matt Caswell*
4711 * Multiblock corrupted pointer fix
4713 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4714 feature only applies on 64 bit x86 architecture platforms that support AES
4715 NI instructions. A defect in the implementation of "multiblock" can cause
4716 OpenSSL's internal write buffer to become incorrectly set to NULL when
4717 using non-blocking IO. Typically, when the user application is using a
4718 socket BIO for writing, this will only result in a failed connection.
4719 However if some other BIO is used then it is likely that a segmentation
4720 fault will be triggered, thus enabling a potential DoS attack.
4722 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4727 * Segmentation fault in DTLSv1_listen fix
4729 The DTLSv1_listen function is intended to be stateless and processes the
4730 initial ClientHello from many peers. It is common for user code to loop
4731 over the call to DTLSv1_listen until a valid ClientHello is received with
4732 an associated cookie. A defect in the implementation of DTLSv1_listen means
4733 that state is preserved in the SSL object from one invocation to the next
4734 that can lead to a segmentation fault. Errors processing the initial
4735 ClientHello can trigger this scenario. An example of such an error could be
4736 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4739 This issue was reported to OpenSSL by Per Allansson.
4744 * Segmentation fault in ASN1_TYPE_cmp fix
4746 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4747 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4748 certificate signature algorithm consistency this can be used to crash any
4749 certificate verification operation and exploited in a DoS attack. Any
4750 application which performs certificate verification is vulnerable including
4751 OpenSSL clients and servers which enable client authentication.
4756 * Segmentation fault for invalid PSS parameters fix
4758 The signature verification routines will crash with a NULL pointer
4759 dereference if presented with an ASN.1 signature using the RSA PSS
4760 algorithm and invalid parameters. Since these routines are used to verify
4761 certificate signature algorithms this can be used to crash any
4762 certificate verification operation and exploited in a DoS attack. Any
4763 application which performs certificate verification is vulnerable including
4764 OpenSSL clients and servers which enable client authentication.
4766 This issue was was reported to OpenSSL by Brian Carpenter.
4771 * ASN.1 structure reuse memory corruption fix
4773 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4774 memory corruption via an invalid write. Such reuse is and has been
4775 strongly discouraged and is believed to be rare.
4777 Applications that parse structures containing CHOICE or ANY DEFINED BY
4778 components may be affected. Certificate parsing (d2i_X509 and related
4779 functions) are however not affected. OpenSSL clients and servers are
4785 * PKCS7 NULL pointer dereferences fix
4787 The PKCS#7 parsing code does not handle missing outer ContentInfo
4788 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4789 missing content and trigger a NULL pointer dereference on parsing.
4791 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4792 otherwise parse PKCS#7 structures from untrusted sources are
4793 affected. OpenSSL clients and servers are not affected.
4795 This issue was reported to OpenSSL by Michal Zalewski (Google).
4800 * DoS via reachable assert in SSLv2 servers fix
4802 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4803 servers that both support SSLv2 and enable export cipher suites by sending
4804 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4806 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4807 (OpenSSL development team).
4812 * Empty CKE with client auth and DHE fix
4814 If client auth is used then a server can seg fault in the event of a DHE
4815 ciphersuite being selected and a zero length ClientKeyExchange message
4816 being sent by the client. This could be exploited in a DoS attack.
4821 * Handshake with unseeded PRNG fix
4823 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4824 with an unseeded PRNG. The conditions are:
4825 - The client is on a platform where the PRNG has not been seeded
4826 automatically, and the user has not seeded manually
4827 - A protocol specific client method version has been used (i.e. not
4828 SSL_client_methodv23)
4829 - A ciphersuite is used that does not require additional random data from
4830 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4832 If the handshake succeeds then the client random that has been used will
4833 have been generated from a PRNG with insufficient entropy and therefore the
4834 output may be predictable.
4836 For example using the following command with an unseeded openssl will
4837 succeed on an unpatched platform:
4839 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4844 * Use After Free following d2i_ECPrivatekey error fix
4846 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4847 could cause a use after free condition. This, in turn, could cause a double
4848 free in several private key parsing functions (such as d2i_PrivateKey
4849 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4850 for applications that receive EC private keys from untrusted
4851 sources. This scenario is considered rare.
4853 This issue was discovered by the BoringSSL project and fixed in their
4859 * X509_to_X509_REQ NULL pointer deref fix
4861 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4862 the certificate key is invalid. This function is rarely used in practice.
4864 This issue was discovered by Brian Carpenter.
4869 * Removed the export ciphers from the DEFAULT ciphers
4873 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4875 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4876 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4877 So far those who have to target multiple platforms would compromise
4878 and argue that binary targeting say ARMv5 would still execute on
4879 ARMv8. "Universal" build resolves this compromise by providing
4880 near-optimal performance even on newer platforms.
4884 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4885 (other platforms pending).
4887 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4889 * Add support for the SignedCertificateTimestampList certificate and
4890 OCSP response extensions from RFC6962.
4894 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4895 for corner cases. (Certain input points at infinity could lead to
4896 bogus results, with non-infinity inputs mapped to infinity too.)
4900 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4901 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4902 common cases are optimized and there still is room for further
4903 improvements. Vector Permutation AES for Altivec is also added.
4907 * Add support for little-endian ppc64 Linux target.
4909 *Marcelo Cerri (IBM)*
4911 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4912 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4913 are optimized and there still is room for further improvements.
4914 Both 32- and 64-bit modes are supported.
4916 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4918 * Improved ARMv7 NEON support.
4922 * Support for SPARC Architecture 2011 crypto extensions, first
4923 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4924 SHA256/512, MD5, GHASH and modular exponentiation.
4926 *Andy Polyakov, David Miller*
4928 * Accelerated modular exponentiation for Intel processors, a.k.a.
4931 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4933 * Support for new and upcoming Intel processors, including AVX2,
4934 BMI and SHA ISA extensions. This includes additional "stitched"
4935 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4938 This work was sponsored by Intel Corp.
4942 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4943 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4944 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4948 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4949 this fixes a limitation in previous versions of OpenSSL.
4953 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4954 MGF1 digest and OAEP label.
4958 * Add EVP support for key wrapping algorithms, to avoid problems with
4959 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4960 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4961 algorithms and include tests cases.
4965 * Add functions to allocate and set the fields of an ECDSA_METHOD
4968 *Douglas E. Engert, Steve Henson*
4970 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4971 difference in days and seconds between two tm or ASN1_TIME structures.
4975 * Add -rev test option to s_server to just reverse order of characters
4976 received by client and send back to server. Also prints an abbreviated
4977 summary of the connection parameters.
4981 * New option -brief for s_client and s_server to print out a brief summary
4982 of connection parameters.
4986 * Add callbacks for arbitrary TLS extensions.
4988 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4990 * New option -crl_download in several openssl utilities to download CRLs
4991 from CRLDP extension in certificates.
4995 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4999 * New function X509_CRL_diff to generate a delta CRL from the difference
5000 of two full CRLs. Add support to "crl" utility.
5004 * New functions to set lookup_crls function and to retrieve
5005 X509_STORE from X509_STORE_CTX.
5009 * Print out deprecated issuer and subject unique ID fields in
5014 * Extend OCSP I/O functions so they can be used for simple general purpose
5015 HTTP as well as OCSP. New wrapper function which can be used to download
5016 CRLs using the OCSP API.
5020 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5024 * `SSL_CONF*` functions. These provide a common framework for application
5025 configuration using configuration files or command lines.
5029 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5030 message callback and prints the results. Needs compile time option
5031 "enable-ssl-trace". New options to s_client and s_server to enable
5036 * New ctrl and macro to retrieve supported points extensions.
5037 Print out extension in s_server and s_client.
5041 * New functions to retrieve certificate signature and signature
5046 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5051 * New Suite B modes for TLS code. These use and enforce the requirements
5052 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5053 only use Suite B curves. The Suite B modes can be set by using the
5054 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5058 * New chain verification flags for Suite B levels of security. Check
5059 algorithms are acceptable when flags are set in X509_verify_cert.
5063 * Make tls1_check_chain return a set of flags indicating checks passed
5064 by a certificate chain. Add additional tests to handle client
5065 certificates: checks for matching certificate type and issuer name
5070 * If an attempt is made to use a signature algorithm not in the peer
5071 preference list abort the handshake. If client has no suitable
5072 signature algorithms in response to a certificate request do not
5073 use the certificate.
5077 * If server EC tmp key is not in client preference list abort handshake.
5081 * Add support for certificate stores in CERT structure. This makes it
5082 possible to have different stores per SSL structure or one store in
5083 the parent SSL_CTX. Include distinct stores for certificate chain
5084 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5085 to build and store a certificate chain in CERT structure: returning
5086 an error if the chain cannot be built: this will allow applications
5087 to test if a chain is correctly configured.
5089 Note: if the CERT based stores are not set then the parent SSL_CTX
5090 store is used to retain compatibility with existing behaviour.
5094 * New function ssl_set_client_disabled to set a ciphersuite disabled
5095 mask based on the current session, check mask when sending client
5096 hello and checking the requested ciphersuite.
5100 * New ctrls to retrieve and set certificate types in a certificate
5101 request message. Print out received values in s_client. If certificate
5102 types is not set with custom values set sensible values based on
5103 supported signature algorithms.
5107 * Support for distinct client and server supported signature algorithms.
5111 * Add certificate callback. If set this is called whenever a certificate
5112 is required by client or server. An application can decide which
5113 certificate chain to present based on arbitrary criteria: for example
5114 supported signature algorithms. Add very simple example to s_server.
5115 This fixes many of the problems and restrictions of the existing client
5116 certificate callback: for example you can now clear an existing
5117 certificate and specify the whole chain.
5121 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5122 the certificate can be used for (if anything). Set valid_flags field
5123 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5124 to have similar checks in it.
5126 Add new "cert_flags" field to CERT structure and include a "strict mode".
5127 This enforces some TLS certificate requirements (such as only permitting
5128 certificate signature algorithms contained in the supported algorithms
5129 extension) which some implementations ignore: this option should be used
5130 with caution as it could cause interoperability issues.
5134 * Update and tidy signature algorithm extension processing. Work out
5135 shared signature algorithms based on preferences and peer algorithms
5136 and print them out in s_client and s_server. Abort handshake if no
5137 shared signature algorithms.
5141 * Add new functions to allow customised supported signature algorithms
5142 for SSL and SSL_CTX structures. Add options to s_client and s_server
5147 * New function SSL_certs_clear() to delete all references to certificates
5148 from an SSL structure. Before this once a certificate had been added
5149 it couldn't be removed.
5153 * Integrate hostname, email address and IP address checking with certificate
5154 verification. New verify options supporting checking in openssl utility.
5158 * Fixes and wildcard matching support to hostname and email checking
5159 functions. Add manual page.
5161 *Florian Weimer (Red Hat Product Security Team)*
5163 * New functions to check a hostname email or IP address against a
5164 certificate. Add options x509 utility to print results of checks against
5169 * Fix OCSP checking.
5171 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5173 * Initial experimental support for explicitly trusted non-root CAs.
5174 OpenSSL still tries to build a complete chain to a root but if an
5175 intermediate CA has a trust setting included that is used. The first
5176 setting is used: whether to trust (e.g., -addtrust option to the x509
5181 * Add -trusted_first option which attempts to find certificates in the
5182 trusted store even if an untrusted chain is also supplied.
5186 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5187 platform support for Linux and Android.
5191 * Support for linux-x32, ILP32 environment in x86_64 framework.
5195 * Experimental multi-implementation support for FIPS capable OpenSSL.
5196 When in FIPS mode the approved implementations are used as normal,
5197 when not in FIPS mode the internal unapproved versions are used instead.
5198 This means that the FIPS capable OpenSSL isn't forced to use the
5199 (often lower performance) FIPS implementations outside FIPS mode.
5203 * Transparently support X9.42 DH parameters when calling
5204 PEM_read_bio_DHparameters. This means existing applications can handle
5205 the new parameter format automatically.
5209 * Initial experimental support for X9.42 DH parameter format: mainly
5210 to support use of 'q' parameter for RFC5114 parameters.
5214 * Add DH parameters from RFC5114 including test data to dhtest.
5218 * Support for automatic EC temporary key parameter selection. If enabled
5219 the most preferred EC parameters are automatically used instead of
5220 hardcoded fixed parameters. Now a server just has to call:
5221 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5222 support ECDH and use the most appropriate parameters.
5226 * Enhance and tidy EC curve and point format TLS extension code. Use
5227 static structures instead of allocation if default values are used.
5228 New ctrls to set curves we wish to support and to retrieve shared curves.
5229 Print out shared curves in s_server. New options to s_server and s_client
5230 to set list of supported curves.
5234 * New ctrls to retrieve supported signature algorithms and
5235 supported curve values as an array of NIDs. Extend openssl utility
5236 to print out received values.
5240 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5241 between NIDs and the more common NIST names such as "P-256". Enhance
5242 ecparam utility and ECC method to recognise the NIST names for curves.
5246 * Enhance SSL/TLS certificate chain handling to support different
5247 chains for each certificate instead of one chain in the parent SSL_CTX.
5251 * Support for fixed DH ciphersuite client authentication: where both
5252 server and client use DH certificates with common parameters.
5256 * Support for fixed DH ciphersuites: those requiring DH server
5261 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5263 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5264 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5265 X509_CINF_get_signature were reverted post internal team review.
5270 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5272 * OCSP Status Request extension unbounded memory growth
5274 A malicious client can send an excessively large OCSP Status Request
5275 extension. If that client continually requests renegotiation, sending a
5276 large OCSP Status Request extension each time, then there will be unbounded
5277 memory growth on the server. This will eventually lead to a Denial Of
5278 Service attack through memory exhaustion. Servers with a default
5279 configuration are vulnerable even if they do not support OCSP. Builds using
5280 the "no-ocsp" build time option are not affected.
5282 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5287 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5290 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5296 * OOB write in MDC2_Update()
5298 An overflow can occur in MDC2_Update() either if called directly or
5299 through the EVP_DigestUpdate() function using MDC2. If an attacker
5300 is able to supply very large amounts of input data after a previous
5301 call to EVP_EncryptUpdate() with a partial block then a length check
5302 can overflow resulting in a heap corruption.
5304 The amount of data needed is comparable to SIZE_MAX which is impractical
5307 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5312 * Malformed SHA512 ticket DoS
5314 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5315 DoS attack where a malformed ticket will result in an OOB read which will
5318 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5319 a custom server callback and ticket lookup mechanism.
5321 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5326 * OOB write in BN_bn2dec()
5328 The function BN_bn2dec() does not check the return value of BN_div_word().
5329 This can cause an OOB write if an application uses this function with an
5330 overly large BIGNUM. This could be a problem if an overly large certificate
5331 or CRL is printed out from an untrusted source. TLS is not affected because
5332 record limits will reject an oversized certificate before it is parsed.
5334 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5339 * OOB read in TS_OBJ_print_bio()
5341 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5342 the total length the OID text representation would use and not the amount
5343 of data written. This will result in OOB reads when large OIDs are
5346 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5351 * Pointer arithmetic undefined behaviour
5353 Avoid some undefined pointer arithmetic
5355 A common idiom in the codebase is to check limits in the following manner:
5358 Where "p" points to some malloc'd data of SIZE bytes and
5361 "len" here could be from some externally supplied data (e.g. from a TLS
5364 The rules of C pointer arithmetic are such that "p + len" is only well
5365 defined where len <= SIZE. Therefore the above idiom is actually
5366 undefined behaviour.
5368 For example this could cause problems if some malloc implementation
5369 provides an address for "p" such that "p + len" actually overflows for
5370 values of len that are too big and therefore p + len < limit.
5372 This issue was reported to OpenSSL by Guido Vranken
5377 * Constant time flag not preserved in DSA signing
5379 Operations in the DSA signing algorithm should run in constant time in
5380 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5381 implementation means that a non-constant time codepath is followed for
5382 certain operations. This has been demonstrated through a cache-timing
5383 attack to be sufficient for an attacker to recover the private DSA key.
5385 This issue was reported by César Pereida (Aalto University), Billy Brumley
5386 (Tampere University of Technology), and Yuval Yarom (The University of
5387 Adelaide and NICTA).
5392 * DTLS buffered message DoS
5394 In a DTLS connection where handshake messages are delivered out-of-order
5395 those messages that OpenSSL is not yet ready to process will be buffered
5396 for later use. Under certain circumstances, a flaw in the logic means that
5397 those messages do not get removed from the buffer even though the handshake
5398 has been completed. An attacker could force up to approx. 15 messages to
5399 remain in the buffer when they are no longer required. These messages will
5400 be cleared when the DTLS connection is closed. The default maximum size for
5401 a message is 100k. Therefore the attacker could force an additional 1500k
5402 to be consumed per connection. By opening many simulataneous connections an
5403 attacker could cause a DoS attack through memory exhaustion.
5405 This issue was reported to OpenSSL by Quan Luo.
5410 * DTLS replay protection DoS
5412 A flaw in the DTLS replay attack protection mechanism means that records
5413 that arrive for future epochs update the replay protection "window" before
5414 the MAC for the record has been validated. This could be exploited by an
5415 attacker by sending a record for the next epoch (which does not have to
5416 decrypt or have a valid MAC), with a very large sequence number. This means
5417 that all subsequent legitimate packets are dropped causing a denial of
5418 service for a specific DTLS connection.
5420 This issue was reported to OpenSSL by the OCAP audit team.
5425 * Certificate message OOB reads
5427 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5428 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5429 theoretical DoS risk but this has not been observed in practice on common
5432 The messages affected are client certificate, client certificate request
5433 and server certificate. As a result the attack can only be performed
5434 against a client or a server which enables client authentication.
5436 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5441 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5443 * Prevent padding oracle in AES-NI CBC MAC check
5445 A MITM attacker can use a padding oracle attack to decrypt traffic
5446 when the connection uses an AES CBC cipher and the server support
5449 This issue was introduced as part of the fix for Lucky 13 padding
5450 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5451 constant time by making sure that always the same bytes are read and
5452 compared against either the MAC or padding bytes. But it no longer
5453 checked that there was enough data to have both the MAC and padding
5456 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5461 * Fix EVP_EncodeUpdate overflow
5463 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5464 Base64 encoding of binary data. If an attacker is able to supply very large
5465 amounts of input data then a length check can overflow resulting in a heap
5468 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5469 the `PEM_write_bio*` family of functions. These are mainly used within the
5470 OpenSSL command line applications, so any application which processes data
5471 from an untrusted source and outputs it as a PEM file should be considered
5472 vulnerable to this issue. User applications that call these APIs directly
5473 with large amounts of untrusted data may also be vulnerable.
5475 This issue was reported by Guido Vranken.
5480 * Fix EVP_EncryptUpdate overflow
5482 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5483 is able to supply very large amounts of input data after a previous call to
5484 EVP_EncryptUpdate() with a partial block then a length check can overflow
5485 resulting in a heap corruption. Following an analysis of all OpenSSL
5486 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5487 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5488 the first called function after an EVP_EncryptInit(), and therefore that
5489 specific call must be safe. The second form is where the length passed to
5490 EVP_EncryptUpdate() can be seen from the code to be some small value and
5491 therefore there is no possibility of an overflow. Since all instances are
5492 one of these two forms, it is believed that there can be no overflows in
5493 internal code due to this problem. It should be noted that
5494 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5495 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5496 of these calls have also been analysed too and it is believed there are no
5497 instances in internal usage where an overflow could occur.
5499 This issue was reported by Guido Vranken.
5504 * Prevent ASN.1 BIO excessive memory allocation
5506 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5507 a short invalid encoding can casuse allocation of large amounts of memory
5508 potentially consuming excessive resources or exhausting memory.
5510 Any application parsing untrusted data through d2i BIO functions is
5511 affected. The memory based functions such as d2i_X509() are *not* affected.
5512 Since the memory based functions are used by the TLS library, TLS
5513 applications are not affected.
5515 This issue was reported by Brian Carpenter.
5522 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5523 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5524 in arbitrary stack data being returned in the buffer.
5526 This issue was reported by Guido Vranken.
5531 * Modify behavior of ALPN to invoke callback after SNI/servername
5532 callback, such that updates to the SSL_CTX affect ALPN.
5536 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5541 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5542 methods are enabled and ssl2 is disabled the methods return NULL.
5546 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5548 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5549 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5550 provide any "EXPORT" or "LOW" strength ciphers.
5554 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5555 is by default disabled at build-time. Builds that are not configured with
5556 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5557 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5558 will need to explicitly call either of:
5560 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5562 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5564 as appropriate. Even if either of those is used, or the application
5565 explicitly uses the version-specific SSLv2_method() or its client and
5566 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5567 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5568 ciphers, and SSLv2 56-bit DES are no longer available.
5573 * Fix a double-free in DSA code
5575 A double free bug was discovered when OpenSSL parses malformed DSA private
5576 keys and could lead to a DoS attack or memory corruption for applications
5577 that receive DSA private keys from untrusted sources. This scenario is
5580 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5586 * Disable SRP fake user seed to address a server memory leak.
5588 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5590 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5591 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5592 was changed to ignore the "fake user" SRP seed, even if the seed
5595 Users should use SRP_VBASE_get1_by_user instead. Note that in
5596 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5597 also that even though configuring the SRP seed attempts to hide
5598 invalid usernames by continuing the handshake with fake
5599 credentials, this behaviour is not constant time and no strong
5600 guarantees are made that the handshake is indistinguishable from
5601 that of a valid user.
5606 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5608 In the BN_hex2bn function the number of hex digits is calculated using an
5609 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5610 large values of |i| this can result in |bn_expand| not allocating any
5611 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5612 field as NULL leading to a subsequent NULL ptr deref. For very large values
5613 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5614 In this case memory is allocated to the internal BIGNUM data field, but it
5615 is insufficiently sized leading to heap corruption. A similar issue exists
5616 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5617 is ever called by user applications with very large untrusted hex/dec data.
5618 This is anticipated to be a rare occurrence.
5620 All OpenSSL internal usage of these functions use data that is not expected
5621 to be untrusted, e.g. config file data or application command line
5622 arguments. If user developed applications generate config file data based
5623 on untrusted data then it is possible that this could also lead to security
5624 consequences. This is also anticipated to be rare.
5626 This issue was reported to OpenSSL by Guido Vranken.
5631 * Fix memory issues in `BIO_*printf` functions
5633 The internal |fmtstr| function used in processing a "%s" format string in
5634 the `BIO_*printf` functions could overflow while calculating the length of a
5635 string and cause an OOB read when printing very long strings.
5637 Additionally the internal |doapr_outch| function can attempt to write to an
5638 OOB memory location (at an offset from the NULL pointer) in the event of a
5639 memory allocation failure. In 1.0.2 and below this could be caused where
5640 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5641 could be in processing a very long "%s" format string. Memory leaks can
5644 The first issue may mask the second issue dependent on compiler behaviour.
5645 These problems could enable attacks where large amounts of untrusted data
5646 is passed to the `BIO_*printf` functions. If applications use these functions
5647 in this way then they could be vulnerable. OpenSSL itself uses these
5648 functions when printing out human-readable dumps of ASN.1 data. Therefore
5649 applications that print this data could be vulnerable if the data is from
5650 untrusted sources. OpenSSL command line applications could also be
5651 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5652 as command line arguments.
5654 Libssl is not considered directly vulnerable. Additionally certificates etc
5655 received via remote connections via libssl are also unlikely to be able to
5656 trigger these issues because of message size limits enforced within libssl.
5658 This issue was reported to OpenSSL Guido Vranken.
5663 * Side channel attack on modular exponentiation
5665 A side-channel attack was found which makes use of cache-bank conflicts on
5666 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5667 of RSA keys. The ability to exploit this issue is limited as it relies on
5668 an attacker who has control of code in a thread running on the same
5669 hyper-threaded core as the victim thread which is performing decryptions.
5671 This issue was reported to OpenSSL by Yuval Yarom, The University of
5672 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5673 Nadia Heninger, University of Pennsylvania with more information at
5674 <http://cachebleed.info>.
5679 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5680 if no keysize is specified with default_bits. This fixes an
5681 omission in an earlier change that changed all RSA/DSA key generation
5682 apps to use 2048 bits by default.
5686 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5688 * Protection for DH small subgroup attacks
5690 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5691 switched on by default and cannot be disabled. This could have some
5696 * SSLv2 doesn't block disabled ciphers
5698 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5699 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5700 been disabled, provided that the SSLv2 protocol was not also disabled via
5703 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5704 and Sebastian Schinzel.
5709 * Reject DH handshakes with parameters shorter than 1024 bits.
5713 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5715 * Certificate verify crash with missing PSS parameter
5717 The signature verification routines will crash with a NULL pointer
5718 dereference if presented with an ASN.1 signature using the RSA PSS
5719 algorithm and absent mask generation function parameter. Since these
5720 routines are used to verify certificate signature algorithms this can be
5721 used to crash any certificate verification operation and exploited in a
5722 DoS attack. Any application which performs certificate verification is
5723 vulnerable including OpenSSL clients and servers which enable client
5726 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5731 * X509_ATTRIBUTE memory leak
5733 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5734 memory. This structure is used by the PKCS#7 and CMS routines so any
5735 application which reads PKCS#7 or CMS data from untrusted sources is
5736 affected. SSL/TLS is not affected.
5738 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5744 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5745 This changes the decoding behaviour for some invalid messages,
5746 though the change is mostly in the more lenient direction, and
5747 legacy behaviour is preserved as much as possible.
5751 * In DSA_generate_parameters_ex, if the provided seed is too short,
5752 use a random seed, as already documented.
5754 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5756 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5758 * Alternate chains certificate forgery
5760 During certificate verfification, OpenSSL will attempt to find an
5761 alternative certificate chain if the first attempt to build such a chain
5762 fails. An error in the implementation of this logic can mean that an
5763 attacker could cause certain checks on untrusted certificates to be
5764 bypassed, such as the CA flag, enabling them to use a valid leaf
5765 certificate to act as a CA and "issue" an invalid certificate.
5767 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5773 * Race condition handling PSK identify hint
5775 If PSK identity hints are received by a multi-threaded client then
5776 the values are wrongly updated in the parent SSL_CTX structure. This can
5777 result in a race condition potentially leading to a double free of the
5783 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5785 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5786 incompatibility in the handling of HMAC. The previous ABI has now been
5789 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5791 * Malformed ECParameters causes infinite loop
5793 When processing an ECParameters structure OpenSSL enters an infinite loop
5794 if the curve specified is over a specially malformed binary polynomial
5797 This can be used to perform denial of service against any
5798 system which processes public keys, certificate requests or
5799 certificates. This includes TLS clients and TLS servers with
5800 client authentication enabled.
5802 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5807 * Exploitable out-of-bounds read in X509_cmp_time
5809 X509_cmp_time does not properly check the length of the ASN1_TIME
5810 string and can read a few bytes out of bounds. In addition,
5811 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5814 An attacker can use this to craft malformed certificates and CRLs of
5815 various sizes and potentially cause a segmentation fault, resulting in
5816 a DoS on applications that verify certificates or CRLs. TLS clients
5817 that verify CRLs are affected. TLS clients and servers with client
5818 authentication enabled may be affected if they use custom verification
5821 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5822 independently by Hanno Böck.
5827 * PKCS7 crash with missing EnvelopedContent
5829 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5830 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5831 with missing content and trigger a NULL pointer dereference on parsing.
5833 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5834 structures from untrusted sources are affected. OpenSSL clients and
5835 servers are not affected.
5837 This issue was reported to OpenSSL by Michal Zalewski (Google).
5842 * CMS verify infinite loop with unknown hash function
5844 When verifying a signedData message the CMS code can enter an infinite loop
5845 if presented with an unknown hash function OID. This can be used to perform
5846 denial of service against any system which verifies signedData messages using
5848 This issue was reported to OpenSSL by Johannes Bauer.
5853 * Race condition handling NewSessionTicket
5855 If a NewSessionTicket is received by a multi-threaded client when attempting to
5856 reuse a previous ticket then a race condition can occur potentially leading to
5857 a double free of the ticket data.
5862 * Reject DH handshakes with parameters shorter than 768 bits.
5864 *Kurt Roeckx and Emilia Kasper*
5866 * dhparam: generate 2048-bit parameters by default.
5868 *Kurt Roeckx and Emilia Kasper*
5870 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5872 * Segmentation fault in ASN1_TYPE_cmp fix
5874 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5875 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5876 certificate signature algorithm consistency this can be used to crash any
5877 certificate verification operation and exploited in a DoS attack. Any
5878 application which performs certificate verification is vulnerable including
5879 OpenSSL clients and servers which enable client authentication.
5884 * ASN.1 structure reuse memory corruption fix
5886 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5887 memory corruption via an invalid write. Such reuse is and has been
5888 strongly discouraged and is believed to be rare.
5890 Applications that parse structures containing CHOICE or ANY DEFINED BY
5891 components may be affected. Certificate parsing (d2i_X509 and related
5892 functions) are however not affected. OpenSSL clients and servers are
5898 * PKCS7 NULL pointer dereferences fix
5900 The PKCS#7 parsing code does not handle missing outer ContentInfo
5901 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5902 missing content and trigger a NULL pointer dereference on parsing.
5904 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5905 otherwise parse PKCS#7 structures from untrusted sources are
5906 affected. OpenSSL clients and servers are not affected.
5908 This issue was reported to OpenSSL by Michal Zalewski (Google).
5913 * DoS via reachable assert in SSLv2 servers fix
5915 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5916 servers that both support SSLv2 and enable export cipher suites by sending
5917 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5919 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5920 (OpenSSL development team).
5925 * Use After Free following d2i_ECPrivatekey error fix
5927 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5928 could cause a use after free condition. This, in turn, could cause a double
5929 free in several private key parsing functions (such as d2i_PrivateKey
5930 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5931 for applications that receive EC private keys from untrusted
5932 sources. This scenario is considered rare.
5934 This issue was discovered by the BoringSSL project and fixed in their
5940 * X509_to_X509_REQ NULL pointer deref fix
5942 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5943 the certificate key is invalid. This function is rarely used in practice.
5945 This issue was discovered by Brian Carpenter.
5950 * Removed the export ciphers from the DEFAULT ciphers
5954 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
5956 * Build fixes for the Windows and OpenVMS platforms
5958 *Matt Caswell and Richard Levitte*
5960 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
5962 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5963 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5964 dereference. This could lead to a Denial Of Service attack. Thanks to
5965 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5970 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5971 dtls1_buffer_record function under certain conditions. In particular this
5972 could occur if an attacker sent repeated DTLS records with the same
5973 sequence number but for the next epoch. The memory leak could be exploited
5974 by an attacker in a Denial of Service attack through memory exhaustion.
5975 Thanks to Chris Mueller for reporting this issue.
5980 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5981 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5982 method would be set to NULL which could later result in a NULL pointer
5983 dereference. Thanks to Frank Schmirler for reporting this issue.
5988 * Abort handshake if server key exchange message is omitted for ephemeral
5991 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5992 reporting this issue.
5997 * Remove non-export ephemeral RSA code on client and server. This code
5998 violated the TLS standard by allowing the use of temporary RSA keys in
5999 non-export ciphersuites and could be used by a server to effectively
6000 downgrade the RSA key length used to a value smaller than the server
6001 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6002 INRIA or reporting this issue.
6007 * Fixed issue where DH client certificates are accepted without verification.
6008 An OpenSSL server will accept a DH certificate for client authentication
6009 without the certificate verify message. This effectively allows a client to
6010 authenticate without the use of a private key. This only affects servers
6011 which trust a client certificate authority which issues certificates
6012 containing DH keys: these are extremely rare and hardly ever encountered.
6013 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6019 * Ensure that the session ID context of an SSL is updated when its
6020 SSL_CTX is updated via SSL_set_SSL_CTX.
6022 The session ID context is typically set from the parent SSL_CTX,
6023 and can vary with the CTX.
6027 * Fix various certificate fingerprint issues.
6029 By using non-DER or invalid encodings outside the signed portion of a
6030 certificate the fingerprint can be changed without breaking the signature.
6031 Although no details of the signed portion of the certificate can be changed
6032 this can cause problems with some applications: e.g. those using the
6033 certificate fingerprint for blacklists.
6035 1. Reject signatures with non zero unused bits.
6037 If the BIT STRING containing the signature has non zero unused bits reject
6038 the signature. All current signature algorithms require zero unused bits.
6040 2. Check certificate algorithm consistency.
6042 Check the AlgorithmIdentifier inside TBS matches the one in the
6043 certificate signature. NB: this will result in signature failure
6044 errors for some broken certificates.
6046 Thanks to Konrad Kraszewski from Google for reporting this issue.
6048 3. Check DSA/ECDSA signatures use DER.
6050 Re-encode DSA/ECDSA signatures and compare with the original received
6051 signature. Return an error if there is a mismatch.
6053 This will reject various cases including garbage after signature
6054 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6055 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6056 (negative or with leading zeroes).
6058 Further analysis was conducted and fixes were developed by Stephen Henson
6059 of the OpenSSL core team.
6065 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6066 results on some platforms, including x86_64. This bug occurs at random
6067 with a very low probability, and is not known to be exploitable in any
6068 way, though its exact impact is difficult to determine. Thanks to Pieter
6069 Wuille (Blockstream) who reported this issue and also suggested an initial
6070 fix. Further analysis was conducted by the OpenSSL development team and
6071 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6072 the OpenSSL core team.
6077 * Do not resume sessions on the server if the negotiated protocol
6078 version does not match the session's version. Resuming with a different
6079 version, while not strictly forbidden by the RFC, is of questionable
6080 sanity and breaks all known clients.
6082 *David Benjamin, Emilia Käsper*
6084 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6085 early CCS messages during renegotiation. (Note that because
6086 renegotiation is encrypted, this early CCS was not exploitable.)
6090 * Tighten client-side session ticket handling during renegotiation:
6091 ensure that the client only accepts a session ticket if the server sends
6092 the extension anew in the ServerHello. Previously, a TLS client would
6093 reuse the old extension state and thus accept a session ticket if one was
6094 announced in the initial ServerHello.
6096 Similarly, ensure that the client requires a session ticket if one
6097 was advertised in the ServerHello. Previously, a TLS client would
6098 ignore a missing NewSessionTicket message.
6102 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6106 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6107 sends a carefully crafted handshake message, to cause OpenSSL to fail
6108 to free up to 64k of memory causing a memory leak. This could be
6109 exploited in a Denial Of Service attack. This issue affects OpenSSL
6110 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6111 whether SRTP is used or configured. Implementations of OpenSSL that
6112 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6114 The fix was developed by the OpenSSL team.
6119 * Session Ticket Memory Leak.
6121 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6122 integrity of that ticket is first verified. In the event of a session
6123 ticket integrity check failing, OpenSSL will fail to free memory
6124 causing a memory leak. By sending a large number of invalid session
6125 tickets an attacker could exploit this issue in a Denial Of Service
6131 * Build option no-ssl3 is incomplete.
6133 When OpenSSL is configured with "no-ssl3" as a build option, servers
6134 could accept and complete a SSL 3.0 handshake, and clients could be
6135 configured to send them.
6138 *Akamai and the OpenSSL team*
6140 * Add support for TLS_FALLBACK_SCSV.
6141 Client applications doing fallback retries should call
6142 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6145 *Adam Langley, Bodo Moeller*
6147 * Add additional DigestInfo checks.
6149 Re-encode DigestInto in DER and check against the original when
6150 verifying RSA signature: this will reject any improperly encoded
6151 DigestInfo structures.
6153 Note: this is a precautionary measure and no attacks are currently known.
6157 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6159 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6160 SRP code can be overrun an internal buffer. Add sanity check that
6161 g, A, B < N to SRP code.
6163 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6164 Group for discovering this issue.
6169 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6170 TLS 1.0 instead of higher protocol versions when the ClientHello message
6171 is badly fragmented. This allows a man-in-the-middle attacker to force a
6172 downgrade to TLS 1.0 even if both the server and the client support a
6173 higher protocol version, by modifying the client's TLS records.
6175 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6176 researching this issue.
6181 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6182 to a denial of service attack. A malicious server can crash the client
6183 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6184 ciphersuite and sending carefully crafted handshake messages.
6186 Thanks to Felix Gröbert (Google) for discovering and researching this
6192 * By sending carefully crafted DTLS packets an attacker could cause openssl
6193 to leak memory. This can be exploited through a Denial of Service attack.
6194 Thanks to Adam Langley for discovering and researching this issue.
6199 * An attacker can force openssl to consume large amounts of memory whilst
6200 processing DTLS handshake messages. This can be exploited through a
6201 Denial of Service attack.
6202 Thanks to Adam Langley for discovering and researching this issue.
6207 * An attacker can force an error condition which causes openssl to crash
6208 whilst processing DTLS packets due to memory being freed twice. This
6209 can be exploited through a Denial of Service attack.
6210 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6216 * If a multithreaded client connects to a malicious server using a resumed
6217 session and the server sends an ec point format extension it could write
6218 up to 255 bytes to freed memory.
6220 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6226 * A malicious server can crash an OpenSSL client with a null pointer
6227 dereference (read) by specifying an SRP ciphersuite even though it was not
6228 properly negotiated with the client. This can be exploited through a
6229 Denial of Service attack.
6231 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6232 discovering and researching this issue.
6237 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6238 X509_name_oneline, X509_name_print_ex et al. to leak some information
6239 from the stack. Applications may be affected if they echo pretty printing
6240 output to the attacker.
6242 Thanks to Ivan Fratric (Google) for discovering this issue.
6245 *Emilia Käsper, and Steve Henson*
6247 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6248 for corner cases. (Certain input points at infinity could lead to
6249 bogus results, with non-infinity inputs mapped to infinity too.)
6253 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6255 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6256 handshake can force the use of weak keying material in OpenSSL
6257 SSL/TLS clients and servers.
6259 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6260 researching this issue. [CVE-2014-0224][]
6262 *KIKUCHI Masashi, Steve Henson*
6264 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6265 OpenSSL DTLS client the code can be made to recurse eventually crashing
6268 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6271 *Imre Rad, Steve Henson*
6273 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6274 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6275 client or server. This is potentially exploitable to run arbitrary
6276 code on a vulnerable client or server.
6278 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6280 *Jüri Aedla, Steve Henson*
6282 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6283 are subject to a denial of service attack.
6285 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6286 this issue. [CVE-2014-3470][]
6288 *Felix Gröbert, Ivan Fratric, Steve Henson*
6290 * Harmonize version and its documentation. -f flag is used to display
6293 *mancha <mancha1@zoho.com>*
6295 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6296 in i2d_ECPrivateKey.
6298 *mancha <mancha1@zoho.com>*
6300 * Fix some double frees. These are not thought to be exploitable.
6302 *mancha <mancha1@zoho.com>*
6304 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6306 * A missing bounds check in the handling of the TLS heartbeat extension
6307 can be used to reveal up to 64k of memory to a connected client or
6310 Thanks for Neel Mehta of Google Security for discovering this bug and to
6311 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6312 preparing the fix [CVE-2014-0160][]
6314 *Adam Langley, Bodo Moeller*
6316 * Fix for the attack described in the paper "Recovering OpenSSL
6317 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6318 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6319 <http://eprint.iacr.org/2014/140>
6321 Thanks to Yuval Yarom and Naomi Benger for discovering this
6322 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6324 *Yuval Yarom and Naomi Benger*
6326 * TLS pad extension: draft-agl-tls-padding-03
6328 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6329 TLS client Hello record length value would otherwise be > 255 and
6330 less that 512 pad with a dummy extension containing zeroes so it
6331 is at least 512 bytes long.
6333 *Adam Langley, Steve Henson*
6335 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6337 * Fix for TLS record tampering bug. A carefully crafted invalid
6338 handshake could crash OpenSSL with a NULL pointer exception.
6339 Thanks to Anton Johansson for reporting this issues.
6342 * Keep original DTLS digest and encryption contexts in retransmission
6343 structures so we can use the previous session parameters if they need
6344 to be resent. [CVE-2013-6450][]
6348 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6349 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6350 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6351 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6352 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6353 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6355 *Rob Stradling, Adam Langley*
6357 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6359 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6360 supporting platforms or when small records were transferred.
6362 *Andy Polyakov, Steve Henson*
6364 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6366 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6368 This addresses the flaw in CBC record processing discovered by
6369 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6370 at: <http://www.isg.rhul.ac.uk/tls/>
6372 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6373 Security Group at Royal Holloway, University of London
6374 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6375 Emilia Käsper for the initial patch.
6378 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6380 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6381 ciphersuites which can be exploited in a denial of service attack.
6382 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6383 and detecting this bug and to Wolfgang Ettlinger
6384 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6389 * Return an error when checking OCSP signatures when key is NULL.
6390 This fixes a DoS attack. [CVE-2013-0166][]
6394 * Make openssl verify return errors.
6396 *Chris Palmer <palmer@google.com> and Ben Laurie*
6398 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6399 the right response is stapled. Also change SSL_get_certificate()
6400 so it returns the certificate actually sent.
6401 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6403 *Rob Stradling <rob.stradling@comodo.com>*
6405 * Fix possible deadlock when decoding public keys.
6409 * Don't use TLS 1.0 record version number in initial client hello
6414 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6416 * Sanity check record length before skipping explicit IV in TLS
6417 1.2, 1.1 and DTLS to fix DoS attack.
6419 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6420 fuzzing as a service testing platform.
6425 * Initialise tkeylen properly when encrypting CMS messages.
6426 Thanks to Solar Designer of Openwall for reporting this issue.
6430 * In FIPS mode don't try to use composite ciphers as they are not
6435 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6437 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6438 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6439 mean any application compiled against OpenSSL 1.0.0 headers setting
6440 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6441 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6442 0x10000000L Any application which was previously compiled against
6443 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6444 will need to be recompiled as a result. Letting be results in
6445 inability to disable specifically TLS 1.1 and in client context,
6446 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6450 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6451 disable just protocol X, but all protocols above X *if* there are
6452 protocols *below* X still enabled. In more practical terms it means
6453 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6454 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6455 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6460 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6462 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6463 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6464 in CRYPTO_realloc_clean.
6466 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6467 issue and to Adam Langley <agl@chromium.org> for fixing it.
6470 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6472 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6476 * Workarounds for some broken servers that "hang" if a client hello
6477 record length exceeds 255 bytes.
6479 1. Do not use record version number > TLS 1.0 in initial client
6480 hello: some (but not all) hanging servers will now work.
6481 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6482 the number of ciphers sent in the client hello. This should be
6483 set to an even number, such as 50, for example by passing:
6484 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6485 Most broken servers should now work.
6486 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6487 TLS 1.2 client support entirely.
6491 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6495 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6497 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6498 STRING form instead of a DigestInfo.
6502 * The format used for MDC2 RSA signatures is inconsistent between EVP
6503 and the RSA_sign/RSA_verify functions. This was made more apparent when
6504 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6505 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6506 the correct format in RSA_verify so both forms transparently work.
6510 * Some servers which support TLS 1.0 can choke if we initially indicate
6511 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6512 encrypted premaster secret. As a workaround use the maximum permitted
6513 client version in client hello, this should keep such servers happy
6514 and still work with previous versions of OpenSSL.
6518 * Add support for TLS/DTLS heartbeats.
6520 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6522 * Add support for SCTP.
6524 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6526 * Improved PRNG seeding for VOS.
6528 *Paul Green <Paul.Green@stratus.com>*
6530 * Extensive assembler packs updates, most notably:
6532 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6533 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6534 - x86_64: bit-sliced AES implementation;
6535 - ARM: NEON support, contemporary platforms optimizations;
6536 - s390x: z196 support;
6537 - `*`: GHASH and GF(2^m) multiplication implementations;
6541 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6542 (removal of unnecessary code)
6544 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6546 * Add TLS key material exporter from RFC 5705.
6550 * Add DTLS-SRTP negotiation from RFC 5764.
6554 * Add Next Protocol Negotiation,
6555 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6556 disabled with a no-npn flag to config or Configure. Code donated
6559 *Adam Langley <agl@google.com> and Ben Laurie*
6561 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6562 NIST-P256, NIST-P521, with constant-time single point multiplication on
6563 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6564 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6565 Code made available under Apache License version 2.0.
6567 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6568 line to include this in your build of OpenSSL, and run "make depend" (or
6569 "make update"). This enables the following EC_METHODs:
6571 EC_GFp_nistp224_method()
6572 EC_GFp_nistp256_method()
6573 EC_GFp_nistp521_method()
6575 EC_GROUP_new_by_curve_name() will automatically use these (while
6576 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6579 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6581 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6582 all platforms. Move ssize_t definition from e_os.h to the public
6583 header file e_os2.h as it now appears in public header file cms.h
6587 * New -sigopt option to the ca, req and x509 utilities. Additional
6588 signature parameters can be passed using this option and in
6593 * Add RSA PSS signing function. This will generate and set the
6594 appropriate AlgorithmIdentifiers for PSS based on those in the
6595 corresponding EVP_MD_CTX structure. No application support yet.
6599 * Support for companion algorithm specific ASN1 signing routines.
6600 New function ASN1_item_sign_ctx() signs a pre-initialised
6601 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6602 the appropriate parameters.
6606 * Add new algorithm specific ASN1 verification initialisation function
6607 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6608 handling will be the same no matter what EVP_PKEY_METHOD is used.
6609 Add a PSS handler to support verification of PSS signatures: checked
6610 against a number of sample certificates.
6614 * Add signature printing for PSS. Add PSS OIDs.
6616 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6618 * Add algorithm specific signature printing. An individual ASN1 method
6619 can now print out signatures instead of the standard hex dump.
6621 More complex signatures (e.g. PSS) can print out more meaningful
6622 information. Include DSA version that prints out the signature
6627 * Password based recipient info support for CMS library: implementing
6632 * Split password based encryption into PBES2 and PBKDF2 functions. This
6633 neatly separates the code into cipher and PBE sections and is required
6634 for some algorithms that split PBES2 into separate pieces (such as
6635 password based CMS).
6639 * Session-handling fixes:
6640 - Fix handling of connections that are resuming with a session ID,
6641 but also support Session Tickets.
6642 - Fix a bug that suppressed issuing of a new ticket if the client
6643 presented a ticket with an expired session.
6644 - Try to set the ticket lifetime hint to something reasonable.
6645 - Make tickets shorter by excluding irrelevant information.
6646 - On the client side, don't ignore renewed tickets.
6648 *Adam Langley, Bodo Moeller (Google)*
6650 * Fix PSK session representation.
6654 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6656 This work was sponsored by Intel.
6660 * Add GCM support to TLS library. Some custom code is needed to split
6661 the IV between the fixed (from PRF) and explicit (from TLS record)
6662 portions. This adds all GCM ciphersuites supported by RFC5288 and
6663 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6664 add a special AESGCM string for GCM only.
6668 * Expand range of ctrls for AES GCM. Permit setting invocation
6669 field on decrypt and retrieval of invocation field only on encrypt.
6673 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6674 As required by RFC5289 these ciphersuites cannot be used if for
6675 versions of TLS earlier than 1.2.
6679 * For FIPS capable OpenSSL interpret a NULL default public key method
6680 as unset and return the appropriate default but do *not* set the default.
6681 This means we can return the appropriate method in applications that
6682 switch between FIPS and non-FIPS modes.
6686 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6687 ENGINE is used then we cannot handle that in the FIPS module so we
6688 keep original code iff non-FIPS operations are allowed.
6692 * Add -attime option to openssl utilities.
6694 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6696 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6700 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6701 FIPS EC methods unconditionally for now.
6705 * New build option no-ec2m to disable characteristic 2 code.
6709 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6710 all cases can be covered as some introduce binary incompatibilities.
6714 * Redirect RSA operations to FIPS module including keygen,
6715 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6719 * Add similar low level API blocking to ciphers.
6723 * Low level digest APIs are not approved in FIPS mode: any attempt
6724 to use these will cause a fatal error. Applications that *really* want
6725 to use them can use the `private_*` version instead.
6729 * Redirect cipher operations to FIPS module for FIPS builds.
6733 * Redirect digest operations to FIPS module for FIPS builds.
6737 * Update build system to add "fips" flag which will link in fipscanister.o
6738 for static and shared library builds embedding a signature if needed.
6742 * Output TLS supported curves in preference order instead of numerical
6743 order. This is currently hardcoded for the highest order curves first.
6744 This should be configurable so applications can judge speed vs strength.
6748 * Add TLS v1.2 server support for client authentication.
6752 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6757 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6758 FIPS modules versions.
6762 * Add TLS v1.2 client side support for client authentication. Keep cache
6763 of handshake records longer as we don't know the hash algorithm to use
6764 until after the certificate request message is received.
6768 * Initial TLS v1.2 client support. Add a default signature algorithms
6769 extension including all the algorithms we support. Parse new signature
6770 format in client key exchange. Relax some ECC signing restrictions for
6771 TLS v1.2 as indicated in RFC5246.
6775 * Add server support for TLS v1.2 signature algorithms extension. Switch
6776 to new signature format when needed using client digest preference.
6777 All server ciphersuites should now work correctly in TLS v1.2. No client
6778 support yet and no support for client certificates.
6782 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6783 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6784 ciphersuites. At present only RSA key exchange ciphersuites work with
6785 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6786 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6787 and version checking.
6791 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6792 with this defined it will not be affected by any changes to ssl internal
6793 structures. Add several utility functions to allow openssl application
6794 to work with OPENSSL_NO_SSL_INTERN defined.
6798 * A long standing patch to add support for SRP from EdelWeb (Peter
6799 Sylvester and Christophe Renou) was integrated.
6800 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6801 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6804 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6808 * Permit abbreviated handshakes when renegotiating using the function
6809 SSL_renegotiate_abbreviated().
6811 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6813 * Add call to ENGINE_register_all_complete() to
6814 ENGINE_load_builtin_engines(), so some implementations get used
6815 automatically instead of needing explicit application support.
6819 * Add support for TLS key exporter as described in RFC5705.
6821 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6823 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6824 a few changes are required:
6826 Add SSL_OP_NO_TLSv1_1 flag.
6827 Add TLSv1_1 methods.
6828 Update version checking logic to handle version 1.1.
6829 Add explicit IV handling (ported from DTLS code).
6830 Add command line options to s_client/s_server.
6837 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6839 * X509_ATTRIBUTE memory leak
6841 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6842 memory. This structure is used by the PKCS#7 and CMS routines so any
6843 application which reads PKCS#7 or CMS data from untrusted sources is
6844 affected. SSL/TLS is not affected.
6846 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6852 * Race condition handling PSK identify hint
6854 If PSK identity hints are received by a multi-threaded client then
6855 the values are wrongly updated in the parent SSL_CTX structure. This can
6856 result in a race condition potentially leading to a double free of the
6862 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6864 * Malformed ECParameters causes infinite loop
6866 When processing an ECParameters structure OpenSSL enters an infinite loop
6867 if the curve specified is over a specially malformed binary polynomial
6870 This can be used to perform denial of service against any
6871 system which processes public keys, certificate requests or
6872 certificates. This includes TLS clients and TLS servers with
6873 client authentication enabled.
6875 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6880 * Exploitable out-of-bounds read in X509_cmp_time
6882 X509_cmp_time does not properly check the length of the ASN1_TIME
6883 string and can read a few bytes out of bounds. In addition,
6884 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6887 An attacker can use this to craft malformed certificates and CRLs of
6888 various sizes and potentially cause a segmentation fault, resulting in
6889 a DoS on applications that verify certificates or CRLs. TLS clients
6890 that verify CRLs are affected. TLS clients and servers with client
6891 authentication enabled may be affected if they use custom verification
6894 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6895 independently by Hanno Böck.
6900 * PKCS7 crash with missing EnvelopedContent
6902 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6903 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6904 with missing content and trigger a NULL pointer dereference on parsing.
6906 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6907 structures from untrusted sources are affected. OpenSSL clients and
6908 servers are not affected.
6910 This issue was reported to OpenSSL by Michal Zalewski (Google).
6915 * CMS verify infinite loop with unknown hash function
6917 When verifying a signedData message the CMS code can enter an infinite loop
6918 if presented with an unknown hash function OID. This can be used to perform
6919 denial of service against any system which verifies signedData messages using
6921 This issue was reported to OpenSSL by Johannes Bauer.
6926 * Race condition handling NewSessionTicket
6928 If a NewSessionTicket is received by a multi-threaded client when attempting to
6929 reuse a previous ticket then a race condition can occur potentially leading to
6930 a double free of the ticket data.
6935 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
6937 * Segmentation fault in ASN1_TYPE_cmp fix
6939 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6940 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6941 certificate signature algorithm consistency this can be used to crash any
6942 certificate verification operation and exploited in a DoS attack. Any
6943 application which performs certificate verification is vulnerable including
6944 OpenSSL clients and servers which enable client authentication.
6949 * ASN.1 structure reuse memory corruption fix
6951 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6952 memory corruption via an invalid write. Such reuse is and has been
6953 strongly discouraged and is believed to be rare.
6955 Applications that parse structures containing CHOICE or ANY DEFINED BY
6956 components may be affected. Certificate parsing (d2i_X509 and related
6957 functions) are however not affected. OpenSSL clients and servers are
6963 * PKCS7 NULL pointer dereferences fix
6965 The PKCS#7 parsing code does not handle missing outer ContentInfo
6966 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6967 missing content and trigger a NULL pointer dereference on parsing.
6969 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6970 otherwise parse PKCS#7 structures from untrusted sources are
6971 affected. OpenSSL clients and servers are not affected.
6973 This issue was reported to OpenSSL by Michal Zalewski (Google).
6978 * DoS via reachable assert in SSLv2 servers fix
6980 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6981 servers that both support SSLv2 and enable export cipher suites by sending
6982 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6984 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6985 (OpenSSL development team).
6990 * Use After Free following d2i_ECPrivatekey error fix
6992 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6993 could cause a use after free condition. This, in turn, could cause a double
6994 free in several private key parsing functions (such as d2i_PrivateKey
6995 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6996 for applications that receive EC private keys from untrusted
6997 sources. This scenario is considered rare.
6999 This issue was discovered by the BoringSSL project and fixed in their
7005 * X509_to_X509_REQ NULL pointer deref fix
7007 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7008 the certificate key is invalid. This function is rarely used in practice.
7010 This issue was discovered by Brian Carpenter.
7015 * Removed the export ciphers from the DEFAULT ciphers
7019 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7021 * Build fixes for the Windows and OpenVMS platforms
7023 *Matt Caswell and Richard Levitte*
7025 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7027 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7028 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7029 dereference. This could lead to a Denial Of Service attack. Thanks to
7030 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7035 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7036 dtls1_buffer_record function under certain conditions. In particular this
7037 could occur if an attacker sent repeated DTLS records with the same
7038 sequence number but for the next epoch. The memory leak could be exploited
7039 by an attacker in a Denial of Service attack through memory exhaustion.
7040 Thanks to Chris Mueller for reporting this issue.
7045 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7046 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7047 method would be set to NULL which could later result in a NULL pointer
7048 dereference. Thanks to Frank Schmirler for reporting this issue.
7053 * Abort handshake if server key exchange message is omitted for ephemeral
7056 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7057 reporting this issue.
7062 * Remove non-export ephemeral RSA code on client and server. This code
7063 violated the TLS standard by allowing the use of temporary RSA keys in
7064 non-export ciphersuites and could be used by a server to effectively
7065 downgrade the RSA key length used to a value smaller than the server
7066 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7067 INRIA or reporting this issue.
7072 * Fixed issue where DH client certificates are accepted without verification.
7073 An OpenSSL server will accept a DH certificate for client authentication
7074 without the certificate verify message. This effectively allows a client to
7075 authenticate without the use of a private key. This only affects servers
7076 which trust a client certificate authority which issues certificates
7077 containing DH keys: these are extremely rare and hardly ever encountered.
7078 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7084 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7085 results on some platforms, including x86_64. This bug occurs at random
7086 with a very low probability, and is not known to be exploitable in any
7087 way, though its exact impact is difficult to determine. Thanks to Pieter
7088 Wuille (Blockstream) who reported this issue and also suggested an initial
7089 fix. Further analysis was conducted by the OpenSSL development team and
7090 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7091 the OpenSSL core team.
7096 * Fix various certificate fingerprint issues.
7098 By using non-DER or invalid encodings outside the signed portion of a
7099 certificate the fingerprint can be changed without breaking the signature.
7100 Although no details of the signed portion of the certificate can be changed
7101 this can cause problems with some applications: e.g. those using the
7102 certificate fingerprint for blacklists.
7104 1. Reject signatures with non zero unused bits.
7106 If the BIT STRING containing the signature has non zero unused bits reject
7107 the signature. All current signature algorithms require zero unused bits.
7109 2. Check certificate algorithm consistency.
7111 Check the AlgorithmIdentifier inside TBS matches the one in the
7112 certificate signature. NB: this will result in signature failure
7113 errors for some broken certificates.
7115 Thanks to Konrad Kraszewski from Google for reporting this issue.
7117 3. Check DSA/ECDSA signatures use DER.
7119 Reencode DSA/ECDSA signatures and compare with the original received
7120 signature. Return an error if there is a mismatch.
7122 This will reject various cases including garbage after signature
7123 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7124 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7125 (negative or with leading zeroes).
7127 Further analysis was conducted and fixes were developed by Stephen Henson
7128 of the OpenSSL core team.
7134 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7136 * Session Ticket Memory Leak.
7138 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7139 integrity of that ticket is first verified. In the event of a session
7140 ticket integrity check failing, OpenSSL will fail to free memory
7141 causing a memory leak. By sending a large number of invalid session
7142 tickets an attacker could exploit this issue in a Denial Of Service
7148 * Build option no-ssl3 is incomplete.
7150 When OpenSSL is configured with "no-ssl3" as a build option, servers
7151 could accept and complete a SSL 3.0 handshake, and clients could be
7152 configured to send them.
7155 *Akamai and the OpenSSL team*
7157 * Add support for TLS_FALLBACK_SCSV.
7158 Client applications doing fallback retries should call
7159 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7162 *Adam Langley, Bodo Moeller*
7164 * Add additional DigestInfo checks.
7166 Reencode DigestInto in DER and check against the original when
7167 verifying RSA signature: this will reject any improperly encoded
7168 DigestInfo structures.
7170 Note: this is a precautionary measure and no attacks are currently known.
7174 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7176 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7177 to a denial of service attack. A malicious server can crash the client
7178 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7179 ciphersuite and sending carefully crafted handshake messages.
7181 Thanks to Felix Gröbert (Google) for discovering and researching this
7187 * By sending carefully crafted DTLS packets an attacker could cause openssl
7188 to leak memory. This can be exploited through a Denial of Service attack.
7189 Thanks to Adam Langley for discovering and researching this issue.
7194 * An attacker can force openssl to consume large amounts of memory whilst
7195 processing DTLS handshake messages. This can be exploited through a
7196 Denial of Service attack.
7197 Thanks to Adam Langley for discovering and researching this issue.
7202 * An attacker can force an error condition which causes openssl to crash
7203 whilst processing DTLS packets due to memory being freed twice. This
7204 can be exploited through a Denial of Service attack.
7205 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7211 * If a multithreaded client connects to a malicious server using a resumed
7212 session and the server sends an ec point format extension it could write
7213 up to 255 bytes to freed memory.
7215 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7221 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7222 X509_name_oneline, X509_name_print_ex et al. to leak some information
7223 from the stack. Applications may be affected if they echo pretty printing
7224 output to the attacker.
7226 Thanks to Ivan Fratric (Google) for discovering this issue.
7229 *Emilia Käsper, and Steve Henson*
7231 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7232 for corner cases. (Certain input points at infinity could lead to
7233 bogus results, with non-infinity inputs mapped to infinity too.)
7237 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7239 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7240 handshake can force the use of weak keying material in OpenSSL
7241 SSL/TLS clients and servers.
7243 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7244 researching this issue. [CVE-2014-0224][]
7246 *KIKUCHI Masashi, Steve Henson*
7248 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7249 OpenSSL DTLS client the code can be made to recurse eventually crashing
7252 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7255 *Imre Rad, Steve Henson*
7257 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7258 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7259 client or server. This is potentially exploitable to run arbitrary
7260 code on a vulnerable client or server.
7262 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7264 *Jüri Aedla, Steve Henson*
7266 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7267 are subject to a denial of service attack.
7269 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7270 this issue. [CVE-2014-3470][]
7272 *Felix Gröbert, Ivan Fratric, Steve Henson*
7274 * Harmonize version and its documentation. -f flag is used to display
7277 *mancha <mancha1@zoho.com>*
7279 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7280 in i2d_ECPrivateKey.
7282 *mancha <mancha1@zoho.com>*
7284 * Fix some double frees. These are not thought to be exploitable.
7286 *mancha <mancha1@zoho.com>*
7288 * Fix for the attack described in the paper "Recovering OpenSSL
7289 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7290 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7291 <http://eprint.iacr.org/2014/140>
7293 Thanks to Yuval Yarom and Naomi Benger for discovering this
7294 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7296 *Yuval Yarom and Naomi Benger*
7298 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7300 * Keep original DTLS digest and encryption contexts in retransmission
7301 structures so we can use the previous session parameters if they need
7302 to be resent. [CVE-2013-6450][]
7306 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7307 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7308 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7309 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7310 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7311 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7313 *Rob Stradling, Adam Langley*
7315 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7317 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7319 This addresses the flaw in CBC record processing discovered by
7320 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7321 at: <http://www.isg.rhul.ac.uk/tls/>
7323 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7324 Security Group at Royal Holloway, University of London
7325 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7326 Emilia Käsper for the initial patch.
7329 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7331 * Return an error when checking OCSP signatures when key is NULL.
7332 This fixes a DoS attack. [CVE-2013-0166][]
7336 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7337 the right response is stapled. Also change SSL_get_certificate()
7338 so it returns the certificate actually sent.
7339 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7340 (This is a backport)
7342 *Rob Stradling <rob.stradling@comodo.com>*
7344 * Fix possible deadlock when decoding public keys.
7348 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7350 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7353 * Sanity check record length before skipping explicit IV in DTLS
7356 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7357 fuzzing as a service testing platform.
7362 * Initialise tkeylen properly when encrypting CMS messages.
7363 Thanks to Solar Designer of Openwall for reporting this issue.
7367 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7369 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7370 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7371 in CRYPTO_realloc_clean.
7373 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7374 issue and to Adam Langley <agl@chromium.org> for fixing it.
7377 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7379 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7381 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7382 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7383 content decryption and always return the same error. Note: this attack
7384 needs on average 2^20 messages so it only affects automated senders. The
7385 old behaviour can be re-enabled in the CMS code by setting the
7386 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7387 an MMA defence is not necessary.
7388 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7389 this issue. [CVE-2012-0884][]
7393 * Fix CVE-2011-4619: make sure we really are receiving a
7394 client hello before rejecting multiple SGC restarts. Thanks to
7395 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7399 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7401 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7402 Thanks to Antonio Martin, Enterprise Secure Access Research and
7403 Development, Cisco Systems, Inc. for discovering this bug and
7404 preparing a fix. [CVE-2012-0050][]
7408 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7410 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7411 of the Vaudenay padding oracle attack on CBC mode encryption
7412 which enables an efficient plaintext recovery attack against
7413 the OpenSSL implementation of DTLS. Their attack exploits timing
7414 differences arising during decryption processing. A research
7415 paper describing this attack can be found at:
7416 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7417 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7418 Security Group at Royal Holloway, University of London
7419 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7420 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7421 for preparing the fix. [CVE-2011-4108][]
7423 *Robin Seggelmann, Michael Tuexen*
7425 * Clear bytes used for block padding of SSL 3.0 records.
7428 *Adam Langley (Google)*
7430 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7431 Kadianakis <desnacked@gmail.com> for discovering this issue and
7432 Adam Langley for preparing the fix. [CVE-2011-4619][]
7434 *Adam Langley (Google)*
7436 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7438 *Andrey Kulikov <amdeich@gmail.com>*
7440 * Prevent malformed RFC3779 data triggering an assertion failure.
7441 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7442 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7444 *Rob Austein <sra@hactrn.net>*
7446 * Improved PRNG seeding for VOS.
7448 *Paul Green <Paul.Green@stratus.com>*
7450 * Fix ssl_ciph.c set-up race.
7452 *Adam Langley (Google)*
7454 * Fix spurious failures in ecdsatest.c.
7456 *Emilia Käsper (Google)*
7458 * Fix the BIO_f_buffer() implementation (which was mixing different
7459 interpretations of the `..._len` fields).
7461 *Adam Langley (Google)*
7463 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7464 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7465 threads won't reuse the same blinding coefficients.
7467 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7468 lock to call BN_BLINDING_invert_ex, and avoids one use of
7469 BN_BLINDING_update for each BN_BLINDING structure (previously,
7470 the last update always remained unused).
7472 *Emilia Käsper (Google)*
7474 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7476 *Bob Buckholz (Google)*
7478 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7480 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7481 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7483 *Kaspar Brand <ossl@velox.ch>*
7485 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7486 for multi-threaded use of ECDH. [CVE-2011-3210][]
7488 *Adam Langley (Google)*
7490 * Fix x509_name_ex_d2i memory leak on bad inputs.
7494 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7495 signature public key algorithm by using OID xref utilities instead.
7496 Before this you could only use some ECC ciphersuites with SHA1 only.
7500 * Add protection against ECDSA timing attacks as mentioned in the paper
7501 by Billy Bob Brumley and Nicola Tuveri, see:
7502 <http://eprint.iacr.org/2011/232.pdf>
7504 *Billy Bob Brumley and Nicola Tuveri*
7506 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7508 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7510 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7512 * Fix bug in string printing code: if *any* escaping is enabled we must
7513 escape the escape character (backslash) or the resulting string is
7518 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7520 * Disable code workaround for ancient and obsolete Netscape browsers
7521 and servers: an attacker can use it in a ciphersuite downgrade attack.
7522 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7526 * Fixed J-PAKE implementation error, originally discovered by
7527 Sebastien Martini, further info and confirmation from Stefan
7528 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7532 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7534 * Fix extension code to avoid race conditions which can result in a buffer
7535 overrun vulnerability: resumed sessions must not be modified as they can
7536 be shared by multiple threads. CVE-2010-3864
7540 * Fix WIN32 build system to correctly link an ENGINE directory into
7545 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7547 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7550 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7552 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7554 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7555 context. The operation can be customised via the ctrl mechanism in
7556 case ENGINEs want to include additional functionality.
7560 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7564 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7565 output hashes compatible with older versions of OpenSSL.
7567 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7569 * Fix compression algorithm handling: if resuming a session use the
7570 compression algorithm of the resumed session instead of determining
7571 it from client hello again. Don't allow server to change algorithm.
7575 * Add load_crls() function to apps tidying load_certs() too. Add option
7576 to verify utility to allow additional CRLs to be included.
7580 * Update OCSP request code to permit adding custom headers to the request:
7581 some responders need this.
7585 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7588 *Julia Lawall <julia@diku.dk>*
7590 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7591 needlessly dereferenced structures, used obsolete functions and
7592 didn't handle all updated verify codes correctly.
7596 * Disable MD2 in the default configuration.
7600 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7601 indicate the initial BIO being pushed or popped. This makes it possible
7602 to determine whether the BIO is the one explicitly called or as a result
7603 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7604 it handles reference counts correctly and doesn't zero out the I/O bio
7605 when it is not being explicitly popped. WARNING: applications which
7606 included workarounds for the old buggy behaviour will need to be modified
7607 or they could free up already freed BIOs.
7611 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7612 renaming to all platforms (within the 0.9.8 branch, this was
7613 done conditionally on Netware platforms to avoid a name clash).
7615 *Guenter <lists@gknw.net>*
7617 * Add ECDHE and PSK support to DTLS.
7619 *Michael Tuexen <tuexen@fh-muenster.de>*
7621 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7626 * Add "missing" function EVP_MD_flags() (without this the only way to
7627 retrieve a digest flags is by accessing the structure directly. Update
7628 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7629 or cipher is registered as in the "from" argument. Print out all
7630 registered digests in the dgst usage message instead of manually
7631 attempting to work them out.
7635 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7636 this allows the use of compression and extensions. Change default cipher
7637 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7638 by default unless an application cipher string requests it.
7642 * Alter match criteria in PKCS12_parse(). It used to try to use local
7643 key ids to find matching certificates and keys but some PKCS#12 files
7644 don't follow the (somewhat unwritten) rules and this strategy fails.
7645 Now just gather all certificates together and the first private key
7646 then look for the first certificate that matches the key.
7650 * Support use of registered digest and cipher names for dgst and cipher
7651 commands instead of having to add each one as a special case. So now
7658 openssl dgst -sha256 foo
7660 and this works for ENGINE based algorithms too.
7664 * Update Gost ENGINE to support parameter files.
7666 *Victor B. Wagner <vitus@cryptocom.ru>*
7668 * Support GeneralizedTime in ca utility.
7670 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7672 * Enhance the hash format used for certificate directory links. The new
7673 form uses the canonical encoding (meaning equivalent names will work
7674 even if they aren't identical) and uses SHA1 instead of MD5. This form
7675 is incompatible with the older format and as a result c_rehash should
7676 be used to rebuild symbolic links.
7680 * Make PKCS#8 the default write format for private keys, replacing the
7681 traditional format. This form is standardised, more secure and doesn't
7682 include an implicit MD5 dependency.
7686 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7687 committed to OpenSSL should pass this lot as a minimum.
7691 * Add session ticket override functionality for use by EAP-FAST.
7693 *Jouni Malinen <j@w1.fi>*
7695 * Modify HMAC functions to return a value. Since these can be implemented
7696 in an ENGINE errors can occur.
7700 * Type-checked OBJ_bsearch_ex.
7704 * Type-checked OBJ_bsearch. Also some constification necessitated
7705 by type-checking. Still to come: TXT_DB, bsearch(?),
7706 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7711 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7712 seconds to a tm structure directly, instead of going through OS
7713 specific date routines. This avoids any issues with OS routines such
7714 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7715 and X509_time_adj_ex() to cover the extended range. The existing
7716 X509_time_adj() is still usable and will no longer have any date issues.
7720 * Delta CRL support. New use deltas option which will attempt to locate
7721 and search any appropriate delta CRLs available.
7723 This work was sponsored by Google.
7727 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7728 code and add additional score elements. Validate alternate CRL paths
7729 as part of the CRL checking and indicate a new error "CRL path validation
7730 error" in this case. Applications wanting additional details can use
7731 the verify callback and check the new "parent" field. If this is not
7732 NULL CRL path validation is taking place. Existing applications won't
7733 see this because it requires extended CRL support which is off by
7736 This work was sponsored by Google.
7740 * Support for freshest CRL extension.
7742 This work was sponsored by Google.
7746 * Initial indirect CRL support. Currently only supported in the CRLs
7747 passed directly and not via lookup. Process certificate issuer
7748 CRL entry extension and lookup CRL entries by bother issuer name
7749 and serial number. Check and process CRL issuer entry in IDP extension.
7751 This work was sponsored by Google.
7755 * Add support for distinct certificate and CRL paths. The CRL issuer
7756 certificate is validated separately in this case. Only enabled if
7757 an extended CRL support flag is set: this flag will enable additional
7758 CRL functionality in future.
7760 This work was sponsored by Google.
7764 * Add support for policy mappings extension.
7766 This work was sponsored by Google.
7770 * Fixes to pathlength constraint, self issued certificate handling,
7771 policy processing to align with RFC3280 and PKITS tests.
7773 This work was sponsored by Google.
7777 * Support for name constraints certificate extension. DN, email, DNS
7778 and URI types are currently supported.
7780 This work was sponsored by Google.
7784 * To cater for systems that provide a pointer-based thread ID rather
7785 than numeric, deprecate the current numeric thread ID mechanism and
7786 replace it with a structure and associated callback type. This
7787 mechanism allows a numeric "hash" to be extracted from a thread ID in
7788 either case, and on platforms where pointers are larger than 'long',
7789 mixing is done to help ensure the numeric 'hash' is usable even if it
7790 can't be guaranteed unique. The default mechanism is to use "&errno"
7791 as a pointer-based thread ID to distinguish between threads.
7793 Applications that want to provide their own thread IDs should now use
7794 CRYPTO_THREADID_set_callback() to register a callback that will call
7795 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7797 Note that ERR_remove_state() is now deprecated, because it is tied
7798 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7799 to free the current thread's error state should be replaced by
7800 ERR_remove_thread_state(NULL).
7802 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7803 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7804 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7805 application was previously providing a numeric thread callback that
7806 was inappropriate for distinguishing threads, then uniqueness might
7807 have been obtained with &errno that happened immediately in the
7808 intermediate development versions of OpenSSL; this is no longer the
7809 case, the numeric thread callback will now override the automatic use
7812 *Geoff Thorpe, with help from Bodo Moeller*
7814 * Initial support for different CRL issuing certificates. This covers a
7815 simple case where the self issued certificates in the chain exist and
7816 the real CRL issuer is higher in the existing chain.
7818 This work was sponsored by Google.
7822 * Removed effectively defunct crypto/store from the build.
7826 * Revamp of STACK to provide stronger type-checking. Still to come:
7827 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7828 ASN1_STRING, CONF_VALUE.
7832 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7833 RAM on SSL connections. This option can save about 34k per idle SSL.
7837 * Revamp of LHASH to provide stronger type-checking. Still to come:
7838 STACK, TXT_DB, bsearch, qsort.
7842 * Initial support for Cryptographic Message Syntax (aka CMS) based
7843 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7844 support for data, signedData, compressedData, digestedData and
7845 encryptedData, envelopedData types included. Scripts to check against
7846 RFC4134 examples draft and interop and consistency checks of many
7847 content types and variants.
7851 * Add options to enc utility to support use of zlib compression BIO.
7855 * Extend mk1mf to support importing of options and assembly language
7856 files from Configure script, currently only included in VC-WIN32.
7857 The assembly language rules can now optionally generate the source
7858 files from the associated perl scripts.
7862 * Implement remaining functionality needed to support GOST ciphersuites.
7863 Interop testing has been performed using CryptoPro implementations.
7865 *Victor B. Wagner <vitus@cryptocom.ru>*
7867 * s390x assembler pack.
7871 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7876 * Implement Opaque PRF Input TLS extension as specified in
7877 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7878 official specification yet and no extension type assignment by
7879 IANA exists, this extension (for now) will have to be explicitly
7880 enabled when building OpenSSL by providing the extension number
7881 to use. For example, specify an option
7883 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7885 to the "config" or "Configure" script to enable the extension,
7886 assuming extension number 0x9527 (which is a completely arbitrary
7887 and unofficial assignment based on the MD5 hash of the Internet
7888 Draft). Note that by doing so, you potentially lose
7889 interoperability with other TLS implementations since these might
7890 be using the same extension number for other purposes.
7892 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7893 opaque PRF input value to use in the handshake. This will create
7894 an internal copy of the length-'len' string at 'src', and will
7895 return non-zero for success.
7897 To get more control and flexibility, provide a callback function
7900 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7901 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7905 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7908 Callback function 'cb' will be called in handshakes, and is
7909 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7910 Argument 'arg' is for application purposes (the value as given to
7911 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7912 be provided to the callback function). The callback function
7913 has to return non-zero to report success: usually 1 to use opaque
7914 PRF input just if possible, or 2 to enforce use of the opaque PRF
7915 input. In the latter case, the library will abort the handshake
7916 if opaque PRF input is not successfully negotiated.
7918 Arguments 'peerinput' and 'len' given to the callback function
7919 will always be NULL and 0 in the case of a client. A server will
7920 see the client's opaque PRF input through these variables if
7921 available (NULL and 0 otherwise). Note that if the server
7922 provides an opaque PRF input, the length must be the same as the
7923 length of the client's opaque PRF input.
7925 Note that the callback function will only be called when creating
7926 a new session (session resumption can resume whatever was
7927 previously negotiated), and will not be called in SSL 2.0
7928 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7929 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7930 for applications that need to enforce opaque PRF input.
7934 * Update ssl code to support digests other than SHA1+MD5 for handshake
7937 *Victor B. Wagner <vitus@cryptocom.ru>*
7939 * Add RFC4507 support to OpenSSL. This includes the corrections in
7940 RFC4507bis. The encrypted ticket format is an encrypted encoded
7941 SSL_SESSION structure, that way new session features are automatically
7944 If a client application caches session in an SSL_SESSION structure
7945 support is transparent because tickets are now stored in the encoded
7948 The SSL_CTX structure automatically generates keys for ticket
7949 protection in servers so again support should be possible
7950 with no application modification.
7952 If a client or server wishes to disable RFC4507 support then the option
7953 SSL_OP_NO_TICKET can be set.
7955 Add a TLS extension debugging callback to allow the contents of any client
7956 or server extensions to be examined.
7958 This work was sponsored by Google.
7962 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7963 OpenSSL should now compile cleanly on gcc 4.2
7965 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7967 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7968 support including streaming MAC support: this is required for GOST
7969 ciphersuite support.
7971 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7973 * Add option -stream to use PKCS#7 streaming in smime utility. New
7974 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7975 to output in BER and PEM format.
7979 * Experimental support for use of HMAC via EVP_PKEY interface. This
7980 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
7981 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7982 ENGINE support for HMAC keys which are unextractable. New -mac and
7983 -macopt options to dgst utility.
7987 * New option -sigopt to dgst utility. Update dgst to use
7988 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
7989 alternative signing parameters such as X9.31 or PSS in the dgst
7994 * Change ssl_cipher_apply_rule(), the internal function that does
7995 the work each time a ciphersuite string requests enabling
7996 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7997 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7998 the order of disabled ciphersuites such that those ciphersuites
7999 that most recently went from enabled to disabled not only stay
8000 in order with respect to each other, but also have higher priority
8001 than other disabled ciphersuites the next time ciphersuites are
8004 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8005 the same ciphersuites as with "HIGH" alone, but in a specific
8006 order where the PSK ciphersuites come first (since they are the
8007 most recently disabled ciphersuites when "HIGH" is parsed).
8009 Also, change ssl_create_cipher_list() (using this new
8010 functionality) such that between otherwise identical
8011 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8016 * Change ssl_create_cipher_list() so that it automatically
8017 arranges the ciphersuites in reasonable order before starting
8018 to process the rule string. Thus, the definition for "DEFAULT"
8019 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8020 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8021 This makes it much easier to arrive at a reasonable default order
8022 in applications for which anonymous ciphers are OK (meaning
8023 that you can't actually use DEFAULT).
8025 *Bodo Moeller; suggested by Victor Duchovni*
8027 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8028 processing) into multiple integers instead of setting
8029 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8030 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8031 (These masks as well as the individual bit definitions are hidden
8032 away into the non-exported interface ssl/ssl_locl.h, so this
8033 change to the definition of the SSL_CIPHER structure shouldn't
8034 affect applications.) This give us more bits for each of these
8035 categories, so there is no longer a need to coagulate AES128 and
8036 AES256 into a single algorithm bit, and to coagulate Camellia128
8037 and Camellia256 into a single algorithm bit, which has led to all
8040 Thus, among other things, the kludge introduced in 0.9.7m and
8041 0.9.8e for masking out AES256 independently of AES128 or masking
8042 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8044 With the change, we also introduce new ciphersuite aliases that
8045 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8050 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8051 Use the leftmost N bytes of the signature input if the input is
8052 larger than the prime q (with N being the size in bytes of q).
8056 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8057 it yet and it is largely untested.
8061 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8065 * Initial incomplete changes to avoid need for function casts in OpenSSL
8066 some compilers (gcc 4.2 and later) reject their use. Safestack is
8067 reimplemented. Update ASN1 to avoid use of legacy functions.
8071 * Win32/64 targets are linked with Winsock2.
8075 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8076 to external functions. This can be used to increase CRL handling
8077 efficiency especially when CRLs are very large by (for example) storing
8078 the CRL revoked certificates in a database.
8082 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8083 new CRLs added to a directory can be used. New command line option
8084 -verify_return_error to s_client and s_server. This causes real errors
8085 to be returned by the verify callback instead of carrying on no matter
8086 what. This reflects the way a "real world" verify callback would behave.
8090 * GOST engine, supporting several GOST algorithms and public key formats.
8091 Kindly donated by Cryptocom.
8095 * Partial support for Issuing Distribution Point CRL extension. CRLs
8096 partitioned by DP are handled but no indirect CRL or reason partitioning
8097 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8098 selected via a scoring technique which handles IDP and AKID in CRLs.
8102 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8103 will ultimately be used for all verify operations: this will remove the
8104 X509_STORE dependency on certificate verification and allow alternative
8105 lookup methods. X509_STORE based implementations of these two callbacks.
8109 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8110 Modify get_crl() to find a valid (unexpired) CRL if possible.
8114 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8115 this would be called X509_CRL_cmp() but that name is already used by
8116 a function that just compares CRL issuer names. Cache several CRL
8117 extensions in X509_CRL structure and cache CRLDP in X509.
8121 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8122 this maps equivalent X509_NAME structures into a consistent structure.
8123 Name comparison can then be performed rapidly using memcmp().
8127 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8132 * Allow digests to supply their own micalg string for S/MIME type using
8133 the ctrl EVP_MD_CTRL_MICALG.
8137 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8138 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8139 ctrl. It can then customise the structure before and/or after signing
8144 * New function OBJ_add_sigid() to allow application defined signature OIDs
8145 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8146 to free up any added signature OIDs.
8150 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8151 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8152 digest and cipher tables. New options added to openssl utility:
8153 list-message-digest-algorithms and list-cipher-algorithms.
8157 * Change the array representation of binary polynomials: the list
8158 of degrees of non-zero coefficients is now terminated with -1.
8159 Previously it was terminated with 0, which was also part of the
8160 value; thus, the array representation was not applicable to
8161 polynomials where t^0 has coefficient zero. This change makes
8162 the array representation useful in a more general context.
8166 * Various modifications and fixes to SSL/TLS cipher string
8167 handling. For ECC, the code now distinguishes between fixed ECDH
8168 with RSA certificates on the one hand and with ECDSA certificates
8169 on the other hand, since these are separate ciphersuites. The
8170 unused code for Fortezza ciphersuites has been removed.
8172 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8173 (not "ECDHE"). For consistency with the code for DH
8174 certificates, use of ECDH certificates is now considered ECDH
8175 authentication, not RSA or ECDSA authentication (the latter is
8176 merely the CA's signing algorithm and not actively used in the
8179 The temporary ciphersuite alias "ECCdraft" is no longer
8180 available, and ECC ciphersuites are no longer excluded from "ALL"
8181 and "DEFAULT". The following aliases now exist for RFC 4492
8182 ciphersuites, most of these by analogy with the DH case:
8184 kECDHr - ECDH cert, signed with RSA
8185 kECDHe - ECDH cert, signed with ECDSA
8186 kECDH - ECDH cert (signed with either RSA or ECDSA)
8187 kEECDH - ephemeral ECDH
8188 ECDH - ECDH cert or ephemeral ECDH
8194 AECDH - anonymous ECDH
8195 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8199 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8200 Use correct micalg parameters depending on digest(s) in signed message.
8204 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8205 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8209 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8210 an engine to register a method. Add ENGINE lookups for methods and
8211 functional reference processing.
8215 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8216 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8221 * New -resign option to smime utility. This adds one or more signers
8222 to an existing PKCS#7 signedData structure. Also -md option to use an
8223 alternative message digest algorithm for signing.
8227 * Tidy up PKCS#7 routines and add new functions to make it easier to
8228 create PKCS7 structures containing multiple signers. Update smime
8229 application to support multiple signers.
8233 * New -macalg option to pkcs12 utility to allow setting of an alternative
8238 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8239 Reorganize PBE internals to lookup from a static table using NIDs,
8240 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8241 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8242 PRF which will be automatically used with PBES2.
8246 * Replace the algorithm specific calls to generate keys in "req" with the
8251 * Update PKCS#7 enveloped data routines to use new API. This is now
8252 supported by any public key method supporting the encrypt operation. A
8253 ctrl is added to allow the public key algorithm to examine or modify
8254 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8259 * Add a ctrl to asn1 method to allow a public key algorithm to express
8260 a default digest type to use. In most cases this will be SHA1 but some
8261 algorithms (such as GOST) need to specify an alternative digest. The
8262 return value indicates how strong the preference is 1 means optional and
8263 2 is mandatory (that is it is the only supported type). Modify
8264 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8265 use the default md. Update openssl utilities to use the default digest
8266 type for signing if it is not explicitly indicated.
8270 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8271 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8272 signing method from the key type. This effectively removes the link
8273 between digests and public key types.
8277 * Add an OID cross reference table and utility functions. Its purpose is to
8278 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8279 rsaEncryption. This will allow some of the algorithm specific hackery
8280 needed to use the correct OID to be removed.
8284 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8285 structures for PKCS7_sign(). They are now set up by the relevant public
8290 * Add provisional EC pkey method with support for ECDSA and ECDH.
8294 * Add support for key derivation (agreement) in the API, DH method and
8299 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8300 public and private key formats. As a side effect these add additional
8301 command line functionality not previously available: DSA signatures can be
8302 generated and verified using pkeyutl and DH key support and generation in
8309 *Oliver Tappe <zooey@hirschkaefer.de>*
8311 * New make target "install_html_docs" installs HTML renditions of the
8314 *Oliver Tappe <zooey@hirschkaefer.de>*
8316 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8317 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8318 support key and parameter generation and add initial key generation
8319 functionality for RSA.
8323 * Add functions for main EVP_PKEY_method operations. The undocumented
8324 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8325 `EVP_PKEY_{encrypt,decrypt}_old`.
8329 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8330 key API, doesn't do much yet.
8334 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8335 public key algorithms. New option to openssl utility:
8336 "list-public-key-algorithms" to print out info.
8340 * Implement the Supported Elliptic Curves Extension for
8341 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8345 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8346 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8350 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8351 utilities such as rsa, dsa, dsaparam etc except they process any key
8356 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8357 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8358 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8363 * Initial support for pluggable public key ASN1.
8364 De-spaghettify the public key ASN1 handling. Move public and private
8365 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8366 algorithm specific handling to a single module within the relevant
8367 algorithm directory. Add functions to allow (near) opaque processing
8368 of public and private key structures.
8372 * Implement the Supported Point Formats Extension for
8373 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8377 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8378 for the psk identity [hint] and the psk callback functions to the
8379 SSL_SESSION, SSL and SSL_CTX structure.
8382 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8386 SSL_CTX_use_psk_identity_hint
8387 SSL_get_psk_identity_hint
8388 SSL_get_psk_identity
8389 SSL_use_psk_identity_hint
8391 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8393 * Add RFC 3161 compliant time stamp request creation, response generation
8394 and response verification functionality.
8396 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8398 * Add initial support for TLS extensions, specifically for the server_name
8399 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8400 have new members for a host name. The SSL data structure has an
8401 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8402 stored in that context to allow for session resumption, even after the
8403 SSL has been switched to a new SSL_CTX in reaction to a client's
8404 server_name extension.
8406 New functions (subject to change):
8408 SSL_get_servername()
8409 SSL_get_servername_type()
8412 New CTRL codes and macros (subject to change):
8414 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8415 - SSL_CTX_set_tlsext_servername_callback()
8416 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8417 - SSL_CTX_set_tlsext_servername_arg()
8418 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8420 openssl s_client has a new '-servername ...' option.
8422 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8423 '-key2 ...', '-servername_fatal' (subject to change). This allows
8424 testing the HostName extension for a specific single host name ('-cert'
8425 and '-key' remain fallbacks for handshakes without HostName
8426 negotiation). If the unrecognized_name alert has to be sent, this by
8427 default is a warning; it becomes fatal with the '-servername_fatal'
8430 *Peter Sylvester, Remy Allais, Christophe Renou*
8432 * Whirlpool hash implementation is added.
8436 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8437 bn(64,32). Because of instruction set limitations it doesn't have
8438 any negative impact on performance. This was done mostly in order
8439 to make it possible to share assembler modules, such as bn_mul_mont
8440 implementations, between 32- and 64-bit builds without hassle.
8444 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8445 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8450 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8451 dedicated Montgomery multiplication procedure, is introduced.
8452 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8453 "64-bit" performance on certain 32-bit targets.
8457 * New option SSL_OP_NO_COMP to disable use of compression selectively
8458 in SSL structures. New SSL ctrl to set maximum send fragment size.
8459 Save memory by setting the I/O buffer sizes dynamically instead of
8460 using the maximum available value.
8464 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8465 in addition to the text details.
8469 * Very, very preliminary EXPERIMENTAL support for printing of general
8470 ASN1 structures. This currently produces rather ugly output and doesn't
8471 handle several customised structures at all.
8475 * Integrated support for PVK file format and some related formats such
8476 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8477 these in the 'rsa' and 'dsa' utilities.
8481 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8485 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8486 place for the (very old) "NETSCAPE" format certificates which are now
8487 handled using new ASN1 code equivalents.
8491 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8492 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8493 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8497 * Modify CRL distribution points extension code to print out previously
8498 unsupported fields. Enhance extension setting code to allow setting of
8503 * Add print and set support for Issuing Distribution Point CRL extension.
8507 * Change 'Configure' script to enable Camellia by default.
8514 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8516 * When rejecting SSL/TLS records due to an incorrect version number, never
8517 update s->server with a new major version number. As of
8518 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8519 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8520 the previous behavior could result in a read attempt at NULL when
8521 receiving specific incorrect SSL/TLS records once record payload
8522 protection is active. [CVE-2010-0740][]
8524 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8526 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8527 could be crashed if the relevant tables were not present (e.g. chrooted).
8529 *Tomas Hoger <thoger@redhat.com>*
8531 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8533 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8535 *Martin Olsson, Neel Mehta*
8537 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8538 accommodate for stack sorting, always a write lock!).
8542 * On some versions of WIN32 Heap32Next is very slow. This can cause
8543 excessive delays in the RAND_poll(): over a minute. As a workaround
8544 include a time check in the inner Heap32Next loop too.
8548 * The code that handled flushing of data in SSL/TLS originally used the
8549 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8550 the problem outlined in PR#1949. The fix suggested there however can
8551 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8552 of Apache). So instead simplify the code to flush unconditionally.
8553 This should be fine since flushing with no data to flush is a no op.
8557 * Handle TLS versions 2.0 and later properly and correctly use the
8558 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8559 off ancient servers have a habit of sticking around for a while...
8563 * Modify compression code so it frees up structures without using the
8564 ex_data callbacks. This works around a problem where some applications
8565 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8566 restarting) then use compression (e.g. SSL with compression) later.
8567 This results in significant per-connection memory leaks and
8568 has caused some security issues including CVE-2008-1678 and
8573 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8574 change when encrypting or decrypting.
8578 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8579 connect and renegotiate with servers which do not support RI.
8580 Until RI is more widely deployed this option is enabled by default.
8584 * Add "missing" ssl ctrls to clear options and mode.
8588 * If client attempts to renegotiate and doesn't support RI respond with
8589 a no_renegotiation alert as required by RFC5746. Some renegotiating
8590 TLS clients will continue a connection gracefully when they receive
8591 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8592 waiting for a server hello which it will never receive. Now we treat a
8593 received no_renegotiation alert as a fatal error. This is because
8594 applications requesting a renegotiation might well expect it to succeed
8595 and would have no code in place to handle the server denying it so the
8596 only safe thing to do is to terminate the connection.
8600 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8601 peer supports secure renegotiation and 0 otherwise. Print out peer
8602 renegotiation support in s_client/s_server.
8606 * Replace the highly broken and deprecated SPKAC certification method with
8607 the updated NID creation version. This should correctly handle UTF8.
8611 * Implement RFC5746. Re-enable renegotiation but require the extension
8612 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8613 turns out to be a bad idea. It has been replaced by
8614 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8615 SSL_CTX_set_options(). This is really not recommended unless you
8616 know what you are doing.
8618 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8620 * Fixes to stateless session resumption handling. Use initial_ctx when
8621 issuing and attempting to decrypt tickets in case it has changed during
8622 servername handling. Use a non-zero length session ID when attempting
8623 stateless session resumption: this makes it possible to determine if
8624 a resumption has occurred immediately after receiving server hello
8625 (several places in OpenSSL subtly assume this) instead of later in
8630 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8631 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8632 fixes for a few places where the return code is not checked
8635 *Julia Lawall <julia@diku.dk>*
8637 * Add --strict-warnings option to Configure script to include devteam
8638 warnings in other configurations.
8642 * Add support for --libdir option and LIBDIR variable in makefiles. This
8643 makes it possible to install openssl libraries in locations which
8644 have names other than "lib", for example "/usr/lib64" which some
8647 *Steve Henson, based on patch from Jeremy Utley*
8649 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8650 X690 8.9.12 and can produce some misleading textual output of OIDs.
8652 *Steve Henson, reported by Dan Kaminsky*
8654 * Delete MD2 from algorithm tables. This follows the recommendation in
8655 several standards that it is not used in new applications due to
8656 several cryptographic weaknesses. For binary compatibility reasons
8657 the MD2 API is still compiled in by default.
8661 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8666 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8667 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8670 *Guenter <lists@gknw.net>*
8672 * Fix the server certificate chain building code to use X509_verify_cert(),
8673 it used to have an ad-hoc builder which was unable to cope with anything
8674 other than a simple chain.
8676 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8678 * Don't check self signed certificate signatures in X509_verify_cert()
8679 by default (a flag can override this): it just wastes time without
8680 adding any security. As a useful side effect self signed root CAs
8681 with non-FIPS digests are now usable in FIPS mode.
8685 * In dtls1_process_out_of_seq_message() the check if the current message
8686 is already buffered was missing. For every new message was memory
8687 allocated, allowing an attacker to perform an denial of service attack
8688 with sending out of seq handshake messages until there is no memory
8689 left. Additionally every future message was buffered, even if the
8690 sequence number made no sense and would be part of another handshake.
8691 So only messages with sequence numbers less than 10 in advance will be
8692 buffered. [CVE-2009-1378][]
8694 *Robin Seggelmann, discovered by Daniel Mentz*
8696 * Records are buffered if they arrive with a future epoch to be
8697 processed after finishing the corresponding handshake. There is
8698 currently no limitation to this buffer allowing an attacker to perform
8699 a DOS attack with sending records with future epochs until there is no
8700 memory left. This patch adds the pqueue_size() function to determine
8701 the size of a buffer and limits the record buffer to 100 entries.
8704 *Robin Seggelmann, discovered by Daniel Mentz*
8706 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8707 parent structure is freed. [CVE-2009-1379][]
8711 * Handle non-blocking I/O properly in SSL_shutdown() call.
8713 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8715 * Add `2.5.4.*` OIDs
8717 *Ilya O. <vrghost@gmail.com>*
8719 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8721 * Disable renegotiation completely - this fixes a severe security
8722 problem [CVE-2009-3555][] at the cost of breaking all
8723 renegotiation. Renegotiation can be re-enabled by setting
8724 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8725 run-time. This is really not recommended unless you know what
8730 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8732 * Don't set val to NULL when freeing up structures, it is freed up by
8733 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8734 zeroing past the valid field. [CVE-2009-0789][]
8736 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8738 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8739 checked correctly. This would allow some invalid signed attributes to
8740 appear to verify correctly. [CVE-2009-0591][]
8742 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8744 * Reject UniversalString and BMPString types with invalid lengths. This
8745 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8746 a legal length. [CVE-2009-0590][]
8750 * Set S/MIME signing as the default purpose rather than setting it
8751 unconditionally. This allows applications to override it at the store
8756 * Permit restricted recursion of ASN1 strings. This is needed in practice
8757 to handle some structures.
8761 * Improve efficiency of mem_gets: don't search whole buffer each time
8764 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8766 * New -hex option for openssl rand.
8770 * Print out UTF8String and NumericString when parsing ASN1.
8774 * Support NumericString type for name components.
8778 * Allow CC in the environment to override the automatically chosen
8779 compiler. Note that nothing is done to ensure flags work with the
8784 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8786 * Properly check EVP_VerifyFinal() and similar return values
8789 *Ben Laurie, Bodo Moeller, Google Security Team*
8791 * Enable TLS extensions by default.
8795 * Allow the CHIL engine to be loaded, whether the application is
8796 multithreaded or not. (This does not release the developer from the
8797 obligation to set up the dynamic locking callbacks.)
8799 *Sander Temme <sander@temme.net>*
8801 * Use correct exit code if there is an error in dgst command.
8803 *Steve Henson; problem pointed out by Roland Dirlewanger*
8805 * Tweak Configure so that you need to say "experimental-jpake" to enable
8806 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8810 * Add experimental JPAKE support, including demo authentication in
8811 s_client and s_server.
8815 * Set the comparison function in v3_addr_canonize().
8817 *Rob Austein <sra@hactrn.net>*
8819 * Add support for XMPP STARTTLS in s_client.
8821 *Philip Paeps <philip@freebsd.org>*
8823 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8824 to ensure that even with this option, only ciphersuites in the
8825 server's preference list will be accepted. (Note that the option
8826 applies only when resuming a session, so the earlier behavior was
8827 just about the algorithm choice for symmetric cryptography.)
8831 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8833 * Fix NULL pointer dereference if a DTLS server received
8834 ChangeCipherSpec as first record [CVE-2009-1386][].
8838 * Fix a state transition in s3_srvr.c and d1_srvr.c
8839 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8843 * The fix in 0.9.8c that supposedly got rid of unsafe
8844 double-checked locking was incomplete for RSA blinding,
8845 addressing just one layer of what turns out to have been
8846 doubly unsafe triple-checked locking.
8848 So now fix this for real by retiring the MONT_HELPER macro
8849 in crypto/rsa/rsa_eay.c.
8851 *Bodo Moeller; problem pointed out by Marius Schilder*
8853 * Various precautionary measures:
8855 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8857 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8858 (NB: This would require knowledge of the secret session ticket key
8859 to exploit, in which case you'd be SOL either way.)
8861 - Change bn_nist.c so that it will properly handle input BIGNUMs
8862 outside the expected range.
8864 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8867 *Neel Mehta, Bodo Moeller*
8869 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8870 the load fails. Useful for distros.
8872 *Ben Laurie and the FreeBSD team*
8874 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8878 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8882 * Expand ENGINE to support engine supplied SSL client certificate functions.
8884 This work was sponsored by Logica.
8888 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8889 keystores. Support for SSL/TLS client authentication too.
8890 Not compiled unless enable-capieng specified to Configure.
8892 This work was sponsored by Logica.
8896 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8897 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8898 attribute creation routines such as certificate requests and PKCS#12
8903 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
8905 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8906 handshake which could lead to a client crash as found using the
8907 Codenomicon TLS test suite [CVE-2008-1672][]
8909 *Steve Henson, Mark Cox*
8911 * Fix double free in TLS server name extensions which could lead to
8912 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8916 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8918 Clear the error queue to ensure that error entries left from
8919 older function calls do not interfere with the correct operation.
8921 *Lutz Jaenicke, Erik de Castro Lopo*
8923 * Remove root CA certificates of commercial CAs:
8925 The OpenSSL project does not recommend any specific CA and does not
8926 have any policy with respect to including or excluding any CA.
8927 Therefore it does not make any sense to ship an arbitrary selection
8928 of root CA certificates with the OpenSSL software.
8932 * RSA OAEP patches to fix two separate invalid memory reads.
8933 The first one involves inputs when 'lzero' is greater than
8934 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8935 before the beginning of from). The second one involves inputs where
8936 the 'db' section contains nothing but zeroes (there is a one-byte
8937 invalid read after the end of 'db').
8939 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8941 * Partial backport from 0.9.9-dev:
8943 Introduce bn_mul_mont (dedicated Montgomery multiplication
8944 procedure) as a candidate for BIGNUM assembler implementation.
8945 While 0.9.9-dev uses assembler for various architectures, only
8946 x86_64 is available by default here in the 0.9.8 branch, and
8947 32-bit x86 is available through a compile-time setting.
8949 To try the 32-bit x86 assembler implementation, use Configure
8950 option "enable-montasm" (which exists only for this backport).
8952 As "enable-montasm" for 32-bit x86 disclaims code stability
8953 anyway, in this constellation we activate additional code
8954 backported from 0.9.9-dev for further performance improvements,
8955 namely BN_from_montgomery_word. (To enable this otherwise,
8956 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8958 *Andy Polyakov (backport partially by Bodo Moeller)*
8960 * Add TLS session ticket callback. This allows an application to set
8961 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8962 values. This is useful for key rollover for example where several key
8963 sets may exist with different names.
8967 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8968 This was broken until now in 0.9.8 releases, such that the only way
8969 a registered ENGINE could be used (assuming it initialises
8970 successfully on the host) was to explicitly set it as the default
8971 for the relevant algorithms. This is in contradiction with 0.9.7
8972 behaviour and the documentation. With this fix, when an ENGINE is
8973 registered into a given algorithm's table of implementations, the
8974 'uptodate' flag is reset so that auto-discovery will be used next
8975 time a new context for that algorithm attempts to select an
8978 *Ian Lister (tweaked by Geoff Thorpe)*
8980 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8981 implementation in the following ways:
8983 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8986 Lack of BER streaming support means one pass streaming processing is
8987 only supported if data is detached: setting the streaming flag is
8988 ignored for embedded content.
8990 CMS support is disabled by default and must be explicitly enabled
8991 with the enable-cms configuration option.
8995 * Update the GMP engine glue to do direct copies between BIGNUM and
8996 mpz_t when openssl and GMP use the same limb size. Otherwise the
8997 existing "conversion via a text string export" trick is still used.
8999 *Paul Sheer <paulsheer@gmail.com>*
9001 * Zlib compression BIO. This is a filter BIO which compressed and
9002 uncompresses any data passed through it.
9006 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9007 RFC3394 compatible AES key wrapping.
9011 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9012 sets string data without copying. X509_ALGOR_set0() and
9013 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9014 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9015 from an X509_ATTRIBUTE structure optionally checking it occurs only
9016 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9021 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9022 to get the expected BN_FLG_CONSTTIME behavior.
9024 *Bodo Moeller (Google)*
9028 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9029 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9030 - added some more tests to do_tests.pl
9031 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9032 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9033 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9034 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9035 - various changes to netware.pl to enable gcc-cross builds on Win32
9037 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9038 - various changes to fix missing prototype warnings
9039 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9040 - added AES, WHIRLPOOL and CPUID assembler code to build files
9041 - added missing AES assembler make rules to mk1mf.pl
9042 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9044 *Guenter Knauf <eflash@gmx.net>*
9046 * Implement certificate status request TLS extension defined in RFC3546.
9047 A client can set the appropriate parameters and receive the encoded
9048 OCSP response via a callback. A server can query the supplied parameters
9049 and set the encoded OCSP response in the callback. Add simplified examples
9050 to s_client and s_server.
9054 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9057 + Binary incompatibility of ssl_ctx_st structure
9058 + DTLS interoperation with non-compliant servers
9059 + Don't call get_session_cb() without proposed session
9060 + Fix ia64 assembler code
9062 *Andy Polyakov, Steve Henson*
9064 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9066 * DTLS Handshake overhaul. There were longstanding issues with
9067 OpenSSL DTLS implementation, which were making it impossible for
9068 RFC 4347 compliant client to communicate with OpenSSL server.
9069 Unfortunately just fixing these incompatibilities would "cut off"
9070 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9071 server keeps tolerating non RFC compliant syntax. The opposite is
9072 not true, 0.9.8f client can not communicate with earlier server.
9073 This update even addresses CVE-2007-4995.
9077 * Changes to avoid need for function casts in OpenSSL: some compilers
9078 (gcc 4.2 and later) reject their use.
9079 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9082 * Add RFC4507 support to OpenSSL. This includes the corrections in
9083 RFC4507bis. The encrypted ticket format is an encrypted encoded
9084 SSL_SESSION structure, that way new session features are automatically
9087 If a client application caches session in an SSL_SESSION structure
9088 support is transparent because tickets are now stored in the encoded
9091 The SSL_CTX structure automatically generates keys for ticket
9092 protection in servers so again support should be possible
9093 with no application modification.
9095 If a client or server wishes to disable RFC4507 support then the option
9096 SSL_OP_NO_TICKET can be set.
9098 Add a TLS extension debugging callback to allow the contents of any client
9099 or server extensions to be examined.
9101 This work was sponsored by Google.
9105 * Add initial support for TLS extensions, specifically for the server_name
9106 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9107 have new members for a host name. The SSL data structure has an
9108 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9109 stored in that context to allow for session resumption, even after the
9110 SSL has been switched to a new SSL_CTX in reaction to a client's
9111 server_name extension.
9113 New functions (subject to change):
9115 SSL_get_servername()
9116 SSL_get_servername_type()
9119 New CTRL codes and macros (subject to change):
9121 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9122 - SSL_CTX_set_tlsext_servername_callback()
9123 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9124 - SSL_CTX_set_tlsext_servername_arg()
9125 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9127 openssl s_client has a new '-servername ...' option.
9129 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9130 '-key2 ...', '-servername_fatal' (subject to change). This allows
9131 testing the HostName extension for a specific single host name ('-cert'
9132 and '-key' remain fallbacks for handshakes without HostName
9133 negotiation). If the unrecognized_name alert has to be sent, this by
9134 default is a warning; it becomes fatal with the '-servername_fatal'
9137 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9139 * Add AES and SSE2 assembly language support to VC++ build.
9143 * Mitigate attack on final subtraction in Montgomery reduction.
9147 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9148 (which previously caused an internal error).
9152 * Squeeze another 10% out of IGE mode when in != out.
9156 * AES IGE mode speedup.
9158 *Dean Gaudet (Google)*
9160 * Add the Korean symmetric 128-bit cipher SEED (see
9161 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9162 add SEED ciphersuites from RFC 4162:
9164 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9165 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9166 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9167 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9169 To minimize changes between patchlevels in the OpenSSL 0.9.8
9170 series, SEED remains excluded from compilation unless OpenSSL
9171 is configured with 'enable-seed'.
9173 *KISA, Bodo Moeller*
9175 * Mitigate branch prediction attacks, which can be practical if a
9176 single processor is shared, allowing a spy process to extract
9177 information. For detailed background information, see
9178 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9179 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9180 and Necessary Software Countermeasures"). The core of the change
9181 are new versions BN_div_no_branch() and
9182 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9183 respectively, which are slower, but avoid the security-relevant
9184 conditional branches. These are automatically called by BN_div()
9185 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9186 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9187 remove a conditional branch.
9189 BN_FLG_CONSTTIME is the new name for the previous
9190 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9191 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9192 in the exponent causes BN_mod_exp_mont() to use the alternative
9193 implementation in BN_mod_exp_mont_consttime().) The old name
9194 remains as a deprecated alias.
9196 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9197 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9198 constant-time implementations for more than just exponentiation.
9199 Here too the old name is kept as a deprecated alias.
9201 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9202 the BN_BLINDING structure gets an independent copy of the
9203 modulus. This means that the previous `BIGNUM *m` argument to
9204 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9205 essentially becomes `const BIGNUM *m`, although we can't actually
9206 change this in the header file before 0.9.9. It allows
9207 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9208 enable BN_FLG_CONSTTIME.
9210 *Matthew D Wood (Intel Corp)*
9212 * In the SSL/TLS server implementation, be strict about session ID
9213 context matching (which matters if an application uses a single
9214 external cache for different purposes). Previously,
9215 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9216 set. This did ensure strict client verification, but meant that,
9217 with applications using a single external cache for quite
9218 different requirements, clients could circumvent ciphersuite
9219 restrictions for a given session ID context by starting a session
9220 in a different context.
9224 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9225 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9226 authentication-only ciphersuites.
9230 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9231 not complete and could lead to a possible single byte overflow
9232 [CVE-2007-5135][] [Ben Laurie]
9234 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9236 * Since AES128 and AES256 (and similarly Camellia128 and
9237 Camellia256) share a single mask bit in the logic of
9238 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9239 kludge to work properly if AES128 is available and AES256 isn't
9240 (or if Camellia128 is available and Camellia256 isn't).
9244 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9245 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9246 When a point or a seed is encoded in a BIT STRING, we need to
9247 prevent the removal of trailing zero bits to get the proper DER
9248 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9249 of a NamedBitList, for which trailing 0 bits need to be removed.)
9253 * Have SSL/TLS server implementation tolerate "mismatched" record
9254 protocol version while receiving ClientHello even if the
9255 ClientHello is fragmented. (The server can't insist on the
9256 particular protocol version it has chosen before the ServerHello
9257 message has informed the client about his choice.)
9261 * Add RFC 3779 support.
9263 *Rob Austein for ARIN, Ben Laurie*
9265 * Load error codes if they are not already present instead of using a
9266 static variable. This allows them to be cleanly unloaded and reloaded.
9267 Improve header file function name parsing.
9271 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9272 or CAPABILITY handshake as required by RFCs.
9276 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9278 * Introduce limits to prevent malicious keys being able to
9279 cause a denial of service. [CVE-2006-2940][]
9281 *Steve Henson, Bodo Moeller*
9283 * Fix ASN.1 parsing of certain invalid structures that can result
9284 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9286 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9287 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9289 * Fix SSL client code which could crash if connecting to a
9290 malicious SSLv2 server. [CVE-2006-4343][]
9292 *Tavis Ormandy and Will Drewry, Google Security Team*
9294 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9295 match only those. Before that, "AES256-SHA" would be interpreted
9296 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9297 the same strength classification in 0.9.7h) as we currently only
9298 have a single AES bit in the ciphersuite description bitmap.
9299 That change, however, also applied to ciphersuite strings such as
9300 "RC4-MD5" that intentionally matched multiple ciphersuites --
9301 namely, SSL 2.0 ciphersuites in addition to the more common ones
9302 from SSL 3.0/TLS 1.0.
9304 So we change the selection algorithm again: Naming an explicit
9305 ciphersuite selects this one ciphersuite, and any other similar
9306 ciphersuite (same bitmap) from *other* protocol versions.
9307 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9308 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9310 Since SSL 2.0 does not have any ciphersuites for which the
9311 128/256 bit distinction would be relevant, this works for now.
9312 The proper fix will be to use different bits for AES128 and
9313 AES256, which would have avoided the problems from the beginning;
9314 however, bits are scarce, so we can only do this in a new release
9315 (not just a patchlevel) when we can change the SSL_CIPHER
9316 definition to split the single 'unsigned long mask' bitmap into
9317 multiple values to extend the available space.
9321 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9323 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9324 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9326 * Add AES IGE and biIGE modes.
9330 * Change the Unix randomness entropy gathering to use poll() when
9331 possible instead of select(), since the latter has some
9332 undesirable limitations.
9334 *Darryl Miles via Richard Levitte and Bodo Moeller*
9336 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9337 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9338 cannot be implicitly activated as part of, e.g., the "AES" alias.
9339 However, please upgrade to OpenSSL 0.9.9[-dev] for
9340 non-experimental use of the ECC ciphersuites to get TLS extension
9341 support, which is required for curve and point format negotiation
9342 to avoid potential handshake problems.
9346 * Disable rogue ciphersuites:
9348 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9349 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9350 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9352 The latter two were purportedly from
9353 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9356 Also deactivate the remaining ciphersuites from
9357 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9358 unofficial, and the ID has long expired.
9362 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9363 dual-core machines) and other potential thread-safety issues.
9367 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9368 versions), which is now available for royalty-free use
9369 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9370 Also, add Camellia TLS ciphersuites from RFC 4132.
9372 To minimize changes between patchlevels in the OpenSSL 0.9.8
9373 series, Camellia remains excluded from compilation unless OpenSSL
9374 is configured with 'enable-camellia'.
9378 * Disable the padding bug check when compression is in use. The padding
9379 bug check assumes the first packet is of even length, this is not
9380 necessarily true if compression is enabled and can result in false
9381 positives causing handshake failure. The actual bug test is ancient
9382 code so it is hoped that implementations will either have fixed it by
9383 now or any which still have the bug do not support compression.
9387 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9389 * When applying a cipher rule check to see if string match is an explicit
9390 cipher suite and only match that one cipher suite if it is.
9394 * Link in manifests for VC++ if needed.
9396 *Austin Ziegler <halostatue@gmail.com>*
9398 * Update support for ECC-based TLS ciphersuites according to
9399 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9400 TLS extensions, which are supported starting with the 0.9.9
9401 branch, not in the OpenSSL 0.9.8 branch).
9405 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9406 opaque EVP_CIPHER_CTX handling.
9410 * Fixes and enhancements to zlib compression code. We now only use
9411 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9412 to conform with the standards mentioned here:
9413 <http://www.zlib.net/DLL_FAQ.txt>
9414 Static zlib linking now works on Windows and the new --with-zlib-include
9415 --with-zlib-lib options to Configure can be used to supply the location
9416 of the headers and library. Gracefully handle case where zlib library
9421 * Several fixes and enhancements to the OID generation code. The old code
9422 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9423 handle numbers larger than ULONG_MAX, truncated printing and had a
9424 non standard OBJ_obj2txt() behaviour.
9428 * Add support for building of engines under engine/ as shared libraries
9429 under VC++ build system.
9433 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9434 Hopefully, we will not see any false combination of paths any more.
9438 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9440 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9441 (part of SSL_OP_ALL). This option used to disable the
9442 countermeasure against man-in-the-middle protocol-version
9443 rollback in the SSL 2.0 server implementation, which is a bad
9444 idea. [CVE-2005-2969][]
9446 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9447 for Information Security, National Institute of Advanced Industrial
9448 Science and Technology [AIST], Japan)*
9450 * Add two function to clear and return the verify parameter flags.
9454 * Keep cipherlists sorted in the source instead of sorting them at
9455 runtime, thus removing the need for a lock.
9459 * Avoid some small subgroup attacks in Diffie-Hellman.
9461 *Nick Mathewson and Ben Laurie*
9463 * Add functions for well-known primes.
9467 * Extended Windows CE support.
9469 *Satoshi Nakamura and Andy Polyakov*
9471 * Initialize SSL_METHOD structures at compile time instead of during
9472 runtime, thus removing the need for a lock.
9476 * Make PKCS7_decrypt() work even if no certificate is supplied by
9477 attempting to decrypt each encrypted key in turn. Add support to
9482 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9484 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9487 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9491 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9492 key into the same file any more.
9496 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9500 * Add -utf8 command line and config file option to 'ca'.
9502 *Stefan <stf@udoma.org*
9504 * Removed the macro des_crypt(), as it seems to conflict with some
9505 libraries. Use DES_crypt().
9509 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9510 involves renaming the source and generated shared-libs for
9511 both. The engines will accept the corrected or legacy ids
9512 ('ncipher' and '4758_cca' respectively) when binding. NB,
9513 this only applies when building 'shared'.
9515 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9517 * Add attribute functions to EVP_PKEY structure. Modify
9518 PKCS12_create() to recognize a CSP name attribute and
9519 use it. Make -CSP option work again in pkcs12 utility.
9523 * Add new functionality to the bn blinding code:
9524 - automatic re-creation of the BN_BLINDING parameters after
9525 a fixed number of uses (currently 32)
9526 - add new function for parameter creation
9527 - introduce flags to control the update behaviour of the
9528 BN_BLINDING parameters
9529 - hide BN_BLINDING structure
9530 Add a second BN_BLINDING slot to the RSA structure to improve
9531 performance when a single RSA object is shared among several
9536 * Add support for DTLS.
9538 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9540 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9541 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9545 * Remove buggy and incomplete DH cert support from
9546 ssl/ssl_rsa.c and ssl/s3_both.c
9550 * Use SHA-1 instead of MD5 as the default digest algorithm for
9551 the apps/openssl applications.
9555 * Compile clean with "-Wall -Wmissing-prototypes
9556 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9557 DEBUG_SAFESTACK must also be set.
9561 * Change ./Configure so that certain algorithms can be disabled by default.
9562 The new counterpiece to "no-xxx" is "enable-xxx".
9564 The patented RC5 and MDC2 algorithms will now be disabled unless
9565 "enable-rc5" and "enable-mdc2", respectively, are specified.
9567 (IDEA remains enabled despite being patented. This is because IDEA
9568 is frequently required for interoperability, and there is no license
9569 fee for non-commercial use. As before, "no-idea" can be used to
9570 avoid this algorithm.)
9574 * Add processing of proxy certificates (see RFC 3820). This work was
9575 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9576 EGEE (Enabling Grids for E-science in Europe).
9580 * RC4 performance overhaul on modern architectures/implementations, such
9581 as Intel P4, IA-64 and AMD64.
9585 * New utility extract-section.pl. This can be used specify an alternative
9586 section number in a pod file instead of having to treat each file as
9587 a separate case in Makefile. This can be done by adding two lines to the
9590 =for comment openssl_section:XXX
9592 The blank line is mandatory.
9596 * New arguments -certform, -keyform and -pass for s_client and s_server
9597 to allow alternative format key and certificate files and passphrase
9602 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9603 update associated structures and add various utility functions.
9605 Add new policy related verify parameters, include policy checking in
9606 standard verify code. Enhance 'smime' application with extra parameters
9607 to support policy checking and print out.
9611 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9612 Nehemiah processors. These extensions support AES encryption in hardware
9613 as well as RNG (though RNG support is currently disabled).
9615 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9617 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9621 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9623 *Andy Polyakov and a number of other people*
9625 * Improved PowerPC platform support. Most notably BIGNUM assembler
9626 implementation contributed by IBM.
9628 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9630 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9631 exponent rather than 'unsigned long'. There is a corresponding change to
9632 the new 'rsa_keygen' element of the RSA_METHOD structure.
9634 *Jelte Jansen, Geoff Thorpe*
9636 * Functionality for creating the initial serial number file is now
9637 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9639 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9640 number file to 1, which is bound to cause problems. To avoid
9641 the problems while respecting compatibility between different 0.9.7
9642 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9643 CA.pl for serial number initialization. With the new release 0.9.8,
9644 we can fix the problem directly in the 'ca' utility.)
9648 * Reduced header interdependencies by declaring more opaque objects in
9649 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9650 give fewer recursive includes, which could break lazy source code - so
9651 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9652 developers should define this symbol when building and using openssl to
9653 ensure they track the recommended behaviour, interfaces, [etc], but
9654 backwards-compatible behaviour prevails when this isn't defined.
9658 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9662 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9663 This will generate a random key of the appropriate length based on the
9664 cipher context. The EVP_CIPHER can provide its own random key generation
9665 routine to support keys of a specific form. This is used in the des and
9666 3des routines to generate a key of the correct parity. Update S/MIME
9667 code to use new functions and hence generate correct parity DES keys.
9668 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9669 valid (weak or incorrect parity).
9673 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9674 as looking them up. This is useful when the verified structure may contain
9675 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9676 present unless the new PKCS7_NO_CRL flag is asserted.
9680 * Extend ASN1 oid configuration module. It now additionally accepts the
9683 shortName = some long name, 1.2.3.4
9687 * Reimplemented the BN_CTX implementation. There is now no more static
9688 limitation on the number of variables it can handle nor the depth of the
9689 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9690 information can now expand as required, and rather than having a single
9691 static array of bignums, BN_CTX now uses a linked-list of such arrays
9692 allowing it to expand on demand whilst maintaining the usefulness of
9693 BN_CTX's "bundling".
9697 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9698 to allow all RSA operations to function using a single BN_CTX.
9702 * Preliminary support for certificate policy evaluation and checking. This
9703 is initially intended to pass the tests outlined in "Conformance Testing
9704 of Relying Party Client Certificate Path Processing Logic" v1.07.
9708 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9709 remained unused and not that useful. A variety of other little bignum
9710 tweaks and fixes have also been made continuing on from the audit (see
9715 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9716 associated ASN1, EVP and SSL functions and old ASN1 macros.
9720 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9721 and this should never fail. So the return value from the use of
9722 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9723 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9727 * BN_CTX_get() should return zero-valued bignums, providing the same
9728 initialised value as BN_new().
9730 *Geoff Thorpe, suggested by Ulf Möller*
9732 * Support for inhibitAnyPolicy certificate extension.
9736 * An audit of the BIGNUM code is underway, for which debugging code is
9737 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9738 is considered valid when processing BIGNUMs, and causes execution to
9739 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9740 further steps are taken to deliberately pollute unused data in BIGNUM
9741 structures to try and expose faulty code further on. For now, openssl will
9742 (in its default mode of operation) continue to tolerate the inconsistent
9743 forms that it has tolerated in the past, but authors and packagers should
9744 consider trying openssl and their own applications when compiled with
9745 these debugging symbols defined. It will help highlight potential bugs in
9746 their own code, and will improve the test coverage for OpenSSL itself. At
9747 some point, these tighter rules will become openssl's default to improve
9748 maintainability, though the assert()s and other overheads will remain only
9749 in debugging configurations. See bn.h for more details.
9751 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9753 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9754 that can only be obtained through BN_CTX_new() (which implicitly
9755 initialises it). The presence of this function only made it possible
9756 to overwrite an existing structure (and cause memory leaks).
9760 * Because of the callback-based approach for implementing LHASH as a
9761 template type, lh_insert() adds opaque objects to hash-tables and
9762 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9763 to clean up those corresponding objects before destroying the hash table
9764 (and losing the object pointers). So some over-zealous constifications in
9765 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9766 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9767 prototyped to have "const" restrictions on the object pointers they are
9768 given (and so aren't required to cast them away any more).
9772 * The tmdiff.h API was so ugly and minimal that our own timing utility
9773 (speed) prefers to use its own implementation. The two implementations
9774 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9775 its object type properly exposed (MS_TM) instead of casting to/from
9776 `char *`. This may still change yet if someone realises MS_TM and
9778 aren't necessarily the greatest nomenclatures - but this is what was used
9779 internally to the implementation so I've used that for now.
9783 * Ensure that deprecated functions do not get compiled when
9784 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9785 the self-tests were still using deprecated key-generation functions so
9786 these have been updated also.
9790 * Reorganise PKCS#7 code to separate the digest location functionality
9791 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9792 New function PKCS7_set_digest() to set the digest type for PKCS#7
9793 digestedData type. Add additional code to correctly generate the
9794 digestedData type and add support for this type in PKCS7 initialization
9799 * New function PKCS7_set0_type_other() this initializes a PKCS7
9800 structure of type "other".
9804 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9805 sure the loop does correctly stop and breaking ("division by zero")
9806 modulus operations are not performed. The (pre-generated) prime
9807 table crypto/bn/bn_prime.h was already correct, but it could not be
9808 re-generated on some platforms because of the "division by zero"
9809 situation in the script.
9811 *Ralf S. Engelschall*
9813 * Update support for ECC-based TLS ciphersuites according to
9814 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9815 SHA-1 now is only used for "small" curves (where the
9816 representation of a field element takes up to 24 bytes); for
9817 larger curves, the field element resulting from ECDH is directly
9818 used as premaster secret.
9820 *Douglas Stebila (Sun Microsystems Laboratories)*
9822 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9823 curve secp160r1 to the tests.
9825 *Douglas Stebila (Sun Microsystems Laboratories)*
9827 * Add the possibility to load symbols globally with DSO.
9829 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9831 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9832 control of the error stack.
9836 * Add support for STORE in ENGINE.
9840 * Add the STORE type. The intention is to provide a common interface
9841 to certificate and key stores, be they simple file-based stores, or
9842 HSM-type store, or LDAP stores, or...
9843 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9847 * Add a generic structure called OPENSSL_ITEM. This can be used to
9848 pass a list of arguments to any function as well as provide a way
9849 for a function to pass data back to the caller.
9853 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9854 works like BUF_strdup() but can be used to duplicate a portion of
9855 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9860 * Add the function sk_find_ex() which works like sk_find(), but will
9861 return an index to an element even if an exact match couldn't be
9862 found. The index is guaranteed to point at the element where the
9863 searched-for key would be inserted to preserve sorting order.
9867 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9868 takes an extra flags argument for optional functionality. Currently,
9869 the following flags are defined:
9871 OBJ_BSEARCH_VALUE_ON_NOMATCH
9872 This one gets OBJ_bsearch_ex() to return a pointer to the first
9873 element where the comparing function returns a negative or zero
9876 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9877 This one gets OBJ_bsearch_ex() to return a pointer to the first
9878 element where the comparing function returns zero. This is useful
9879 if there are more than one element where the comparing function
9884 * Make it possible to create self-signed certificates with 'openssl ca'
9885 in such a way that the self-signed certificate becomes part of the
9886 CA database and uses the same mechanisms for serial number generation
9887 as all other certificate signing. The new flag '-selfsign' enables
9888 this functionality. Adapt CA.sh and CA.pl.in.
9892 * Add functionality to check the public key of a certificate request
9893 against a given private. This is useful to check that a certificate
9894 request can be signed by that key (self-signing).
9898 * Make it possible to have multiple active certificates with the same
9899 subject in the CA index file. This is done only if the keyword
9900 'unique_subject' is set to 'no' in the main CA section (default
9901 if 'CA_default') of the configuration file. The value is saved
9902 with the database itself in a separate index attribute file,
9903 named like the index file with '.attr' appended to the name.
9907 * Generate multi-valued AVAs using '+' notation in config files for
9912 * Support for nameConstraints certificate extension.
9916 * Support for policyConstraints certificate extension.
9920 * Support for policyMappings certificate extension.
9924 * Make sure the default DSA_METHOD implementation only uses its
9925 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9926 and change its own handlers to be NULL so as to remove unnecessary
9927 indirection. This lets alternative implementations fallback to the
9928 default implementation more easily.
9932 * Support for directoryName in GeneralName related extensions
9937 * Make it possible to link applications using Makefile.shared.
9938 Make that possible even when linking against static libraries!
9942 * Support for single pass processing for S/MIME signing. This now
9943 means that S/MIME signing can be done from a pipe, in addition
9944 cleartext signing (multipart/signed type) is effectively streaming
9945 and the signed data does not need to be all held in memory.
9947 This is done with a new flag PKCS7_STREAM. When this flag is set
9948 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9949 is done after the data is output (and digests calculated) in
9950 SMIME_write_PKCS7().
9954 * Add full support for -rpath/-R, both in shared libraries and
9955 applications, at least on the platforms where it's known how
9960 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9961 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9962 will now compute a table of multiples of the generator that
9963 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9964 faster (notably in the case of a single point multiplication,
9965 scalar * generator).
9967 *Nils Larsch, Bodo Moeller*
9969 * IPv6 support for certificate extensions. The various extensions
9970 which use the IP:a.b.c.d can now take IPv6 addresses using the
9971 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9976 * Added an ENGINE that implements RSA by performing private key
9977 exponentiations with the GMP library. The conversions to and from
9978 GMP's mpz_t format aren't optimised nor are any montgomery forms
9979 cached, and on x86 it appears OpenSSL's own performance has caught up.
9980 However there are likely to be other architectures where GMP could
9981 provide a boost. This ENGINE is not built in by default, but it can be
9982 specified at Configure time and should be accompanied by the necessary
9983 linker additions, eg;
9984 ./config -DOPENSSL_USE_GMP -lgmp
9988 * "openssl engine" will not display ENGINE/DSO load failure errors when
9989 testing availability of engines with "-t" - the old behaviour is
9990 produced by increasing the feature's verbosity with "-tt".
9994 * ECDSA routines: under certain error conditions uninitialized BN objects
9995 could be freed. Solution: make sure initialization is performed early
9996 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10001 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10002 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10003 software implementations. For DSA and DH, parameter generation can
10004 also be overridden by providing the appropriate method callbacks.
10008 * Change the "progress" mechanism used in key-generation and
10009 primality testing to functions that take a new BN_GENCB pointer in
10010 place of callback/argument pairs. The new API functions have `_ex`
10011 postfixes and the older functions are reimplemented as wrappers for
10012 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10013 declarations of the old functions to help (graceful) attempts to
10014 migrate to the new functions. Also, the new key-generation API
10015 functions operate on a caller-supplied key-structure and return
10016 success/failure rather than returning a key or NULL - this is to
10017 help make "keygen" another member function of RSA_METHOD etc.
10019 Example for using the new callback interface:
10021 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10022 void *my_arg = ...;
10025 BN_GENCB_set(&my_cb, my_callback, my_arg);
10027 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10028 /* For the meaning of a, b in calls to my_callback(), see the
10029 * documentation of the function that calls the callback.
10030 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10031 * my_callback should return 1 if it wants BN_is_prime_ex()
10032 * to continue, or 0 to stop.
10037 * Change the ZLIB compression method to be stateful, and make it
10038 available to TLS with the number defined in
10039 draft-ietf-tls-compression-04.txt.
10043 * Add the ASN.1 structures and functions for CertificatePair, which
10044 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10046 CertificatePair ::= SEQUENCE {
10047 forward [0] Certificate OPTIONAL,
10048 reverse [1] Certificate OPTIONAL,
10049 -- at least one of the pair shall be present -- }
10051 Also implement the PEM functions to read and write certificate
10052 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10054 This needed to be defined, mostly for the sake of the LDAP
10055 attribute crossCertificatePair, but may prove useful elsewhere as
10060 * Make it possible to inhibit symlinking of shared libraries in
10061 Makefile.shared, for Cygwin's sake.
10065 * Extend the BIGNUM API by creating a function
10066 void BN_set_negative(BIGNUM *a, int neg);
10067 and a macro that behave like
10068 int BN_is_negative(const BIGNUM *a);
10070 to avoid the need to access 'a->neg' directly in applications.
10074 * Implement fast modular reduction for pseudo-Mersenne primes
10075 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10076 EC_GROUP_new_curve_GFp() will now automatically use this
10079 *Nils Larsch <nla@trustcenter.de>*
10081 * Add new lock type (CRYPTO_LOCK_BN).
10085 * Change the ENGINE framework to automatically load engines
10086 dynamically from specific directories unless they could be
10087 found to already be built in or loaded. Move all the
10088 current engines except for the cryptodev one to a new
10089 directory engines/.
10090 The engines in engines/ are built as shared libraries if
10091 the "shared" options was given to ./Configure or ./config.
10092 Otherwise, they are inserted in libcrypto.a.
10093 /usr/local/ssl/engines is the default directory for dynamic
10094 engines, but that can be overridden at configure time through
10095 the usual use of --prefix and/or --openssldir, and at run
10096 time with the environment variable OPENSSL_ENGINES.
10098 *Geoff Thorpe and Richard Levitte*
10100 * Add Makefile.shared, a helper makefile to build shared
10101 libraries. Adapt Makefile.org.
10105 * Add version info to Win32 DLLs.
10107 *Peter 'Luna' Runestig" <peter@runestig.com>*
10109 * Add new 'medium level' PKCS#12 API. Certificates and keys
10110 can be added using this API to created arbitrary PKCS#12
10111 files while avoiding the low level API.
10113 New options to PKCS12_create(), key or cert can be NULL and
10114 will then be omitted from the output file. The encryption
10115 algorithm NIDs can be set to -1 for no encryption, the mac
10116 iteration count can be set to 0 to omit the mac.
10118 Enhance pkcs12 utility by making the -nokeys and -nocerts
10119 options work when creating a PKCS#12 file. New option -nomac
10120 to omit the mac, NONE can be set for an encryption algorithm.
10121 New code is modified to use the enhanced PKCS12_create()
10122 instead of the low level API.
10126 * Extend ASN1 encoder to support indefinite length constructed
10127 encoding. This can output sequences tags and octet strings in
10128 this form. Modify pk7_asn1.c to support indefinite length
10129 encoding. This is experimental and needs additional code to
10130 be useful, such as an ASN1 bio and some enhanced streaming
10133 Extend template encode functionality so that tagging is passed
10134 down to the template encoder.
10138 * Let 'openssl req' fail if an argument to '-newkey' is not
10139 recognized instead of using RSA as a default.
10143 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10144 As these are not official, they are not included in "ALL";
10145 the "ECCdraft" ciphersuite group alias can be used to select them.
10147 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10149 * Add ECDH engine support.
10151 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10153 * Add ECDH in new directory crypto/ecdh/.
10155 *Douglas Stebila (Sun Microsystems Laboratories)*
10157 * Let BN_rand_range() abort with an error after 100 iterations
10158 without success (which indicates a broken PRNG).
10162 * Change BN_mod_sqrt() so that it verifies that the input value
10163 is really the square of the return value. (Previously,
10164 BN_mod_sqrt would show GIGO behaviour.)
10168 * Add named elliptic curves over binary fields from X9.62, SECG,
10169 and WAP/WTLS; add OIDs that were still missing.
10171 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10173 * Extend the EC library for elliptic curves over binary fields
10174 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10177 EC_GF2m_simple_method
10181 EC_GROUP_new_curve_GF2m
10182 EC_GROUP_set_curve_GF2m
10183 EC_GROUP_get_curve_GF2m
10184 EC_POINT_set_affine_coordinates_GF2m
10185 EC_POINT_get_affine_coordinates_GF2m
10186 EC_POINT_set_compressed_coordinates_GF2m
10188 Point compression for binary fields is disabled by default for
10189 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10192 As binary polynomials are represented as BIGNUMs, various members
10193 of the EC_GROUP and EC_POINT data structures can be shared
10194 between the implementations for prime fields and binary fields;
10195 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10196 are essentially identical to their `..._GFp` counterparts.
10197 (For simplicity, the `..._GFp` prefix has been dropped from
10198 various internal method names.)
10200 An internal 'field_div' method (similar to 'field_mul' and
10201 'field_sqr') has been added; this is used only for binary fields.
10203 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10205 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10206 through methods ('mul', 'precompute_mult').
10208 The generic implementations (now internally called 'ec_wNAF_mul'
10209 and 'ec_wNAF_precomputed_mult') remain the default if these
10210 methods are undefined.
10212 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10214 * New function EC_GROUP_get_degree, which is defined through
10215 EC_METHOD. For curves over prime fields, this returns the bit
10216 length of the modulus.
10218 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10220 * New functions EC_GROUP_dup, EC_POINT_dup.
10221 (These simply call ..._new and ..._copy).
10223 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10225 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10226 Polynomials are represented as BIGNUMs (where the sign bit is not
10227 used) in the following functions [macros]:
10230 BN_GF2m_sub [= BN_GF2m_add]
10231 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10232 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10233 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10235 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10236 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10237 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10238 BN_GF2m_cmp [= BN_ucmp]
10240 (Note that only the 'mod' functions are actually for fields GF(2^m).
10241 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10243 For some functions, an the irreducible polynomial defining a
10244 field can be given as an 'unsigned int[]' with strictly
10245 decreasing elements giving the indices of those bits that are set;
10246 i.e., p[] represents the polynomial
10247 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10249 p[0] > p[1] > ... > p[k] = 0.
10250 This applies to the following functions:
10253 BN_GF2m_mod_mul_arr
10254 BN_GF2m_mod_sqr_arr
10255 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10256 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10257 BN_GF2m_mod_exp_arr
10258 BN_GF2m_mod_sqrt_arr
10259 BN_GF2m_mod_solve_quad_arr
10263 Conversion can be performed by the following functions:
10268 bntest.c has additional tests for binary polynomial arithmetic.
10270 Two implementations for BN_GF2m_mod_div() are available.
10271 The default algorithm simply uses BN_GF2m_mod_inv() and
10272 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10273 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10274 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10276 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10278 * Add new error code 'ERR_R_DISABLED' that can be used when some
10279 functionality is disabled at compile-time.
10281 *Douglas Stebila <douglas.stebila@sun.com>*
10283 * Change default behaviour of 'openssl asn1parse' so that more
10284 information is visible when viewing, e.g., a certificate:
10286 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10287 mode the content of non-printable OCTET STRINGs is output in a
10288 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10289 avoid the appearance of a printable string.
10291 *Nils Larsch <nla@trustcenter.de>*
10293 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10295 EC_GROUP_set_asn1_flag()
10296 EC_GROUP_get_asn1_flag()
10297 EC_GROUP_set_point_conversion_form()
10298 EC_GROUP_get_point_conversion_form()
10299 These control ASN1 encoding details:
10300 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10301 has been set to OPENSSL_EC_NAMED_CURVE.
10302 - Points are encoded in uncompressed form by default; options for
10303 asn1_for are as for point2oct, namely
10304 POINT_CONVERSION_COMPRESSED
10305 POINT_CONVERSION_UNCOMPRESSED
10306 POINT_CONVERSION_HYBRID
10308 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10310 EC_GROUP_set_seed()
10311 EC_GROUP_get0_seed()
10312 EC_GROUP_get_seed_len()
10313 This is used only for ASN1 purposes (so far).
10315 *Nils Larsch <nla@trustcenter.de>*
10317 * Add 'field_type' member to EC_METHOD, which holds the NID
10318 of the appropriate field type OID. The new function
10319 EC_METHOD_get_field_type() returns this value.
10321 *Nils Larsch <nla@trustcenter.de>*
10324 EC_POINT_point2bn()
10325 EC_POINT_bn2point()
10326 EC_POINT_point2hex()
10327 EC_POINT_hex2point()
10328 providing useful interfaces to EC_POINT_point2oct() and
10329 EC_POINT_oct2point().
10331 *Nils Larsch <nla@trustcenter.de>*
10333 * Change internals of the EC library so that the functions
10334 EC_GROUP_set_generator()
10335 EC_GROUP_get_generator()
10336 EC_GROUP_get_order()
10337 EC_GROUP_get_cofactor()
10338 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10339 to methods, which would lead to unnecessary code duplication when
10340 adding different types of curves.
10342 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10344 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10345 arithmetic, and such that modified wNAFs are generated
10346 (which avoid length expansion in many cases).
10350 * Add a function EC_GROUP_check_discriminant() (defined via
10351 EC_METHOD) that verifies that the curve discriminant is non-zero.
10353 Add a function EC_GROUP_check() that makes some sanity tests
10354 on a EC_GROUP, its generator and order. This includes
10355 EC_GROUP_check_discriminant().
10357 *Nils Larsch <nla@trustcenter.de>*
10359 * Add ECDSA in new directory crypto/ecdsa/.
10361 Add applications 'openssl ecparam' and 'openssl ecdsa'
10362 (these are based on 'openssl dsaparam' and 'openssl dsa').
10364 ECDSA support is also included in various other files across the
10365 library. Most notably,
10366 - 'openssl req' now has a '-newkey ecdsa:file' option;
10367 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10368 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10369 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10370 them suitable for ECDSA where domain parameters must be
10371 extracted before the specific public key;
10372 - ECDSA engine support has been added.
10374 *Nils Larsch <nla@trustcenter.de>*
10376 * Include some named elliptic curves, and add OIDs from X9.62,
10377 SECG, and WAP/WTLS. Each curve can be obtained from the new
10379 EC_GROUP_new_by_curve_name(),
10380 and the list of available named curves can be obtained with
10381 EC_get_builtin_curves().
10382 Also add a 'curve_name' member to EC_GROUP objects, which can be
10384 EC_GROUP_set_curve_name()
10385 EC_GROUP_get_curve_name()
10387 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10389 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10390 was actually never needed) and in BN_mul(). The removal in BN_mul()
10391 required a small change in bn_mul_part_recursive() and the addition
10392 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10393 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10394 bn_sub_words() and bn_add_words() except they take arrays with
10399 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10401 * Cleanse PEM buffers before freeing them since they may contain
10404 *Benjamin Bennett <ben@psc.edu>*
10406 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10407 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10408 authentication-only ciphersuites.
10412 * Since AES128 and AES256 share a single mask bit in the logic of
10413 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10414 kludge to work properly if AES128 is available and AES256 isn't.
10418 * Expand security boundary to match 1.1.1 module.
10422 * Remove redundant features: hash file source, editing of test vectors
10423 modify fipsld to use external fips_premain.c signature.
10427 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10428 run algorithm test programs.
10432 * Make algorithm test programs more tolerant of whitespace.
10436 * Have SSL/TLS server implementation tolerate "mismatched" record
10437 protocol version while receiving ClientHello even if the
10438 ClientHello is fragmented. (The server can't insist on the
10439 particular protocol version it has chosen before the ServerHello
10440 message has informed the client about his choice.)
10444 * Load error codes if they are not already present instead of using a
10445 static variable. This allows them to be cleanly unloaded and reloaded.
10449 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10451 * Introduce limits to prevent malicious keys being able to
10452 cause a denial of service. [CVE-2006-2940][]
10454 *Steve Henson, Bodo Moeller*
10456 * Fix ASN.1 parsing of certain invalid structures that can result
10457 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10459 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10460 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10462 * Fix SSL client code which could crash if connecting to a
10463 malicious SSLv2 server. [CVE-2006-4343][]
10465 *Tavis Ormandy and Will Drewry, Google Security Team*
10467 * Change ciphersuite string processing so that an explicit
10468 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10469 will no longer include "AES128-SHA"), and any other similar
10470 ciphersuite (same bitmap) from *other* protocol versions (so that
10471 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10472 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10473 changes from 0.9.8b and 0.9.8d.
10477 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10479 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10480 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10482 * Change the Unix randomness entropy gathering to use poll() when
10483 possible instead of select(), since the latter has some
10484 undesirable limitations.
10486 *Darryl Miles via Richard Levitte and Bodo Moeller*
10488 * Disable rogue ciphersuites:
10490 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10491 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10492 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10494 The latter two were purportedly from
10495 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10498 Also deactivate the remaining ciphersuites from
10499 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10500 unofficial, and the ID has long expired.
10504 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10505 dual-core machines) and other potential thread-safety issues.
10509 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10511 * Adapt fipsld and the build system to link against the validated FIPS
10512 module in FIPS mode.
10516 * Fixes for VC++ 2005 build under Windows.
10520 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10521 from a Windows bash shell such as MSYS. It is autodetected from the
10522 "config" script when run from a VC++ environment. Modify standard VC++
10523 build to use fipscanister.o from the GNU make build.
10527 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10529 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10530 The value now differs depending on if you build for FIPS or not.
10531 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10532 safely run with a non-FIPSed libcrypto, as it may crash because of
10533 the difference induced by this change.
10537 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10539 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10540 (part of SSL_OP_ALL). This option used to disable the
10541 countermeasure against man-in-the-middle protocol-version
10542 rollback in the SSL 2.0 server implementation, which is a bad
10543 idea. [CVE-2005-2969][]
10545 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10546 for Information Security, National Institute of Advanced Industrial
10547 Science and Technology [AIST, Japan)]*
10549 * Minimal support for X9.31 signatures and PSS padding modes. This is
10550 mainly for FIPS compliance and not fully integrated at this stage.
10554 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10555 the exponentiation using a fixed-length exponent. (Otherwise,
10556 the information leaked through timing could expose the secret key
10557 after many signatures; cf. Bleichenbacher's attack on DSA with
10562 * Make a new fixed-window mod_exp implementation the default for
10563 RSA, DSA, and DH private-key operations so that the sequence of
10564 squares and multiplies and the memory access pattern are
10565 independent of the particular secret key. This will mitigate
10566 cache-timing and potential related attacks.
10568 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10569 and this is automatically used by BN_mod_exp_mont() if the new flag
10570 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10571 will use this BN flag for private exponents unless the flag
10572 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10573 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10575 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10577 * Change the client implementation for SSLv23_method() and
10578 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10579 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10580 (Previously, the SSL 2.0 backwards compatible Client Hello
10581 message format would be used even with SSL_OP_NO_SSLv2.)
10585 * Add support for smime-type MIME parameter in S/MIME messages which some
10590 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10591 a threadsafe manner. Modify rsa code to use new function and add calls
10592 to dsa and dh code (which had race conditions before).
10596 * Include the fixed error library code in the C error file definitions
10597 instead of fixing them up at runtime. This keeps the error code
10598 structures constant.
10602 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10604 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10607 * Fixes for newer kerberos headers. NB: the casts are needed because
10608 the 'length' field is signed on one version and unsigned on another
10609 with no (?) obvious way to tell the difference, without these VC++
10610 complains. Also the "definition" of FAR (blank) is no longer included
10611 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10612 some needed definitions.
10616 * Undo Cygwin change.
10620 * Added support for proxy certificates according to RFC 3820.
10621 Because they may be a security thread to unaware applications,
10622 they must be explicitly allowed in run-time. See
10623 docs/HOWTO/proxy_certificates.txt for further information.
10627 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10629 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10630 server and client random values. Previously
10631 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10632 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10634 This change has negligible security impact because:
10636 1. Server and client random values still have 24 bytes of pseudo random
10639 2. Server and client random values are sent in the clear in the initial
10642 3. The master secret is derived using the premaster secret (48 bytes in
10643 size for static RSA ciphersuites) as well as client server and random
10646 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10649 *Stephen Henson, reported by UK NISCC*
10651 * Use Windows randomness collection on Cygwin.
10655 * Fix hang in EGD/PRNGD query when communication socket is closed
10656 prematurely by EGD/PRNGD.
10658 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10660 * Prompt for pass phrases when appropriate for PKCS12 input format.
10664 * Back-port of selected performance improvements from development
10665 branch, as well as improved support for PowerPC platforms.
10669 * Add lots of checks for memory allocation failure, error codes to indicate
10670 failure and freeing up memory if a failure occurs.
10672 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10674 * Add new -passin argument to dgst.
10678 * Perform some character comparisons of different types in X509_NAME_cmp:
10679 this is needed for some certificates that re-encode DNs into UTF8Strings
10680 (in violation of RFC3280) and can't or won't issue name rollover
10685 * Make an explicit check during certificate validation to see that
10686 the CA setting in each certificate on the chain is correct. As a
10687 side effect always do the following basic checks on extensions,
10688 not just when there's an associated purpose to the check:
10690 - if there is an unhandled critical extension (unless the user
10691 has chosen to ignore this fault)
10692 - if the path length has been exceeded (if one is set at all)
10693 - that certain extensions fit the associated purpose (if one has
10698 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10700 * Avoid a race condition when CRLs are checked in a multi threaded
10701 environment. This would happen due to the reordering of the revoked
10702 entries during signature checking and serial number lookup. Now the
10703 encoding is cached and the serial number sort performed under a lock.
10704 Add new STACK function sk_is_sorted().
10708 * Add Delta CRL to the extension code.
10712 * Various fixes to s3_pkt.c so alerts are sent properly.
10714 *David Holmes <d.holmes@f5.com>*
10716 * Reduce the chances of duplicate issuer name and serial numbers (in
10717 violation of RFC3280) using the OpenSSL certificate creation utilities.
10718 This is done by creating a random 64 bit value for the initial serial
10719 number when a serial number file is created or when a self signed
10720 certificate is created using 'openssl req -x509'. The initial serial
10721 number file is created using 'openssl x509 -next_serial' in CA.pl
10722 rather than being initialized to 1.
10726 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10728 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10729 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10731 *Joe Orton, Steve Henson*
10733 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10736 *Joe Orton, Steve Henson*
10738 * Make it possible to have multiple active certificates with the same
10739 subject in the CA index file. This is done only if the keyword
10740 'unique_subject' is set to 'no' in the main CA section (default
10741 if 'CA_default') of the configuration file. The value is saved
10742 with the database itself in a separate index attribute file,
10743 named like the index file with '.attr' appended to the name.
10747 * X509 verify fixes. Disable broken certificate workarounds when
10748 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10749 keyUsage extension present. Don't accept CRLs with unhandled critical
10750 extensions: since verify currently doesn't process CRL extensions this
10751 rejects a CRL with *any* critical extensions. Add new verify error codes
10756 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10757 A clarification of RFC2560 will require the use of OCTET STRINGs and
10758 some implementations cannot handle the current raw format. Since OpenSSL
10759 copies and compares OCSP nonces as opaque blobs without any attempt at
10760 parsing them this should not create any compatibility issues.
10764 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10765 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10766 this HMAC (and other) operations are several times slower than OpenSSL
10771 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10773 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10775 * Use the correct content when signing type "other".
10779 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10781 * Fix various bugs revealed by running the NISCC test suite:
10783 Stop out of bounds reads in the ASN1 code when presented with
10784 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10786 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10788 If verify callback ignores invalid public key errors don't try to check
10789 certificate signature with the NULL public key.
10793 * New -ignore_err option in ocsp application to stop the server
10794 exiting on the first error in a request.
10798 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10799 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10804 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10805 extra data after the compression methods not only for TLS 1.0
10806 but also for SSL 3.0 (as required by the specification).
10808 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10810 * Change X509_certificate_type() to mark the key as exported/exportable
10811 when it's 512 *bits* long, not 512 bytes.
10815 * Change AES_cbc_encrypt() so it outputs exact multiple of
10816 blocks during encryption.
10820 * Various fixes to base64 BIO and non blocking I/O. On write
10821 flushes were not handled properly if the BIO retried. On read
10822 data was not being buffered properly and had various logic bugs.
10823 This also affects blocking I/O when the data being decoded is a
10828 * Various S/MIME bugfixes and compatibility changes:
10829 output correct application/pkcs7 MIME type if
10830 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10831 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10832 of files as .eml work). Correctly handle very long lines in MIME
10837 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10839 * Countermeasure against the Klima-Pokorny-Rosa extension of
10840 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10841 a protocol version number mismatch like a decryption error
10842 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10846 * Turn on RSA blinding by default in the default implementation
10847 to avoid a timing attack. Applications that don't want it can call
10848 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10849 They would be ill-advised to do so in most cases.
10851 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10853 * Change RSA blinding code so that it works when the PRNG is not
10854 seeded (in this case, the secret RSA exponent is abused as
10855 an unpredictable seed -- if it is not unpredictable, there
10856 is no point in blinding anyway). Make RSA blinding thread-safe
10857 by remembering the creator's thread ID in rsa->blinding and
10858 having all other threads use local one-time blinding factors
10859 (this requires more computation than sharing rsa->blinding, but
10860 avoids excessive locking; and if an RSA object is not shared
10861 between threads, blinding will still be very fast).
10865 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10866 ENGINE as defaults for all supported algorithms irrespective of
10867 the 'flags' parameter. 'flags' is now honoured, so applications
10868 should make sure they are passing it correctly.
10872 * Target "mingw" now allows native Windows code to be generated in
10873 the Cygwin environment as well as with the MinGW compiler.
10877 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10879 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10880 via timing by performing a MAC computation even if incorrect
10881 block cipher padding has been found. This is a countermeasure
10882 against active attacks where the attacker has to distinguish
10883 between bad padding and a MAC verification error. [CVE-2003-0078][]
10885 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10886 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10887 Martin Vuagnoux (EPFL, Ilion)*
10889 * Make the no-err option work as intended. The intention with no-err
10890 is not to have the whole error stack handling routines removed from
10891 libcrypto, it's only intended to remove all the function name and
10892 reason texts, thereby removing some of the footprint that may not
10893 be interesting if those errors aren't displayed anyway.
10895 NOTE: it's still possible for any application or module to have its
10896 own set of error texts inserted. The routines are there, just not
10897 used by default when no-err is given.
10901 * Add support for FreeBSD on IA64.
10903 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10905 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10906 Kerberos function mit_des_cbc_cksum(). Before this change,
10907 the value returned by DES_cbc_cksum() was like the one from
10908 mit_des_cbc_cksum(), except the bytes were swapped.
10910 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10912 * Allow an application to disable the automatic SSL chain building.
10913 Before this a rather primitive chain build was always performed in
10914 ssl3_output_cert_chain(): an application had no way to send the
10915 correct chain if the automatic operation produced an incorrect result.
10917 Now the chain builder is disabled if either:
10919 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10921 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10923 The reasoning behind this is that an application would not want the
10924 auto chain building to take place if extra chain certificates are
10925 present and it might also want a means of sending no additional
10926 certificates (for example the chain has two certificates and the
10931 * Add the possibility to build without the ENGINE framework.
10933 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10935 * Under Win32 gmtime() can return NULL: check return value in
10936 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10940 * DSA routines: under certain error conditions uninitialized BN objects
10941 could be freed. Solution: make sure initialization is performed early
10942 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10943 Nils Larsch <nla@trustcenter.de> via PR#459)
10947 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10948 checked on reconnect on the client side, therefore session resumption
10949 could still fail with a "ssl session id is different" error. This
10950 behaviour is masked when SSL_OP_ALL is used due to
10951 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10952 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10953 followup to PR #377.
10957 * IA-32 assembler support enhancements: unified ELF targets, support
10958 for SCO/Caldera platforms, fix for Cygwin shared build.
10962 * Add support for FreeBSD on sparc64. As a consequence, support for
10963 FreeBSD on non-x86 processors is separate from x86 processors on
10964 the config script, much like the NetBSD support.
10966 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10968 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
10970 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10973 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10974 code (06) was taken as the first octet of the session ID and the last
10975 octet was ignored consequently. As a result SSLv2 client side session
10976 caching could not have worked due to the session ID mismatch between
10978 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10983 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10984 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10989 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10990 seems that in spite of existing for more than a year, many application
10991 author have done nothing to provide the necessary callbacks, which
10992 means that this particular engine will not work properly anywhere.
10993 This is a very unfortunate situation which forces us, in the name
10994 of usability, to give the hw_ncipher.c a static lock, which is part
10996 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10997 appear in 0.9.8 or later. We EXPECT application authors to have
10998 dealt properly with this when 0.9.8 is released (unless we actually
10999 make such changes in the libcrypto locking code that changes will
11000 have to be made anyway).
11004 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11005 octets have been read, EOF or an error occurs. Without this change
11006 some truncated ASN1 structures will not produce an error.
11010 * Disable Heimdal support, since it hasn't been fully implemented.
11011 Still give the possibility to force the use of Heimdal, but with
11012 warnings and a request that patches get sent to openssl-dev.
11016 * Add the VC-CE target, introduce the WINCE sysname, and add
11017 INSTALL.WCE and appropriate conditionals to make it build.
11019 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11021 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11022 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11023 edit numbers of the version.
11025 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11027 * Introduce safe string copy and catenation functions
11028 (BUF_strlcpy() and BUF_strlcat()).
11030 *Ben Laurie (CHATS) and Richard Levitte*
11032 * Avoid using fixed-size buffers for one-line DNs.
11034 *Ben Laurie (CHATS)*
11036 * Add BUF_MEM_grow_clean() to avoid information leakage when
11037 resizing buffers containing secrets, and use where appropriate.
11039 *Ben Laurie (CHATS)*
11041 * Avoid using fixed size buffers for configuration file location.
11043 *Ben Laurie (CHATS)*
11045 * Avoid filename truncation for various CA files.
11047 *Ben Laurie (CHATS)*
11049 * Use sizeof in preference to magic numbers.
11051 *Ben Laurie (CHATS)*
11053 * Avoid filename truncation in cert requests.
11055 *Ben Laurie (CHATS)*
11057 * Add assertions to check for (supposedly impossible) buffer
11060 *Ben Laurie (CHATS)*
11062 * Don't cache truncated DNS entries in the local cache (this could
11063 potentially lead to a spoofing attack).
11065 *Ben Laurie (CHATS)*
11067 * Fix various buffers to be large enough for hex/decimal
11068 representations in a platform independent manner.
11070 *Ben Laurie (CHATS)*
11072 * Add CRYPTO_realloc_clean() to avoid information leakage when
11073 resizing buffers containing secrets, and use where appropriate.
11075 *Ben Laurie (CHATS)*
11077 * Add BIO_indent() to avoid much slightly worrying code to do
11080 *Ben Laurie (CHATS)*
11082 * Convert sprintf()/BIO_puts() to BIO_printf().
11084 *Ben Laurie (CHATS)*
11086 * buffer_gets() could terminate with the buffer only half
11089 *Ben Laurie (CHATS)*
11091 * Add assertions to prevent user-supplied crypto functions from
11092 overflowing internal buffers by having large block sizes, etc.
11094 *Ben Laurie (CHATS)*
11096 * New OPENSSL_assert() macro (similar to assert(), but enabled
11099 *Ben Laurie (CHATS)*
11101 * Eliminate unused copy of key in RC4.
11103 *Ben Laurie (CHATS)*
11105 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11107 *Ben Laurie (CHATS)*
11109 * Fix off-by-one error in EGD path.
11111 *Ben Laurie (CHATS)*
11113 * If RANDFILE path is too long, ignore instead of truncating.
11115 *Ben Laurie (CHATS)*
11117 * Eliminate unused and incorrectly sized X.509 structure
11120 *Ben Laurie (CHATS)*
11122 * Eliminate unused and dangerous function knumber().
11124 *Ben Laurie (CHATS)*
11126 * Eliminate unused and dangerous structure, KSSL_ERR.
11128 *Ben Laurie (CHATS)*
11130 * Protect against overlong session ID context length in an encoded
11131 session object. Since these are local, this does not appear to be
11134 *Ben Laurie (CHATS)*
11136 * Change from security patch (see 0.9.6e below) that did not affect
11137 the 0.9.6 release series:
11139 Remote buffer overflow in SSL3 protocol - an attacker could
11140 supply an oversized master key in Kerberos-enabled versions.
11143 *Ben Laurie (CHATS)*
11145 * Change the SSL kerb5 codes to match RFC 2712.
11149 * Make -nameopt work fully for req and add -reqopt switch.
11151 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11153 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11155 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11157 * Make sure tests can be performed even if the corresponding algorithms
11158 have been removed entirely. This was also the last step to make
11159 OpenSSL compilable with DJGPP under all reasonable conditions.
11161 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11163 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11164 to allow version independent disabling of normally unselected ciphers,
11165 which may be activated as a side-effect of selecting a single cipher.
11167 (E.g., cipher list string "RSA" enables ciphersuites that are left
11168 out of "ALL" because they do not provide symmetric encryption.
11169 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11171 *Lutz Jaenicke, Bodo Moeller*
11173 * Add appropriate support for separate platform-dependent build
11174 directories. The recommended way to make a platform-dependent
11175 build directory is the following (tested on Linux), maybe with
11178 # Place yourself outside of the OpenSSL source tree. In
11179 # this example, the environment variable OPENSSL_SOURCE
11180 # is assumed to contain the absolute OpenSSL source directory.
11181 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11182 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11183 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11184 mkdir -p `dirname $F`
11185 ln -s $OPENSSL_SOURCE/$F $F
11188 To be absolutely sure not to disturb the source tree, a "make clean"
11189 is a good thing. If it isn't successful, don't worry about it,
11190 it probably means the source directory is very clean.
11194 * Make sure any ENGINE control commands make local copies of string
11195 pointers passed to them whenever necessary. Otherwise it is possible
11196 the caller may have overwritten (or deallocated) the original string
11197 data when a later ENGINE operation tries to use the stored values.
11199 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11201 * Improve diagnostics in file reading and command-line digests.
11203 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11205 * Add AES modes CFB and OFB to the object database. Correct an
11206 error in AES-CFB decryption.
11210 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11211 allows existing EVP_CIPHER_CTX structures to be reused after
11212 calling `EVP_*Final()`. This behaviour is used by encryption
11213 BIOs and some applications. This has the side effect that
11214 applications must explicitly clean up cipher contexts with
11215 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11219 * Check the values of dna and dnb in bn_mul_recursive before calling
11220 bn_mul_comba (a non zero value means the a or b arrays do not contain
11221 n2 elements) and fallback to bn_mul_normal if either is not zero.
11225 * Fix escaping of non-ASCII characters when using the -subj option
11226 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11230 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11231 form for "surname", serialNumber has no short form.
11232 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11233 therefore remove "mail" short name for "internet 7".
11234 The OID for unique identifiers in X509 certificates is
11235 x500UniqueIdentifier, not uniqueIdentifier.
11236 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11240 * Add an "init" command to the ENGINE config module and auto initialize
11241 ENGINEs. Without any "init" command the ENGINE will be initialized
11242 after all ctrl commands have been executed on it. If init=1 the
11243 ENGINE is initialized at that point (ctrls before that point are run
11244 on the uninitialized ENGINE and after on the initialized one). If
11245 init=0 then the ENGINE will not be initialized at all.
11249 * Fix the 'app_verify_callback' interface so that the user-defined
11250 argument is actually passed to the callback: In the
11251 SSL_CTX_set_cert_verify_callback() prototype, the callback
11252 declaration has been changed from
11255 int (*cb)(X509_STORE_CTX *,void *);
11256 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11257 i=s->ctx->app_verify_callback(&ctx)
11258 has been changed into
11259 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11261 To update applications using SSL_CTX_set_cert_verify_callback(),
11262 a dummy argument can be added to their callback functions.
11264 *D. K. Smetters <smetters@parc.xerox.com>*
11266 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11268 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11270 * Add and OPENSSL_LOAD_CONF define which will cause
11271 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11272 This allows older applications to transparently support certain
11273 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11274 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11275 load the config file and OPENSSL_add_all_algorithms_conf() which will
11276 always load it have also been added.
11280 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11281 Adjust NIDs and EVP layer.
11283 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11285 * Config modules support in openssl utility.
11287 Most commands now load modules from the config file,
11288 though in a few (such as version) this isn't done
11289 because it couldn't be used for anything.
11291 In the case of ca and req the config file used is
11292 the same as the utility itself: that is the -config
11293 command line option can be used to specify an
11298 * Move default behaviour from OPENSSL_config(). If appname is NULL
11299 use "openssl_conf" if filename is NULL use default openssl config file.
11303 * Add an argument to OPENSSL_config() to allow the use of an alternative
11304 config section name. Add a new flag to tolerate a missing config file
11305 and move code to CONF_modules_load_file().
11309 * Support for crypto accelerator cards from Accelerated Encryption
11310 Processing, www.aep.ie. (Use engine 'aep')
11311 The support was copied from 0.9.6c [engine] and adapted/corrected
11312 to work with the new engine framework.
11314 *AEP Inc. and Richard Levitte*
11316 * Support for SureWare crypto accelerator cards from Baltimore
11317 Technologies. (Use engine 'sureware')
11318 The support was copied from 0.9.6c [engine] and adapted
11319 to work with the new engine framework.
11323 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11324 make the newer ENGINE framework commands for the CHIL engine work.
11326 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11328 * Make it possible to produce shared libraries on ReliantUNIX.
11330 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11332 * Add the configuration target debug-linux-ppro.
11333 Make 'openssl rsa' use the general key loading routines
11334 implemented in apps.c, and make those routines able to
11335 handle the key format FORMAT_NETSCAPE and the variant
11338 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11340 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11342 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11344 * Add -keyform to rsautl, and document -engine.
11346 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11348 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11349 BIO_R_NO_SUCH_FILE error code rather than the generic
11350 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11354 * Add new functions
11355 ERR_peek_last_error
11356 ERR_peek_last_error_line
11357 ERR_peek_last_error_line_data.
11358 These are similar to
11360 ERR_peek_error_line
11361 ERR_peek_error_line_data,
11362 but report on the latest error recorded rather than the first one
11363 still in the error queue.
11365 *Ben Laurie, Bodo Moeller*
11367 * default_algorithms option in ENGINE config module. This allows things
11369 default_algorithms = ALL
11370 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11374 * Preliminary ENGINE config module.
11378 * New experimental application configuration code.
11382 * Change the AES code to follow the same name structure as all other
11383 symmetric ciphers, and behave the same way. Move everything to
11384 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11386 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11388 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11390 *Ben Laurie and Theo de Raadt*
11392 * Add option to output public keys in req command.
11394 *Massimiliano Pala madwolf@openca.org*
11396 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11397 (up to about 10% better than before for P-192 and P-224).
11401 * New functions/macros
11403 SSL_CTX_set_msg_callback(ctx, cb)
11404 SSL_CTX_set_msg_callback_arg(ctx, arg)
11405 SSL_set_msg_callback(ssl, cb)
11406 SSL_set_msg_callback_arg(ssl, arg)
11408 to request calling a callback function
11410 void cb(int write_p, int version, int content_type,
11411 const void *buf, size_t len, SSL *ssl, void *arg)
11413 whenever a protocol message has been completely received
11414 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11415 protocol version according to which the SSL library interprets
11416 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11417 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11418 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11419 specification (change_cipher_spec(20), alert(21), handshake(22)).
11420 'buf' and 'len' point to the actual message, 'ssl' to the
11421 SSL object, and 'arg' is the application-defined value set by
11422 SSL[_CTX]_set_msg_callback_arg().
11424 'openssl s_client' and 'openssl s_server' have new '-msg' options
11425 to enable a callback that displays all protocol messages.
11429 * Change the shared library support so shared libraries are built as
11430 soon as the corresponding static library is finished, and thereby get
11431 openssl and the test programs linked against the shared library.
11432 This still only happens when the keyword "shard" has been given to
11433 the configuration scripts.
11435 NOTE: shared library support is still an experimental thing, and
11436 backward binary compatibility is still not guaranteed.
11438 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11440 * Add support for Subject Information Access extension.
11442 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11444 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11445 additional bytes when new memory had to be allocated, not just
11446 when reusing an existing buffer.
11450 * New command line and configuration option 'utf8' for the req command.
11451 This allows field values to be specified as UTF8 strings.
11455 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11456 runs for the former and machine-readable output for the latter.
11460 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11461 of the e-mail address in the DN (i.e., it will go into a certificate
11462 extension only). The new configuration file option 'email_in_dn = no'
11463 has the same effect.
11465 *Massimiliano Pala madwolf@openca.org*
11467 * Change all functions with names starting with `des_` to be starting
11468 with `DES_` instead. Add wrappers that are compatible with libdes,
11469 but are named `_ossl_old_des_*`. Finally, add macros that map the
11470 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11471 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11472 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11475 Since we provide two compatibility mappings, the user needs to
11476 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11477 compatibility is desired. The default (i.e., when that macro
11478 isn't defined) is OpenSSL 0.9.6c compatibility.
11480 There are also macros that enable and disable the support of old
11481 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11482 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11483 are defined, the default will apply: to support the old des routines.
11485 In either case, one must include openssl/des.h to get the correct
11486 definitions. Do not try to just include openssl/des_old.h, that
11489 NOTE: This is a major break of an old API into a new one. Software
11490 authors are encouraged to switch to the `DES_` style functions. Some
11491 time in the future, des_old.h and the libdes compatibility functions
11492 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11493 default), and then completely removed.
11497 * Test for certificates which contain unsupported critical extensions.
11498 If such a certificate is found during a verify operation it is
11499 rejected by default: this behaviour can be overridden by either
11500 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11501 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11502 X509_supported_extension() has also been added which returns 1 if a
11503 particular extension is supported.
11507 * Modify the behaviour of EVP cipher functions in similar way to digests
11508 to retain compatibility with existing code.
11512 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11513 compatibility with existing code. In particular the 'ctx' parameter does
11514 not have to be to be initialized before the call to EVP_DigestInit() and
11515 it is tidied up after a call to EVP_DigestFinal(). New function
11516 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11517 EVP_MD_CTX_copy() changed to not require the destination to be
11518 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11519 requires the destination to be valid.
11521 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11522 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11526 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11527 so that complete 'Handshake' protocol structures are kept in memory
11528 instead of overwriting 'msg_type' and 'length' with 'body' data.
11532 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11534 *Massimo Santin via Richard Levitte*
11536 * Major restructuring to the underlying ENGINE code. This includes
11537 reduction of linker bloat, separation of pure "ENGINE" manipulation
11538 (initialisation, etc) from functionality dealing with implementations
11539 of specific crypto interfaces. This change also introduces integrated
11540 support for symmetric ciphers and digest implementations - so ENGINEs
11541 can now accelerate these by providing EVP_CIPHER and EVP_MD
11542 implementations of their own. This is detailed in crypto/engine/README
11543 as it couldn't be adequately described here. However, there are a few
11544 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11545 were changed in the original introduction of ENGINE code have now
11546 reverted back - the hooking from this code to ENGINE is now a good
11547 deal more passive and at run-time, operations deal directly with
11548 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11549 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11550 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11551 they were not being used by the framework as there is no concept of a
11552 BIGNUM_METHOD and they could not be generalised to the new
11553 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11554 ENGINE_cpy() has been removed as it cannot be consistently defined in
11559 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11563 * Change mkdef.pl to sort symbols that get the same entry number,
11564 and make sure the automatically generated functions `ERR_load_*`
11565 become part of libeay.num as well.
11569 * New function SSL_renegotiate_pending(). This returns true once
11570 renegotiation has been requested (either SSL_renegotiate() call
11571 or HelloRequest/ClientHello received from the peer) and becomes
11572 false once a handshake has been completed.
11573 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11574 sends a HelloRequest, but does not ensure that a handshake takes
11575 place. SSL_renegotiate_pending() is useful for checking if the
11576 client has followed the request.)
11580 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11581 By default, clients may request session resumption even during
11582 renegotiation (if session ID contexts permit); with this option,
11583 session resumption is possible only in the first handshake.
11585 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11586 more bits available for options that should not be part of
11587 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11591 * Add some demos for certificate and certificate request creation.
11595 * Make maximum certificate chain size accepted from the peer application
11596 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11597 "Douglas E. Engert" <deengert@anl.gov>.
11601 * Add support for shared libraries for Unixware-7
11602 (Boyd Lynn Gerber <gerberb@zenez.com>).
11606 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11607 be done prior to destruction. Use this to unload error strings from
11608 ENGINEs that load their own error strings. NB: This adds two new API
11609 functions to "get" and "set" this destroy handler in an ENGINE.
11613 * Alter all existing ENGINE implementations (except "openssl" and
11614 "openbsd") to dynamically instantiate their own error strings. This
11615 makes them more flexible to be built both as statically-linked ENGINEs
11616 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11617 Also, add stub code to each that makes building them as self-contained
11618 shared-libraries easier (see README.ENGINE).
11622 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11623 implementations into applications that are completely implemented in
11624 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11625 commands that can be used to configure what shared-library to load and
11626 to control aspects of the way it is handled. Also, made an update to
11627 the README.ENGINE file that brings its information up-to-date and
11628 provides some information and instructions on the "dynamic" ENGINE
11629 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11633 * Make it possible to unload ranges of ERR strings with a new
11634 "ERR_unload_strings" function.
11638 * Add a copy() function to EVP_MD.
11642 * Make EVP_MD routines take a context pointer instead of just the
11643 md_data void pointer.
11647 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11648 that the digest can only process a single chunk of data
11649 (typically because it is provided by a piece of
11650 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11651 is only going to provide a single chunk of data, and hence the
11652 framework needn't accumulate the data for oneshot drivers.
11656 * As with "ERR", make it possible to replace the underlying "ex_data"
11657 functions. This change also alters the storage and management of global
11658 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11659 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11660 index counters. The API functions that use this state have been changed
11661 to take a "class_index" rather than pointers to the class's local STACK
11662 and counter, and there is now an API function to dynamically create new
11663 classes. This centralisation allows us to (a) plug a lot of the
11664 thread-safety problems that existed, and (b) makes it possible to clean
11665 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11666 such data would previously have always leaked in application code and
11667 workarounds were in place to make the memory debugging turn a blind eye
11668 to it. Application code that doesn't use this new function will still
11669 leak as before, but their memory debugging output will announce it now
11670 rather than letting it slide.
11672 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11673 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11674 has a return value to indicate success or failure.
11678 * Make it possible to replace the underlying "ERR" functions such that the
11679 global state (2 LHASH tables and 2 locks) is only used by the "default"
11680 implementation. This change also adds two functions to "get" and "set"
11681 the implementation prior to it being automatically set the first time
11682 any other ERR function takes place. Ie. an application can call "get",
11683 pass the return value to a module it has just loaded, and that module
11684 can call its own "set" function using that value. This means the
11685 module's "ERR" operations will use (and modify) the error state in the
11686 application and not in its own statically linked copy of OpenSSL code.
11690 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11691 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11692 the operation, and provides a more encapsulated way for external code
11693 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11694 to use these functions rather than manually incrementing the counts.
11696 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11700 * Add EVP test program.
11704 * Add symmetric cipher support to ENGINE. Expect the API to change!
11708 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11709 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11710 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11711 These allow a CRL to be built without having to access X509_CRL fields
11712 directly. Modify 'ca' application to use new functions.
11716 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11717 bug workarounds. Rollback attack detection is a security feature.
11718 The problem will only arise on OpenSSL servers when TLSv1 is not
11719 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11720 Software authors not wanting to support TLSv1 will have special reasons
11721 for their choice and can explicitly enable this option.
11723 *Bodo Moeller, Lutz Jaenicke*
11725 * Rationalise EVP so it can be extended: don't include a union of
11726 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11727 (similar to those existing for EVP_CIPHER_CTX).
11732 EVP_MD_CTX_init(&md); /* new function call */
11733 EVP_DigestInit(&md, EVP_sha1());
11734 EVP_DigestUpdate(&md, in, len);
11735 EVP_DigestFinal(&md, out, NULL);
11736 EVP_MD_CTX_cleanup(&md); /* new function call */
11740 * Make DES key schedule conform to the usual scheme, as well as
11741 correcting its structure. This means that calls to DES functions
11742 now have to pass a pointer to a des_key_schedule instead of a
11743 plain des_key_schedule (which was actually always a pointer
11746 des_key_schedule ks;
11748 des_set_key_checked(..., &ks);
11749 des_ncbc_encrypt(..., &ks, ...);
11751 (Note that a later change renames 'des_...' into 'DES_...'.)
11755 * Initial reduction of linker bloat: the use of some functions, such as
11756 PEM causes large amounts of unused functions to be linked in due to
11757 poor organisation. For example pem_all.c contains every PEM function
11758 which has a knock on effect of linking in large amounts of (unused)
11759 ASN1 code. Grouping together similar functions and splitting unrelated
11760 functions prevents this.
11764 * Cleanup of EVP macros.
11768 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11769 correct `_ecb suffix`.
11773 * Add initial OCSP responder support to ocsp application. The
11774 revocation information is handled using the text based index
11775 use by the ca application. The responder can either handle
11776 requests generated internally, supplied in files (for example
11777 via a CGI script) or using an internal minimal server.
11781 * Add configuration choices to get zlib compression for TLS.
11785 * Changes to Kerberos SSL for RFC 2712 compliance:
11786 1. Implemented real KerberosWrapper, instead of just using
11787 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11788 2. Implemented optional authenticator field of KerberosWrapper.
11790 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11791 and authenticator structs; see crypto/krb5/.
11793 Generalized Kerberos calls to support multiple Kerberos libraries.
11794 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11795 via Richard Levitte*
11797 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11798 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11799 values for each of the key sizes rather than having just
11800 parameters (and 'speed' generating keys each time).
11804 * Speed up EVP routines.
11807 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11808 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11809 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11810 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11812 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11813 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11814 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11817 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11819 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11823 * Added the OS2-EMX target.
11825 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11827 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11828 to support NCONF routines in extension code. New function CONF_set_nconf()
11829 to allow functions which take an NCONF to also handle the old LHASH
11830 structure: this means that the old CONF compatible routines can be
11831 retained (in particular wrt extensions) without having to duplicate the
11832 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11836 * Enhance the general user interface with mechanisms for inner control
11837 and with possibilities to have yes/no kind of prompts.
11841 * Change all calls to low level digest routines in the library and
11842 applications to use EVP. Add missing calls to HMAC_cleanup() and
11843 don't assume HMAC_CTX can be copied using memcpy().
11845 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11847 * Add the possibility to control engines through control names but with
11848 arbitrary arguments instead of just a string.
11849 Change the key loaders to take a UI_METHOD instead of a callback
11850 function pointer. NOTE: this breaks binary compatibility with earlier
11851 versions of OpenSSL [engine].
11852 Adapt the nCipher code for these new conditions and add a card insertion
11857 * Enhance the general user interface with mechanisms to better support
11858 dialog box interfaces, application-defined prompts, the possibility
11859 to use defaults (for example default passwords from somewhere else)
11860 and interrupts/cancellations.
11864 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11865 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11869 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11870 tidy up some unnecessarily weird code in 'sk_new()').
11872 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11874 * Change the key loading routines for ENGINEs to use the same kind
11875 callback (pem_password_cb) as all other routines that need this
11880 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11881 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11882 than this minimum value is recommended.
11886 * New random seeder for OpenVMS, using the system process statistics
11887 that are easily reachable.
11891 * Windows apparently can't transparently handle global
11892 variables defined in DLLs. Initialisations such as:
11894 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11896 won't compile. This is used by the any applications that need to
11897 declare their own ASN1 modules. This was fixed by adding the option
11898 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11899 needed for static libraries under Win32.
11903 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11904 setting of purpose and trust fields. New X509_STORE trust and
11905 purpose functions and tidy up setting in other SSL functions.
11909 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11910 structure. These are inherited by X509_STORE_CTX when it is
11911 initialised. This allows various defaults to be set in the
11912 X509_STORE structure (such as flags for CRL checking and custom
11913 purpose or trust settings) for functions which only use X509_STORE_CTX
11914 internally such as S/MIME.
11916 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11917 trust settings if they are not set in X509_STORE. This allows X509_STORE
11918 purposes and trust (in S/MIME for example) to override any set by default.
11920 Add command line options for CRL checking to smime, s_client and s_server
11925 * Initial CRL based revocation checking. If the CRL checking flag(s)
11926 are set then the CRL is looked up in the X509_STORE structure and
11927 its validity and signature checked, then if the certificate is found
11928 in the CRL the verify fails with a revoked error.
11930 Various new CRL related callbacks added to X509_STORE_CTX structure.
11932 Command line options added to 'verify' application to support this.
11934 This needs some additional work, such as being able to handle multiple
11935 CRLs with different times, extension based lookup (rather than just
11936 by subject name) and ultimately more complete V2 CRL extension
11941 * Add a general user interface API (crypto/ui/). This is designed
11942 to replace things like des_read_password and friends (backward
11943 compatibility functions using this new API are provided).
11944 The purpose is to remove prompting functions from the DES code
11945 section as well as provide for prompting through dialog boxes in
11946 a window system and the like.
11950 * Add "ex_data" support to ENGINE so implementations can add state at a
11951 per-structure level rather than having to store it globally.
11955 * Make it possible for ENGINE structures to be copied when retrieved by
11956 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11957 This causes the "original" ENGINE structure to act like a template,
11958 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11959 operational state can be localised to each ENGINE structure, despite the
11960 fact they all share the same "methods". New ENGINE structures returned in
11961 this case have no functional references and the return value is the single
11962 structural reference. This matches the single structural reference returned
11963 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11968 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11969 needs to match any other type at all we need to manually clear the
11974 * Changes to the "openssl engine" utility to include;
11975 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11976 about an ENGINE's available control commands.
11977 - executing control commands from command line arguments using the
11978 '-pre' and '-post' switches. '-post' is only used if '-t' is
11979 specified and the ENGINE is successfully initialised. The syntax for
11980 the individual commands are colon-separated, for example;
11981 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11985 * New dynamic control command support for ENGINEs. ENGINEs can now
11986 declare their own commands (numbers), names (strings), descriptions,
11987 and input types for run-time discovery by calling applications. A
11988 subset of these commands are implicitly classed as "executable"
11989 depending on their input type, and only these can be invoked through
11990 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11991 can be based on user input, config files, etc). The distinction is
11992 that "executable" commands cannot return anything other than a boolean
11993 result and can only support numeric or string input, whereas some
11994 discoverable commands may only be for direct use through
11995 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11996 pointers, or other custom uses. The "executable" commands are to
11997 support parameterisations of ENGINE behaviour that can be
11998 unambiguously defined by ENGINEs and used consistently across any
11999 OpenSSL-based application. Commands have been added to all the
12000 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12001 control over shared-library paths without source code alterations.
12005 * Changed all ENGINE implementations to dynamically allocate their
12006 ENGINEs rather than declaring them statically. Apart from this being
12007 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12008 this also allows the implementations to compile without using the
12009 internal engine_int.h header.
12013 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12014 'const' value. Any code that should be able to modify a RAND_METHOD
12015 should already have non-const pointers to it (ie. they should only
12016 modify their own ones).
12020 * Made a variety of little tweaks to the ENGINE code.
12021 - "atalla" and "ubsec" string definitions were moved from header files
12022 to C code. "nuron" string definitions were placed in variables
12023 rather than hard-coded - allowing parameterisation of these values
12024 later on via ctrl() commands.
12025 - Removed unused "#if 0"'d code.
12026 - Fixed engine list iteration code so it uses ENGINE_free() to release
12027 structural references.
12028 - Constified the RAND_METHOD element of ENGINE structures.
12029 - Constified various get/set functions as appropriate and added
12030 missing functions (including a catch-all ENGINE_cpy that duplicates
12031 all ENGINE values onto a new ENGINE except reference counts/state).
12032 - Removed NULL parameter checks in get/set functions. Setting a method
12033 or function to NULL is a way of cancelling out a previously set
12034 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12035 and doesn't justify the extra error symbols and code.
12036 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12037 flags from engine_int.h to engine.h.
12038 - Changed prototypes for ENGINE handler functions (init(), finish(),
12039 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12043 * Implement binary inversion algorithm for BN_mod_inverse in addition
12044 to the algorithm using long division. The binary algorithm can be
12045 used only if the modulus is odd. On 32-bit systems, it is faster
12046 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12047 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12048 up to 450 bits. In 64-bit environments, the binary algorithm
12049 appears to be advantageous for much longer moduli; here we use it
12050 for moduli up to 2048 bits.
12054 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12055 could not support the combine flag in choice fields.
12059 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12060 extensions from a certificate request to the certificate.
12064 * Allow multiple 'certopt' and 'nameopt' options to be separated
12065 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12066 file: this allows the display of the certificate about to be
12067 signed to be customised, to allow certain fields to be included
12068 or excluded and extension details. The old system didn't display
12069 multicharacter strings properly, omitted fields not in the policy
12070 and couldn't display additional details such as extensions.
12074 * Function EC_POINTs_mul for multiple scalar multiplication
12075 of an arbitrary number of elliptic curve points
12076 \sum scalars[i]*points[i],
12077 optionally including the generator defined for the EC_GROUP:
12078 scalar*generator + \sum scalars[i]*points[i].
12080 EC_POINT_mul is a simple wrapper function for the typical case
12081 that the point list has just one item (besides the optional
12086 * First EC_METHODs for curves over GF(p):
12088 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12089 operations and provides various method functions that can also
12090 operate with faster implementations of modular arithmetic.
12092 EC_GFp_mont_method() reuses most functions that are part of
12093 EC_GFp_simple_method, but uses Montgomery arithmetic.
12095 *Bodo Moeller; point addition and point doubling
12096 implementation directly derived from source code provided by
12097 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12099 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12100 crypto/ec/ec_lib.c):
12102 Curves are EC_GROUP objects (with an optional group generator)
12103 based on EC_METHODs that are built into the library.
12105 Points are EC_POINT objects based on EC_GROUP objects.
12107 Most of the framework would be able to handle curves over arbitrary
12108 finite fields, but as there are no obvious types for fields other
12109 than GF(p), some functions are limited to that for now.
12113 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12114 that the file contains a complete HTTP response.
12118 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12119 change the def and num file printf format specifier from "%-40sXXX"
12120 to "%-39s XXX". The latter will always guarantee a space after the
12121 field while the former will cause them to run together if the field
12122 is 40 of more characters long.
12126 * Constify the cipher and digest 'method' functions and structures
12127 and modify related functions to take constant EVP_MD and EVP_CIPHER
12132 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12133 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12137 * Modify `EVP_Digest*()` routines so they now return values. Although the
12138 internal software routines can never fail additional hardware versions
12143 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12145 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12146 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12149 ERR_R_NESTED_ASN1_ERROR
12151 ERR_R_MISSING_ASN1_EOS
12152 were 4 .. 9, conflicting with
12153 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12155 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12156 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12158 Add new error code 'ERR_R_INTERNAL_ERROR'.
12162 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12167 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12168 sets the subject name for a new request or supersedes the
12169 subject name in a given request. Formats that can be parsed are
12170 'CN=Some Name, OU=myOU, C=IT'
12172 'CN=Some Name/OU=myOU/C=IT'.
12174 Add options '-batch' and '-verbose' to 'openssl req'.
12176 *Massimiliano Pala <madwolf@hackmasters.net>*
12178 * Introduce the possibility to access global variables through
12179 functions on platform were that's the best way to handle exporting
12180 global variables in shared libraries. To enable this functionality,
12181 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12182 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12183 is normally done by Configure or something similar).
12185 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12186 in the source file (foo.c) like this:
12188 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12189 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12191 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12192 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12194 OPENSSL_DECLARE_GLOBAL(int,foo);
12195 #define foo OPENSSL_GLOBAL_REF(foo)
12196 OPENSSL_DECLARE_GLOBAL(double,bar);
12197 #define bar OPENSSL_GLOBAL_REF(bar)
12199 The #defines are very important, and therefore so is including the
12200 header file everywhere where the defined globals are used.
12202 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12203 of ASN.1 items, but that structure is a bit different.
12205 The largest change is in util/mkdef.pl which has been enhanced with
12206 better and easier to understand logic to choose which symbols should
12207 go into the Windows .def files as well as a number of fixes and code
12208 cleanup (among others, algorithm keywords are now sorted
12209 lexicographically to avoid constant rewrites).
12213 * In BN_div() keep a copy of the sign of 'num' before writing the
12214 result to 'rm' because if rm==num the value will be overwritten
12215 and produce the wrong result if 'num' is negative: this caused
12216 problems with BN_mod() and BN_nnmod().
12220 * Function OCSP_request_verify(). This checks the signature on an
12221 OCSP request and verifies the signer certificate. The signer
12222 certificate is just checked for a generic purpose and OCSP request
12227 * Add OCSP_check_validity() function to check the validity of OCSP
12228 responses. OCSP responses are prepared in real time and may only
12229 be a few seconds old. Simply checking that the current time lies
12230 between thisUpdate and nextUpdate max reject otherwise valid responses
12231 caused by either OCSP responder or client clock inaccuracy. Instead
12232 we allow thisUpdate and nextUpdate to fall within a certain period of
12233 the current time. The age of the response can also optionally be
12234 checked. Two new options -validity_period and -status_age added to
12239 * If signature or public key algorithm is unrecognized print out its
12240 OID rather that just UNKNOWN.
12244 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12245 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12246 ID to be generated from the issuer certificate alone which can then be
12247 passed to OCSP_id_issuer_cmp().
12251 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12252 ASN1 modules to export functions returning ASN1_ITEM pointers
12253 instead of the ASN1_ITEM structures themselves. This adds several
12254 new macros which allow the underlying ASN1 function/structure to
12255 be accessed transparently. As a result code should not use ASN1_ITEM
12256 references directly (such as &X509_it) but instead use the relevant
12257 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12258 use of the new ASN1 code on platforms where exporting structures
12259 is problematical (for example in shared libraries) but exporting
12260 functions returning pointers to structures is not.
12264 * Add support for overriding the generation of SSL/TLS session IDs.
12265 These callbacks can be registered either in an SSL_CTX or per SSL.
12266 The purpose of this is to allow applications to control, if they wish,
12267 the arbitrary values chosen for use as session IDs, particularly as it
12268 can be useful for session caching in multiple-server environments. A
12269 command-line switch for testing this (and any client code that wishes
12270 to use such a feature) has been added to "s_server".
12272 *Geoff Thorpe, Lutz Jaenicke*
12274 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12275 of the form '#if defined(...) || defined(...) || ...' and
12276 '#if !defined(...) && !defined(...) && ...'. This also avoids
12277 the growing number of special cases it was previously handling.
12281 * Make all configuration macros available for application by making
12282 sure they are available in opensslconf.h, by giving them names starting
12283 with `OPENSSL_` to avoid conflicts with other packages and by making
12284 sure e_os2.h will cover all platform-specific cases together with
12286 Additionally, it is now possible to define configuration/platform-
12287 specific names (called "system identities"). In the C code, these
12288 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12289 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12290 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12295 * New option -set_serial to 'req' and 'x509' this allows the serial
12296 number to use to be specified on the command line. Previously self
12297 signed certificates were hard coded with serial number 0 and the
12298 CA options of 'x509' had to use a serial number in a file which was
12303 * New options to 'ca' utility to support V2 CRL entry extensions.
12304 Currently CRL reason, invalidity date and hold instruction are
12305 supported. Add new CRL extensions to V3 code and some new objects.
12309 * New function EVP_CIPHER_CTX_set_padding() this is used to
12310 disable standard block padding (aka PKCS#5 padding) in the EVP
12311 API, which was previously mandatory. This means that the data is
12312 not padded in any way and so the total length much be a multiple
12313 of the block size, otherwise an error occurs.
12317 * Initial (incomplete) OCSP SSL support.
12321 * New function OCSP_parse_url(). This splits up a URL into its host,
12322 port and path components: primarily to parse OCSP URLs. New -url
12323 option to ocsp utility.
12327 * New nonce behavior. The return value of OCSP_check_nonce() now
12328 reflects the various checks performed. Applications can decide
12329 whether to tolerate certain situations such as an absent nonce
12330 in a response when one was present in a request: the ocsp application
12331 just prints out a warning. New function OCSP_add1_basic_nonce()
12332 this is to allow responders to include a nonce in a response even if
12333 the request is nonce-less.
12337 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12338 skipped when using openssl x509 multiple times on a single input file,
12339 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12343 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12344 set string type: to handle setting ASN1_TIME structures. Fix ca
12345 utility to correctly initialize revocation date of CRLs.
12349 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12350 the clients preferred ciphersuites and rather use its own preferences.
12351 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12352 Internet Explorer by ensuring unchanged hash method during stepup.
12353 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12357 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12358 to aes and add a new 'exist' option to print out symbols that don't
12363 * Additional options to ocsp utility to allow flags to be set and
12364 additional certificates supplied.
12368 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12369 OCSP client a number of certificate to only verify the response
12374 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12375 handle the new API. Currently only ECB, CBC modes supported. Add new
12378 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12379 Encryption Standard (AES) Ciphersuites for Transport Layer
12380 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12381 not enabled by default and were not part of the "ALL" ciphersuite
12382 alias because they were not yet official; they could be
12383 explicitly requested by specifying the "AESdraft" ciphersuite
12384 group alias. In the final release of OpenSSL 0.9.7, the group
12385 alias is called "AES" and is part of "ALL".)
12387 *Ben Laurie, Steve Henson, Bodo Moeller*
12389 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12390 request to response.
12394 * Functions for OCSP responders. OCSP_request_onereq_count(),
12395 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12396 extract information from a certificate request. OCSP_response_create()
12397 creates a response and optionally adds a basic response structure.
12398 OCSP_basic_add1_status() adds a complete single response to a basic
12399 response and returns the OCSP_SINGLERESP structure just added (to allow
12400 extensions to be included for example). OCSP_basic_add1_cert() adds a
12401 certificate to a basic response and OCSP_basic_sign() signs a basic
12402 response with various flags. New helper functions ASN1_TIME_check()
12403 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12404 (converts ASN1_TIME to GeneralizedTime).
12408 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12409 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12410 structure from a certificate. X509_pubkey_digest() digests the public_key
12411 contents: this is used in various key identifiers.
12415 * Make sk_sort() tolerate a NULL argument.
12417 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12419 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12420 passed by the function are trusted implicitly. If any of them signed the
12421 response then it is assumed to be valid and is not verified.
12425 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12426 to data. This was previously part of the PKCS7 ASN1 code. This
12427 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12428 *Steve Henson, reported by Kenneth R. Robinette
12429 <support@securenetterm.com>*
12431 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12432 routines: without these tracing memory leaks is very painful.
12433 Fix leaks in PKCS12 and PKCS7 routines.
12437 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12438 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12439 effectively meant GeneralizedTime would never be used. Now it
12440 is initialised to -1 but X509_time_adj() now has to check the value
12441 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12442 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12443 *Steve Henson, reported by Kenneth R. Robinette
12444 <support@securenetterm.com>*
12446 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12447 result in a zero length in the ASN1_INTEGER structure which was
12448 not consistent with the structure when d2i_ASN1_INTEGER() was used
12449 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12450 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12451 where it did not print out a minus for negative ASN1_INTEGER.
12455 * Add summary printout to ocsp utility. The various functions which
12456 convert status values to strings have been renamed to:
12457 OCSP_response_status_str(), OCSP_cert_status_str() and
12458 OCSP_crl_reason_str() and are no longer static. New options
12459 to verify nonce values and to disable verification. OCSP response
12460 printout format cleaned up.
12464 * Add additional OCSP certificate checks. These are those specified
12465 in RFC2560. This consists of two separate checks: the CA of the
12466 certificate being checked must either be the OCSP signer certificate
12467 or the issuer of the OCSP signer certificate. In the latter case the
12468 OCSP signer certificate must contain the OCSP signing extended key
12469 usage. This check is performed by attempting to match the OCSP
12470 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12471 in the OCSP_CERTID structures of the response.
12475 * Initial OCSP certificate verification added to OCSP_basic_verify()
12476 and related routines. This uses the standard OpenSSL certificate
12477 verify routines to perform initial checks (just CA validity) and
12478 to obtain the certificate chain. Then additional checks will be
12479 performed on the chain. Currently the root CA is checked to see
12480 if it is explicitly trusted for OCSP signing. This is used to set
12481 a root CA as a global signing root: that is any certificate that
12482 chains to that CA is an acceptable OCSP signing certificate.
12486 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12487 extensions from a separate configuration file.
12488 As when reading extensions from the main configuration file,
12489 the '-extensions ...' option may be used for specifying the
12492 *Massimiliano Pala <madwolf@comune.modena.it>*
12494 * New OCSP utility. Allows OCSP requests to be generated or
12495 read. The request can be sent to a responder and the output
12496 parsed, outputted or printed in text form. Not complete yet:
12497 still needs to check the OCSP response validity.
12501 * New subcommands for 'openssl ca':
12502 `openssl ca -status <serial>` prints the status of the cert with
12503 the given serial number (according to the index file).
12504 `openssl ca -updatedb` updates the expiry status of certificates
12507 *Massimiliano Pala <madwolf@comune.modena.it>*
12509 * New '-newreq-nodes' command option to CA.pl. This is like
12510 '-newreq', but calls 'openssl req' with the '-nodes' option
12511 so that the resulting key is not encrypted.
12513 *Damien Miller <djm@mindrot.org>*
12515 * New configuration for the GNU Hurd.
12517 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12519 * Initial code to implement OCSP basic response verify. This
12520 is currently incomplete. Currently just finds the signer's
12521 certificate and verifies the signature on the response.
12525 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12526 value of OPENSSLDIR. This is available via the new '-d' option
12527 to 'openssl version', and is also included in 'openssl version -a'.
12531 * Allowing defining memory allocation callbacks that will be given
12532 file name and line number information in additional arguments
12533 (a `const char*` and an int). The basic functionality remains, as
12534 well as the original possibility to just replace malloc(),
12535 realloc() and free() by functions that do not know about these
12536 additional arguments. To register and find out the current
12537 settings for extended allocation functions, the following
12538 functions are provided:
12540 CRYPTO_set_mem_ex_functions
12541 CRYPTO_set_locked_mem_ex_functions
12542 CRYPTO_get_mem_ex_functions
12543 CRYPTO_get_locked_mem_ex_functions
12545 These work the same way as CRYPTO_set_mem_functions and friends.
12546 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12547 extended allocation function is enabled.
12548 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12549 a conventional allocation function is enabled.
12551 *Richard Levitte, Bodo Moeller*
12553 * Finish off removing the remaining LHASH function pointer casts.
12554 There should no longer be any prototype-casting required when using
12555 the LHASH abstraction, and any casts that remain are "bugs". See
12556 the callback types and macros at the head of lhash.h for details
12557 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12561 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12562 If /dev/[u]random devices are not available or do not return enough
12563 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12565 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12566 /etc/entropy will be queried once each in this sequence, querying stops
12567 when enough entropy was collected without querying more sockets.
12571 * Change the Unix RAND_poll() variant to be able to poll several
12572 random devices, as specified by DEVRANDOM, until a sufficient amount
12573 of data has been collected. We spend at most 10 ms on each file
12574 (select timeout) and read in non-blocking mode. DEVRANDOM now
12575 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12576 (previously it was just the string "/dev/urandom"), so on typical
12577 platforms the 10 ms delay will never occur.
12578 Also separate out the Unix variant to its own file, rand_unix.c.
12579 For VMS, there's a currently-empty rand_vms.c.
12583 * Move OCSP client related routines to ocsp_cl.c. These
12584 provide utility functions which an application needing
12585 to issue a request to an OCSP responder and analyse the
12586 response will typically need: as opposed to those which an
12587 OCSP responder itself would need which will be added later.
12589 OCSP_request_sign() signs an OCSP request with an API similar
12590 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12591 response. OCSP_response_get1_basic() extracts basic response
12592 from response. OCSP_resp_find_status(): finds and extracts status
12593 information from an OCSP_CERTID structure (which will be created
12594 when the request structure is built). These are built from lower
12595 level functions which work on OCSP_SINGLERESP structures but
12596 won't normally be used unless the application wishes to examine
12597 extensions in the OCSP response for example.
12599 Replace nonce routines with a pair of functions.
12600 OCSP_request_add1_nonce() adds a nonce value and optionally
12601 generates a random value. OCSP_check_nonce() checks the
12602 validity of the nonce in an OCSP response.
12606 * Change function OCSP_request_add() to OCSP_request_add0_id().
12607 This doesn't copy the supplied OCSP_CERTID and avoids the
12608 need to free up the newly created id. Change return type
12609 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12610 This can then be used to add extensions to the request.
12611 Deleted OCSP_request_new(), since most of its functionality
12612 is now in OCSP_REQUEST_new() (and the case insensitive name
12613 clash) apart from the ability to set the request name which
12614 will be added elsewhere.
12618 * Update OCSP API. Remove obsolete extensions argument from
12619 various functions. Extensions are now handled using the new
12620 OCSP extension code. New simple OCSP HTTP function which
12621 can be used to send requests and parse the response.
12625 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12626 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12627 uses the special reorder version of SET OF to sort the attributes
12628 and reorder them to match the encoded order. This resolves a long
12629 standing problem: a verify on a PKCS7 structure just after signing
12630 it used to fail because the attribute order did not match the
12631 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12632 it uses the received order. This is necessary to tolerate some broken
12633 software that does not order SET OF. This is handled by encoding
12634 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12635 to produce the required SET OF.
12639 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12640 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12641 files to get correct declarations of the ASN.1 item variables.
12645 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12646 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12647 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12648 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12649 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12650 ASN1_ITEM and no wrapper functions.
12654 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12655 replace the old function pointer based I/O routines. Change most of
12656 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12660 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12661 lines, recognize more "algorithms" that can be deselected, and make
12662 it complain about algorithm deselection that isn't recognised.
12666 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12667 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12668 to use new functions. Add NO_ASN1_OLD which can be set to remove
12669 some old style ASN1 functions: this can be used to determine if old
12670 code will still work when these eventually go away.
12674 * New extension functions for OCSP structures, these follow the
12675 same conventions as certificates and CRLs.
12679 * New function X509V3_add1_i2d(). This automatically encodes and
12680 adds an extension. Its behaviour can be customised with various
12681 flags to append, replace or delete. Various wrappers added for
12682 certificates and CRLs.
12686 * Fix to avoid calling the underlying ASN1 print routine when
12687 an extension cannot be parsed. Correct a typo in the
12688 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12692 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12693 entries for variables.
12697 * Add functionality to apps/openssl.c for detecting locking
12698 problems: As the program is single-threaded, all we have
12699 to do is register a locking callback using an array for
12700 storing which locks are currently held by the program.
12704 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12705 SSL_get_ex_data_X509_STORE_idx(), which is used in
12706 ssl_verify_cert_chain() and thus can be called at any time
12707 during TLS/SSL handshakes so that thread-safety is essential.
12708 Unfortunately, the ex_data design is not at all suited
12709 for multi-threaded use, so it probably should be abolished.
12713 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12715 *Broadcom, tweaked and integrated by Geoff Thorpe*
12717 * Move common extension printing code to new function
12718 X509V3_print_extensions(). Reorganise OCSP print routines and
12719 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12723 * New function X509_signature_print() to remove duplication in some
12728 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12729 set (this was treated exactly the same as SET OF previously). This
12730 is used to reorder the STACK representing the structure to match the
12731 encoding. This will be used to get round a problem where a PKCS7
12732 structure which was signed could not be verified because the STACK
12733 order did not reflect the encoded order.
12737 * Reimplement the OCSP ASN1 module using the new code.
12741 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12742 for its ASN1 operations. The old style function pointers still exist
12743 for now but they will eventually go away.
12747 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12748 completely replaces the old ASN1 functionality with a table driven
12749 encoder and decoder which interprets an ASN1_ITEM structure describing
12750 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12751 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12752 has also been converted to the new form.
12756 * Change BN_mod_exp_recp so that negative moduli are tolerated
12757 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12758 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12759 for negative moduli.
12763 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12764 of not touching the result's sign bit.
12768 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12773 * Changed the LHASH code to use prototypes for callbacks, and created
12774 macros to declare and implement thin (optionally static) functions
12775 that provide type-safety and avoid function pointer casting for the
12776 type-specific callbacks.
12780 * Added Kerberos Cipher Suites to be used with TLS, as written in
12782 *Veers Staats <staatsvr@asc.hpc.mil>,
12783 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12785 * Reformat the FAQ so the different questions and answers can be divided
12786 in sections depending on the subject.
12790 * Have the zlib compression code load ZLIB.DLL dynamically under
12795 * New function BN_mod_sqrt for computing square roots modulo a prime
12796 (using the probabilistic Tonelli-Shanks algorithm unless
12797 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12798 be handled deterministically).
12800 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12802 * Make BN_mod_inverse faster by explicitly handling small quotients
12803 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12804 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12808 * New function BN_kronecker.
12812 * Fix BN_gcd so that it works on negative inputs; the result is
12813 positive unless both parameters are zero.
12814 Previously something reasonably close to an infinite loop was
12815 possible because numbers could be growing instead of shrinking
12816 in the implementation of Euclid's algorithm.
12820 * Fix BN_is_word() and BN_is_one() macros to take into account the
12821 sign of the number in question.
12823 Fix BN_is_word(a,w) to work correctly for w == 0.
12825 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12826 because its test if the absolute value of 'a' equals 'w'.
12827 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12828 it exists mostly for use in the implementations of BN_is_zero(),
12829 BN_is_one(), and BN_is_word().
12833 * New function BN_swap.
12837 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12838 the exponentiation functions are more likely to produce reasonable
12839 results on negative inputs.
12843 * Change BN_mod_mul so that the result is always non-negative.
12844 Previously, it could be negative if one of the factors was negative;
12845 I don't think anyone really wanted that behaviour.
12849 * Move `BN_mod_...` functions into new file crypto/bn/bn_mod.c
12850 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12851 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12852 and add new functions:
12861 BN_mod_lshift1_quick
12863 BN_mod_lshift_quick
12865 These functions always generate non-negative results.
12867 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12868 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12870 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12871 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12872 be reduced modulo m.
12874 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12877 The following entry accidentally appeared in the CHANGES file
12878 distributed with OpenSSL 0.9.7. The modifications described in
12879 it do *not* apply to OpenSSL 0.9.7.
12881 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12882 was actually never needed) and in BN_mul(). The removal in BN_mul()
12883 required a small change in bn_mul_part_recursive() and the addition
12884 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12885 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12886 bn_sub_words() and bn_add_words() except they take arrays with
12892 * In 'openssl passwd', verify passwords read from the terminal
12893 unless the '-salt' option is used (which usually means that
12894 verification would just waste user's time since the resulting
12895 hash is going to be compared with some given password hash)
12896 or the new '-noverify' option is used.
12898 This is an incompatible change, but it does not affect
12899 non-interactive use of 'openssl passwd' (passwords on the command
12900 line, '-stdin' option, '-in ...' option) and thus should not
12901 cause any problems.
12905 * Remove all references to RSAref, since there's no more need for it.
12909 * Make DSO load along a path given through an environment variable
12910 (SHLIB_PATH) with shl_load().
12914 * Constify the ENGINE code as a result of BIGNUM constification.
12915 Also constify the RSA code and most things related to it. In a
12916 few places, most notable in the depth of the ASN.1 code, ugly
12917 casts back to non-const were required (to be solved at a later
12922 * Make it so the openssl application has all engines loaded by default.
12926 * Constify the BIGNUM routines a little more.
12930 * Add the following functions:
12932 ENGINE_load_cswift()
12934 ENGINE_load_atalla()
12935 ENGINE_load_nuron()
12936 ENGINE_load_builtin_engines()
12938 That way, an application can itself choose if external engines that
12939 are built-in in OpenSSL shall ever be used or not. The benefit is
12940 that applications won't have to be linked with libdl or other dso
12941 libraries unless it's really needed.
12943 Changed 'openssl engine' to load all engines on demand.
12944 Changed the engine header files to avoid the duplication of some
12945 declarations (they differed!).
12949 * 'openssl engine' can now list capabilities.
12953 * Better error reporting in 'openssl engine'.
12957 * Never call load_dh_param(NULL) in s_server.
12961 * Add engine application. It can currently list engines by name and
12962 identity, and test if they are actually available.
12966 * Improve RPM specification file by forcing symbolic linking and making
12967 sure the installed documentation is also owned by root.root.
12969 *Damien Miller <djm@mindrot.org>*
12971 * Give the OpenSSL applications more possibilities to make use of
12972 keys (public as well as private) handled by engines.
12976 * Add OCSP code that comes from CertCo.
12980 * Add VMS support for the Rijndael code.
12984 * Added untested support for Nuron crypto accelerator.
12988 * Add support for external cryptographic devices. This code was
12989 previously distributed separately as the "engine" branch.
12991 *Geoff Thorpe, Richard Levitte*
12993 * Rework the filename-translation in the DSO code. It is now possible to
12994 have far greater control over how a "name" is turned into a filename
12995 depending on the operating environment and any oddities about the
12996 different shared library filenames on each system.
13000 * Support threads on FreeBSD-elf in Configure.
13004 * Fix for SHA1 assembly problem with MASM: it produces
13005 warnings about corrupt line number information when assembling
13006 with debugging information. This is caused by the overlapping
13009 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13012 NCONF_get_number() has no error checking at all. As a replacement,
13013 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13014 promoted strongly. The old NCONF_get_number is kept around for
13015 binary backward compatibility.
13016 Make it possible for methods to load from something other than a BIO,
13017 by providing a function pointer that is given a name instead of a BIO.
13018 For example, this could be used to load configuration data from an
13023 * Fix for non blocking accept BIOs. Added new I/O special reason
13024 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13025 with non blocking I/O was not possible because no retry code was
13026 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13031 * Added the beginnings of Rijndael support.
13035 * Fix for bug in DirectoryString mask setting. Add support for
13036 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13037 to allow certificate printing to more controllable, additional
13038 'certopt' option to 'x509' to allow new printing options to be
13043 * Clean old EAY MD5 hack from e_os.h.
13047 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13049 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13050 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13052 *Joe Orton, Steve Henson*
13054 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13056 * Fix additional bug revealed by the NISCC test suite:
13058 Stop bug triggering large recursion when presented with
13059 certain ASN.1 tags [CVE-2003-0851][]
13063 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13065 * Fix various bugs revealed by running the NISCC test suite:
13067 Stop out of bounds reads in the ASN1 code when presented with
13068 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13070 If verify callback ignores invalid public key errors don't try to check
13071 certificate signature with the NULL public key.
13075 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13076 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13081 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13082 extra data after the compression methods not only for TLS 1.0
13083 but also for SSL 3.0 (as required by the specification).
13085 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13087 * Change X509_certificate_type() to mark the key as exported/exportable
13088 when it's 512 *bits* long, not 512 bytes.
13092 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13094 * Countermeasure against the Klima-Pokorny-Rosa extension of
13095 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13096 a protocol version number mismatch like a decryption error
13097 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13101 * Turn on RSA blinding by default in the default implementation
13102 to avoid a timing attack. Applications that don't want it can call
13103 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13104 They would be ill-advised to do so in most cases.
13106 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13108 * Change RSA blinding code so that it works when the PRNG is not
13109 seeded (in this case, the secret RSA exponent is abused as
13110 an unpredictable seed -- if it is not unpredictable, there
13111 is no point in blinding anyway). Make RSA blinding thread-safe
13112 by remembering the creator's thread ID in rsa->blinding and
13113 having all other threads use local one-time blinding factors
13114 (this requires more computation than sharing rsa->blinding, but
13115 avoids excessive locking; and if an RSA object is not shared
13116 between threads, blinding will still be very fast).
13120 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13122 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13123 via timing by performing a MAC computation even if incorrect
13124 block cipher padding has been found. This is a countermeasure
13125 against active attacks where the attacker has to distinguish
13126 between bad padding and a MAC verification error. [CVE-2003-0078][]
13128 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13129 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13130 Martin Vuagnoux (EPFL, Ilion)*
13132 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13134 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13135 memory from its contents. This is done with a counter that will
13136 place alternating values in each byte. This can be used to solve
13137 two issues: 1) the removal of calls to memset() by highly optimizing
13138 compilers, and 2) cleansing with other values than 0, since those can
13139 be read through on certain media, for example a swap space on disk.
13143 * Bugfix: client side session caching did not work with external caching,
13144 because the session->cipher setting was not restored when reloading
13145 from the external cache. This problem was masked, when
13146 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13147 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13151 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13152 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13154 *Zeev Lieber <zeev-l@yahoo.com>*
13156 * Undo an undocumented change introduced in 0.9.6e which caused
13157 repeated calls to OpenSSL_add_all_ciphers() and
13158 OpenSSL_add_all_digests() to be ignored, even after calling
13163 * Change the default configuration reader to deal with last line not
13164 being properly terminated.
13168 * Change X509_NAME_cmp() so it applies the special rules on handling
13169 DN values that are of type PrintableString, as well as RDNs of type
13170 emailAddress where the value has the type ia5String.
13172 *stefank@valicert.com via Richard Levitte*
13174 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13175 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13176 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13177 the bitwise-OR of the two for use by the majority of applications
13178 wanting this behaviour, and update the docs. The documented
13179 behaviour and actual behaviour were inconsistent and had been
13180 changing anyway, so this is more a bug-fix than a behavioural
13183 *Geoff Thorpe, diagnosed by Nadav Har'El*
13185 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13186 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13190 * Fix initialization code race conditions in
13191 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13192 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13193 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13194 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13195 ssl2_get_cipher_by_char(),
13196 ssl3_get_cipher_by_char().
13198 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13200 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13201 the cached sessions are flushed, as the remove_cb() might use ex_data
13202 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13203 (see [openssl.org #212]).
13205 *Geoff Thorpe, Lutz Jaenicke*
13207 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13208 length, instead of the encoding length to d2i_ASN1_OBJECT.
13212 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13214 * [In 0.9.6g-engine release:]
13215 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13217 *Lynn Gazis <lgazis@rainbow.com>*
13219 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13221 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13222 and get fix the header length calculation.
13223 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13224 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13226 * Use proper error handling instead of 'assertions' in buffer
13227 overflow checks added in 0.9.6e. This prevents DoS (the
13228 assertions could call abort()).
13230 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13232 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13234 * Add various sanity checks to asn1_get_length() to reject
13235 the ASN1 length bytes if they exceed sizeof(long), will appear
13236 negative or the content length exceeds the length of the
13239 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13241 * Fix cipher selection routines: ciphers without encryption had no flags
13242 for the cipher strength set and where therefore not handled correctly
13243 by the selection routines (PR #130).
13247 * Fix EVP_dsa_sha macro.
13252 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13253 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13254 that was added in OpenSSL 0.9.6d.
13256 As the countermeasure turned out to be incompatible with some
13257 broken SSL implementations, the new option is part of SSL_OP_ALL.
13258 SSL_OP_ALL is usually employed when compatibility with weird SSL
13259 implementations is desired (e.g. '-bugs' option to 's_client' and
13260 's_server'), so the new option is automatically set in many
13265 * Changes in security patch:
13267 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13268 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13269 Air Force Materiel Command, USAF, under agreement number
13272 * Add various sanity checks to asn1_get_length() to reject
13273 the ASN1 length bytes if they exceed sizeof(long), will appear
13274 negative or the content length exceeds the length of the
13275 supplied buffer. [CVE-2002-0659][]
13277 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13279 * Assertions for various potential buffer overflows, not known to
13280 happen in practice.
13282 *Ben Laurie (CHATS)*
13284 * Various temporary buffers to hold ASCII versions of integers were
13285 too small for 64 bit platforms. [CVE-2002-0655][]
13286 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13288 * Remote buffer overflow in SSL3 protocol - an attacker could
13289 supply an oversized session ID to a client. [CVE-2002-0656][]
13291 *Ben Laurie (CHATS)*
13293 * Remote buffer overflow in SSL2 protocol - an attacker could
13294 supply an oversized client master key. [CVE-2002-0656][]
13296 *Ben Laurie (CHATS)*
13298 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13300 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13301 encoded as NULL) with id-dsa-with-sha1.
13303 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13305 * Check various `X509_...()` return values in apps/req.c.
13307 *Nils Larsch <nla@trustcenter.de>*
13309 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13310 an end-of-file condition would erroneously be flagged, when the CRLF
13311 was just at the end of a processed block. The bug was discovered when
13312 processing data through a buffering memory BIO handing the data to a
13313 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13314 <ptsekov@syntrex.com> and Nedelcho Stanev.
13318 * Implement a countermeasure against a vulnerability recently found
13319 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13320 before application data chunks to avoid the use of known IVs
13321 with data potentially chosen by the attacker.
13325 * Fix length checks in ssl3_get_client_hello().
13329 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13330 to prevent ssl3_read_internal() from incorrectly assuming that
13331 ssl3_read_bytes() found application data while handshake
13332 processing was enabled when in fact s->s3->in_read_app_data was
13333 merely automatically cleared during the initial handshake.
13335 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13337 * Fix object definitions for Private and Enterprise: they were not
13338 recognized in their shortname (=lowercase) representation. Extend
13339 obj_dat.pl to issue an error when using undefined keywords instead
13340 of silently ignoring the problem (Svenning Sorensen
13341 <sss@sss.dnsalias.net>).
13345 * Fix DH_generate_parameters() so that it works for 'non-standard'
13346 generators, i.e. generators other than 2 and 5. (Previously, the
13347 code did not properly initialise the 'add' and 'rem' values to
13348 BN_generate_prime().)
13350 In the new general case, we do not insist that 'generator' is
13351 actually a primitive root: This requirement is rather pointless;
13352 a generator of the order-q subgroup is just as good, if not
13357 * Map new X509 verification errors to alerts. Discovered and submitted by
13358 Tom Wu <tom@arcot.com>.
13362 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13363 returning non-zero before the data has been completely received
13364 when using non-blocking I/O.
13366 *Bodo Moeller; problem pointed out by John Hughes*
13368 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13370 *Ben Laurie, Lutz Jaenicke*
13372 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13373 Yoram Zahavi <YoramZ@gilian.com>).
13377 * Add information about CygWin 1.3 and on, and preserve proper
13378 configuration for the versions before that.
13380 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13382 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13383 check whether we deal with a copy of a session and do not delete from
13384 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13385 <izhar@checkpoint.com>.
13389 * Do not store session data into the internal session cache, if it
13390 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13391 flag is set). Proposed by Aslam <aslam@funk.com>.
13395 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13400 * [In 0.9.6d-engine release:]
13401 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13403 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13405 * Add the configuration target linux-s390x.
13407 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13409 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13410 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13411 variable as an indication that a ClientHello message has been
13412 received. As the flag value will be lost between multiple
13413 invocations of ssl3_accept when using non-blocking I/O, the
13414 function may not be aware that a handshake has actually taken
13415 place, thus preventing a new session from being added to the
13418 To avoid this problem, we now set s->new_session to 2 instead of
13419 using a local variable.
13421 *Lutz Jaenicke, Bodo Moeller*
13423 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13424 if the SSL_R_LENGTH_MISMATCH error is detected.
13426 *Geoff Thorpe, Bodo Moeller*
13428 * New 'shared_ldflag' column in Configure platform table.
13432 * Fix EVP_CIPHER_mode macro.
13434 *"Dan S. Camper" <dan@bti.net>*
13436 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13437 type, we must throw them away by setting rr->length to 0.
13439 *D P Chang <dpc@qualys.com>*
13441 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13443 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13444 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13445 worked incorrectly for those cases where range = `10..._2` and
13446 `3*range` is two bits longer than range.)
13450 * Only add signing time to PKCS7 structures if it is not already
13455 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13456 OBJ_ld_ce should be OBJ_id_ce.
13457 Also some ip-pda OIDs in crypto/objects/objects.txt were
13458 incorrect (cf. RFC 3039).
13460 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13462 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13463 returns early because it has nothing to do.
13465 *Andy Schneider <andy.schneider@bjss.co.uk>*
13467 * [In 0.9.6c-engine release:]
13468 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13470 *Andy Schneider <andy.schneider@bjss.co.uk>*
13472 * [In 0.9.6c-engine release:]
13473 Add support for Cryptographic Appliance's keyserver technology.
13474 (Use engine 'keyclient')
13476 *Cryptographic Appliances and Geoff Thorpe*
13478 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13479 is called via tools/c89.sh because arguments have to be
13480 rearranged (all '-L' options must appear before the first object
13483 *Richard Shapiro <rshapiro@abinitio.com>*
13485 * [In 0.9.6c-engine release:]
13486 Add support for Broadcom crypto accelerator cards, backported
13489 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13491 * [In 0.9.6c-engine release:]
13492 Add support for SureWare crypto accelerator cards from
13493 Baltimore Technologies. (Use engine 'sureware')
13495 *Baltimore Technologies and Mark Cox*
13497 * [In 0.9.6c-engine release:]
13498 Add support for crypto accelerator cards from Accelerated
13499 Encryption Processing, www.aep.ie. (Use engine 'aep')
13501 *AEP Inc. and Mark Cox*
13503 * Add a configuration entry for gcc on UnixWare.
13505 *Gary Benson <gbenson@redhat.com>*
13507 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13508 messages are stored in a single piece (fixed-length part and
13509 variable-length part combined) and fix various bugs found on the way.
13513 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13514 instead. BIO_gethostbyname() does not know what timeouts are
13515 appropriate, so entries would stay in cache even when they have
13517 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13519 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13520 faced with a pathologically small ClientHello fragment that does
13521 not contain client_version: Instead of aborting with an error,
13522 simply choose the highest available protocol version (i.e.,
13523 TLS 1.0 unless it is disabled). In practice, ClientHello
13524 messages are never sent like this, but this change gives us
13525 strictly correct behaviour at least for TLS.
13529 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13530 never resets s->method to s->ctx->method when called from within
13531 one of the SSL handshake functions.
13533 *Bodo Moeller; problem pointed out by Niko Baric*
13535 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13536 (sent using the client's version number) if client_version is
13537 smaller than the protocol version in use. Also change
13538 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13539 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13540 the client will at least see that alert.
13544 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13549 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13550 client receives HelloRequest while in a handshake.
13552 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13554 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13555 should end in 'break', not 'goto end' which circumvents various
13556 cleanups done in state SSL_ST_OK. But session related stuff
13557 must be disabled for SSL_ST_OK in the case that we just sent a
13560 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13561 before just sending a HelloRequest.
13563 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13565 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13566 reveal whether illegal block cipher padding was found or a MAC
13567 verification error occurred. (Neither SSLerr() codes nor alerts
13568 are directly visible to potential attackers, but the information
13569 may leak via logfiles.)
13571 Similar changes are not required for the SSL 2.0 implementation
13572 because the number of padding bytes is sent in clear for SSL 2.0,
13573 and the extra bytes are just ignored. However ssl/s2_pkt.c
13574 failed to verify that the purported number of padding bytes is in
13579 * Add OpenUNIX-8 support including shared libraries
13580 (Boyd Lynn Gerber <gerberb@zenez.com>).
13584 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13585 'wristwatch attack' using huge encoding parameters (cf.
13586 James H. Manger's CRYPTO 2001 paper). Note that the
13587 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13588 encoding parameters and hence was not vulnerable.
13592 * BN_sqr() bug fix.
13594 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13596 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13597 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13598 followed by modular reduction.
13600 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13602 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13603 equivalent based on BN_pseudo_rand() instead of BN_rand().
13607 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13608 This function was broken, as the check for a new client hello message
13609 to handle SGC did not allow these large messages.
13610 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13614 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13618 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13619 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13623 * Rework the configuration and shared library support for Tru64 Unix.
13624 The configuration part makes use of modern compiler features and
13625 still retains old compiler behavior for those that run older versions
13626 of the OS. The shared library support part includes a variant that
13627 uses the RPATH feature, and is available through the special
13628 configuration target "alpha-cc-rpath", which will never be selected
13631 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13633 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13634 with the same message size as in ssl3_get_certificate_request().
13635 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13636 messages might inadvertently be reject as too long.
13638 *Petr Lampa <lampa@fee.vutbr.cz>*
13640 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13644 * Modified SSL library such that the verify_callback that has been set
13645 specifically for an SSL object with SSL_set_verify() is actually being
13646 used. Before the change, a verify_callback set with this function was
13647 ignored and the verify_callback() set in the SSL_CTX at the time of
13648 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13649 to allow the necessary settings.
13653 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13654 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13655 done automatically (in contradiction to the requirements of the C
13656 standard). This made problems when used from OpenSSH.
13660 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13661 dh->length and always used
13663 BN_rand_range(priv_key, dh->p).
13665 BN_rand_range() is not necessary for Diffie-Hellman, and this
13666 specific range makes Diffie-Hellman unnecessarily inefficient if
13667 dh->length (recommended exponent length) is much smaller than the
13668 length of dh->p. We could use BN_rand_range() if the order of
13669 the subgroup was stored in the DH structure, but we only have
13674 BN_rand(priv_key, l, ...)
13676 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13683 RSA_eay_public_encrypt
13684 RSA_eay_private_decrypt
13685 RSA_eay_private_encrypt (signing)
13686 RSA_eay_public_decrypt (signature verification)
13688 (default implementations for RSA_public_encrypt,
13689 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13690 always reject numbers >= n.
13694 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13695 to synchronize access to 'locking_thread'. This is necessary on
13696 systems where access to 'locking_thread' (an 'unsigned long'
13697 variable) is not atomic.
13701 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13702 *before* setting the 'crypto_lock_rand' flag. The previous code had
13703 a race condition if 0 is a valid thread ID.
13705 *Travis Vitek <vitek@roguewave.com>*
13707 * Add support for shared libraries under Irix.
13709 *Albert Chin-A-Young <china@thewrittenword.com>*
13711 * Add configuration option to build on Linux on both big-endian and
13712 little-endian MIPS.
13714 *Ralf Baechle <ralf@uni-koblenz.de>*
13716 * Add the possibility to create shared libraries on HP-UX.
13720 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13722 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13723 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13724 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13725 PRNG state recovery was possible based on the output of
13726 one PRNG request appropriately sized to gain knowledge on
13727 'md' followed by enough consecutive 1-byte PRNG requests
13728 to traverse all of 'state'.
13730 1. When updating 'md_local' (the current thread's copy of 'md')
13731 during PRNG output generation, hash all of the previous
13732 'md_local' value, not just the half used for PRNG output.
13734 2. Make the number of bytes from 'state' included into the hash
13735 independent from the number of PRNG bytes requested.
13737 The first measure alone would be sufficient to avoid
13738 Markku-Juhani's attack. (Actually it had never occurred
13739 to me that the half of 'md_local' used for chaining was the
13740 half from which PRNG output bytes were taken -- I had always
13741 assumed that the secret half would be used.) The second
13742 measure makes sure that additional data from 'state' is never
13743 mixed into 'md_local' in small portions; this heuristically
13744 further strengthens the PRNG.
13748 * Fix crypto/bn/asm/mips3.s.
13752 * When only the key is given to "enc", the IV is undefined. Print out
13753 an error message in this case.
13757 * Handle special case when X509_NAME is empty in X509 printing routines.
13761 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13762 positive and less than q.
13766 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13767 used: it isn't thread safe and the add_lock_callback should handle
13770 *Paul Rose <Paul.Rose@bridge.com>*
13772 * Verify that incoming data obeys the block size in
13773 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13779 *Ulf Möller, Bodo Möller*
13781 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13782 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13783 when fixing the server behaviour for backwards-compatible 'client
13784 hello' messages. (Note that the attack is impractical against
13785 SSL 3.0 and TLS 1.0 anyway because length and version checking
13786 means that the probability of guessing a valid ciphertext is
13787 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13790 Before 0.9.5, the countermeasure (hide the error by generating a
13791 random 'decryption result') did not work properly because
13792 ERR_clear_error() was missing, meaning that SSL_get_error() would
13793 detect the supposedly ignored error.
13795 Both problems are now fixed.
13799 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13800 (previously it was 1024).
13804 * Fix for compatibility mode trust settings: ignore trust settings
13805 unless some valid trust or reject settings are present.
13809 * Fix for blowfish EVP: its a variable length cipher.
13813 * Fix various bugs related to DSA S/MIME verification. Handle missing
13814 parameters in DSA public key structures and return an error in the
13815 DSA routines if parameters are absent.
13819 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13820 in the current directory if neither $RANDFILE nor $HOME was set.
13821 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13822 caused some confusion to Windows users who haven't defined $HOME.
13823 Thus RAND_file_name() is changed again: e_os.h can define a
13824 DEFAULT_HOME, which will be used if $HOME is not set.
13825 For Windows, we use "C:"; on other platforms, we still require
13826 environment variables.
13828 * Move 'if (!initialized) RAND_poll()' into regions protected by
13829 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13830 having multiple threads call RAND_poll() concurrently.
13834 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13835 combination of a flag and a thread ID variable.
13836 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13837 flag), *other* threads can enter ssleay_add_bytes without obeying
13838 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13839 that they do not hold after the first thread unsets add_do_not_lock).
13843 * Change bctest again: '-x' expressions are not available in all
13844 versions of 'test'.
13848 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13850 * Fix a couple of memory leaks in PKCS7_dataDecode()
13852 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13854 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13855 the default extension for executables, if any. Also, make the perl
13856 scripts that use symlink() to test if it really exists and use "cp"
13857 if it doesn't. All this made OpenSSL compilable and installable in
13862 * Fix for asn1_GetSequence() for indefinite length constructed data.
13863 If SEQUENCE is length is indefinite just set c->slen to the total
13864 amount of data available.
13866 *Steve Henson, reported by shige@FreeBSD.org*
13868 *This change does not apply to 0.9.7.*
13870 * Change bctest to avoid here-documents inside command substitution
13871 (workaround for FreeBSD /bin/sh bug).
13872 For compatibility with Ultrix, avoid shell functions (introduced
13873 in the bctest version that searches along $PATH).
13877 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13878 with des_encrypt() defined on some operating systems, like Solaris
13883 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13884 On the Importance of Eliminating Errors in Cryptographic
13885 Computations, J. Cryptology 14 (2001) 2, 101-119,
13886 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
13890 * MIPS assembler BIGNUM division bug fix.
13894 * Disabled incorrect Alpha assembler code.
13898 * Fix PKCS#7 decode routines so they correctly update the length
13899 after reading an EOC for the EXPLICIT tag.
13903 *This change does not apply to 0.9.7.*
13905 * Fix bug in PKCS#12 key generation routines. This was triggered
13906 if a 3DES key was generated with a 0 initial byte. Include
13907 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13908 (but broken) behaviour.
13912 * Enhance bctest to search for a working bc along $PATH and print
13915 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13917 * Fix memory leaks in err.c: free err_data string if necessary;
13918 don't write to the wrong index in ERR_set_error_data.
13922 * Implement ssl23_peek (analogous to ssl23_read), which previously
13927 * Replace rdtsc with `_emit` statements for VC++ version 5.
13929 *Jeremy Cooper <jeremy@baymoo.org>*
13931 * Make it possible to reuse SSLv2 sessions.
13935 * In copy_email() check for >= 0 as a return value for
13936 X509_NAME_get_index_by_NID() since 0 is a valid index.
13938 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13940 * Avoid coredump with unsupported or invalid public keys by checking if
13941 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13942 PKCS7_verify() fails with non detached data.
13946 * Don't use getenv in library functions when run as setuid/setgid.
13947 New function OPENSSL_issetugid().
13951 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13952 due to incorrect handling of multi-threading:
13954 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13956 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13958 3. Count how many times MemCheck_off() has been called so that
13959 nested use can be treated correctly. This also avoids
13960 inband-signalling in the previous code (which relied on the
13961 assumption that thread ID 0 is impossible).
13965 * Add "-rand" option also to s_client and s_server.
13969 * Fix CPU detection on Irix 6.x.
13970 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13971 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13973 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13978 *This change does not apply to 0.9.7.*
13980 * Use the cached encoding of an X509_NAME structure rather than
13981 copying it. This is apparently the reason for the libsafe "errors"
13982 but the code is actually correct.
13986 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13987 Bleichenbacher's DSA attack.
13988 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13989 to be set and top=0 forces the highest bit to be set; top=-1 is new
13990 and leaves the highest bit random.
13992 *Ulf Moeller, Bodo Moeller*
13994 * In the `NCONF_...`-based implementations for `CONF_...` queries
13995 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13996 a temporary CONF structure with the data component set to NULL
13997 (which gives segmentation faults in lh_retrieve).
13998 Instead, use NULL for the CONF pointer in CONF_get_string and
13999 CONF_get_number (which may use environment variables) and directly
14000 return NULL from CONF_get_section.
14004 * Fix potential buffer overrun for EBCDIC.
14008 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14009 keyUsage if basicConstraints absent for a CA.
14013 * Make SMIME_write_PKCS7() write mail header values with a format that
14014 is more generally accepted (no spaces before the semicolon), since
14015 some programs can't parse those values properly otherwise. Also make
14016 sure BIO's that break lines after each write do not create invalid
14021 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14022 macros previously used would not encode an empty SEQUENCE OF
14023 and break the signature.
14027 *This change does not apply to 0.9.7.*
14029 * Zero the premaster secret after deriving the master secret in
14034 * Add some EVP_add_digest_alias registrations (as found in
14035 OpenSSL_add_all_digests()) to SSL_library_init()
14036 aka OpenSSL_add_ssl_algorithms(). This provides improved
14037 compatibility with peers using X.509 certificates
14038 with unconventional AlgorithmIdentifier OIDs.
14042 * Fix for Irix with NO_ASM.
14044 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14046 * ./config script fixes.
14048 *Ulf Moeller, Richard Levitte*
14050 * Fix 'openssl passwd -1'.
14054 * Change PKCS12_key_gen_asc() so it can cope with non null
14055 terminated strings whose length is passed in the passlen
14056 parameter, for example from PEM callbacks. This was done
14057 by adding an extra length parameter to asc2uni().
14059 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14061 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14062 call failed, free the DSA structure.
14066 * Fix to uni2asc() to cope with zero length Unicode strings.
14067 These are present in some PKCS#12 files.
14071 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14072 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14073 when writing a 32767 byte record.
14075 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14077 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14078 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14080 (RSA objects have a reference count access to which is protected
14081 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14082 so they are meant to be shared between threads.)
14083 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14084 "Reddie, Steven" <Steven.Reddie@ca.com>*
14086 * Fix a deadlock in CRYPTO_mem_leaks().
14090 * Use better test patterns in bntest.
14094 * rand_win.c fix for Borland C.
14098 * BN_rshift bugfix for n == 0.
14102 * Add a 'bctest' script that checks for some known 'bc' bugs
14103 so that 'make test' does not abort just because 'bc' is broken.
14107 * Store verify_result within SSL_SESSION also for client side to
14108 avoid potential security hole. (Re-used sessions on the client side
14109 always resulted in verify_result==X509_V_OK, not using the original
14110 result of the server certificate verification.)
14114 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14115 SSL3_RT_APPLICATION_DATA, return 0.
14116 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14121 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14122 releases, have been re-implemented by renaming the previous
14123 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14124 and ssl3_read_internal, respectively, and adding 'peek' parameters
14125 to them. The new ssl[23]_{read,peek} functions are calls to
14126 ssl[23]_read_internal with the 'peek' flag set appropriately.
14127 A 'peek' parameter has also been added to ssl3_read_bytes, which
14128 does the actual work for ssl3_read_internal.
14132 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14133 the method-specific "init()" handler. Also clean up ex_data after
14134 calling the method-specific "finish()" handler. Previously, this was
14135 happening the other way round.
14139 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14140 The previous value, 12, was not always sufficient for BN_mod_exp().
14144 * Make sure that shared libraries get the internal name engine with
14145 the full version number and not just 0. This should mark the
14146 shared libraries as not backward compatible. Of course, this should
14147 be changed again when we can guarantee backward binary compatibility.
14151 * Fix typo in get_cert_by_subject() in by_dir.c
14153 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14155 * Rework the system to generate shared libraries:
14157 - Make note of the expected extension for the shared libraries and
14158 if there is a need for symbolic links from for example libcrypto.so.0
14159 to libcrypto.so.0.9.7. There is extended info in Configure for
14162 - Make as few rebuilds of the shared libraries as possible.
14164 - Still avoid linking the OpenSSL programs with the shared libraries.
14166 - When installing, install the shared libraries separately from the
14171 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14173 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14174 and not in SSL_clear because the latter is also used by the
14175 accept/connect functions; previously, the settings made by
14176 SSL_set_read_ahead would be lost during the handshake.
14178 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14180 * Correct util/mkdef.pl to be selective about disabled algorithms.
14181 Previously, it would create entries for disabled algorithms no
14186 * Added several new manual pages for SSL_* function.
14190 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14192 * In ssl23_get_client_hello, generate an error message when faced
14193 with an initial SSL 3.0/TLS record that is too small to contain the
14194 first two bytes of the ClientHello message, i.e. client_version.
14195 (Note that this is a pathologic case that probably has never happened
14196 in real life.) The previous approach was to use the version number
14197 from the record header as a substitute; but our protocol choice
14198 should not depend on that one because it is not authenticated
14199 by the Finished messages.
14203 * More robust randomness gathering functions for Windows.
14205 *Jeffrey Altman <jaltman@columbia.edu>*
14207 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14208 not set then we don't setup the error code for issuer check errors
14209 to avoid possibly overwriting other errors which the callback does
14210 handle. If an application does set the flag then we assume it knows
14211 what it is doing and can handle the new informational codes
14216 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14217 a general "ANY" type, as such it should be able to decode anything
14218 including tagged types. However it didn't check the class so it would
14219 wrongly interpret tagged types in the same way as their universal
14220 counterpart and unknown types were just rejected. Changed so that the
14221 tagged and unknown types are handled in the same way as a SEQUENCE:
14222 that is the encoding is stored intact. There is also a new type
14223 "V_ASN1_OTHER" which is used when the class is not universal, in this
14224 case we have no idea what the actual type is so we just lump them all
14229 * On VMS, stdout may very well lead to a file that is written to
14230 in a record-oriented fashion. That means that every write() will
14231 write a separate record, which will be read separately by the
14232 programs trying to read from it. This can be very confusing.
14234 The solution is to put a BIO filter in the way that will buffer
14235 text until a linefeed is reached, and then write everything a
14236 line at a time, so every record written will be an actual line,
14237 not chunks of lines and not (usually doesn't happen, but I've
14238 seen it once) several lines in one record. BIO_f_linebuffer() is
14241 Currently, it's a VMS-only method, because that's where it has
14242 been tested well enough.
14246 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14247 it can return incorrect results.
14248 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14249 but it was in 0.9.6-beta[12].)
14253 * Disable the check for content being present when verifying detached
14254 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14255 include zero length content when signing messages.
14259 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14260 BIO_ctrl (for BIO pairs).
14264 * Add DSO method for VMS.
14268 * Bug fix: Montgomery multiplication could produce results with the
14273 * Add RPM specification openssl.spec and modify it to build three
14274 packages. The default package contains applications, application
14275 documentation and run-time libraries. The devel package contains
14276 include files, static libraries and function documentation. The
14277 doc package contains the contents of the doc directory. The original
14278 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14282 * Add a large number of documentation files for many SSL routines.
14284 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14286 * Add a configuration entry for Sony News 4.
14288 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14290 * Don't set the two most significant bits to one when generating a
14291 random number < q in the DSA library.
14295 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14296 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14297 the underlying transport is blocking) if a handshake took place.
14298 (The default behaviour is needed by applications such as s_client
14299 and s_server that use select() to determine when to use SSL_read;
14300 but for applications that know in advance when to expect data, it
14301 just makes things more complicated.)
14305 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14310 * Add a few more EBCDIC conditionals that make `req` and `x509`
14311 work better on such systems.
14313 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14315 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14316 Update PKCS12_parse() so it copies the friendlyName and the
14317 keyid to the certificates aux info.
14321 * Fix bug in PKCS7_verify() which caused an infinite loop
14322 if there was more than one signature.
14324 *Sven Uszpelkat <su@celocom.de>*
14326 * Major change in util/mkdef.pl to include extra information
14327 about each symbol, as well as presenting variables as well
14328 as functions. This change means that there's n more need
14329 to rebuild the .num files when some algorithms are excluded.
14333 * Allow the verify time to be set by an application,
14334 rather than always using the current time.
14338 * Phase 2 verify code reorganisation. The certificate
14339 verify code now looks up an issuer certificate by a
14340 number of criteria: subject name, authority key id
14341 and key usage. It also verifies self signed certificates
14342 by the same criteria. The main comparison function is
14343 X509_check_issued() which performs these checks.
14345 Lot of changes were necessary in order to support this
14346 without completely rewriting the lookup code.
14348 Authority and subject key identifier are now cached.
14350 The LHASH 'certs' is X509_STORE has now been replaced
14351 by a STACK_OF(X509_OBJECT). This is mainly because an
14352 LHASH can't store or retrieve multiple objects with
14353 the same hash value.
14355 As a result various functions (which were all internal
14356 use only) have changed to handle the new X509_STORE
14357 structure. This will break anything that messed round
14358 with X509_STORE internally.
14360 The functions X509_STORE_add_cert() now checks for an
14361 exact match, rather than just subject name.
14363 The X509_STORE API doesn't directly support the retrieval
14364 of multiple certificates matching a given criteria, however
14365 this can be worked round by performing a lookup first
14366 (which will fill the cache with candidate certificates)
14367 and then examining the cache for matches. This is probably
14368 the best we can do without throwing out X509_LOOKUP
14369 entirely (maybe later...).
14371 The X509_VERIFY_CTX structure has been enhanced considerably.
14373 All certificate lookup operations now go via a get_issuer()
14374 callback. Although this currently uses an X509_STORE it
14375 can be replaced by custom lookups. This is a simple way
14376 to bypass the X509_STORE hackery necessary to make this
14377 work and makes it possible to use more efficient techniques
14378 in future. A very simple version which uses a simple
14379 STACK for its trusted certificate store is also provided
14380 using X509_STORE_CTX_trusted_stack().
14382 The verify_cb() and verify() callbacks now have equivalents
14383 in the X509_STORE_CTX structure.
14385 X509_STORE_CTX also has a 'flags' field which can be used
14386 to customise the verify behaviour.
14390 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14391 excludes S/MIME capabilities.
14395 * When a certificate request is read in keep a copy of the
14396 original encoding of the signed data and use it when outputting
14397 again. Signatures then use the original encoding rather than
14398 a decoded, encoded version which may cause problems if the
14399 request is improperly encoded.
14403 * For consistency with other BIO_puts implementations, call
14404 buffer_write(b, ...) directly in buffer_puts instead of calling
14407 In BIO_puts, increment b->num_write as in BIO_write.
14409 *Peter.Sylvester@EdelWeb.fr*
14411 * Fix BN_mul_word for the case where the word is 0. (We have to use
14412 BN_zero, we may not return a BIGNUM with an array consisting of
14413 words set to zero.)
14417 * Avoid calling abort() from within the library when problems are
14418 detected, except if preprocessor symbols have been defined
14419 (such as REF_CHECK, BN_DEBUG etc.).
14423 * New openssl application 'rsautl'. This utility can be
14424 used for low level RSA operations. DER public key
14425 BIO/fp routines also added.
14429 * New Configure entry and patches for compiling on QNX 4.
14431 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14433 * A demo state-machine implementation was sponsored by
14434 Nuron (<http://www.nuron.com/>) and is now available in
14435 demos/state_machine.
14439 * New options added to the 'dgst' utility for signature
14440 generation and verification.
14444 * Unrecognized PKCS#7 content types are now handled via a
14445 catch all ASN1_TYPE structure. This allows unsupported
14446 types to be stored as a "blob" and an application can
14447 encode and decode it manually.
14451 * Fix various signed/unsigned issues to make a_strex.c
14452 compile under VC++.
14454 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14456 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14457 length if passed a buffer. ASN1_INTEGER_to_BN failed
14458 if passed a NULL BN and its argument was negative.
14460 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14462 * Modification to PKCS#7 encoding routines to output definite
14463 length encoding. Since currently the whole structures are in
14464 memory there's not real point in using indefinite length
14465 constructed encoding. However if OpenSSL is compiled with
14466 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14470 * Added BIO_vprintf() and BIO_vsnprintf().
14474 * Added more prefixes to parse for in the strings written
14475 through a logging bio, to cover all the levels that are available
14476 through syslog. The prefixes are now:
14478 PANIC, EMERG, EMR => LOG_EMERG
14479 ALERT, ALR => LOG_ALERT
14480 CRIT, CRI => LOG_CRIT
14481 ERROR, ERR => LOG_ERR
14482 WARNING, WARN, WAR => LOG_WARNING
14483 NOTICE, NOTE, NOT => LOG_NOTICE
14484 INFO, INF => LOG_INFO
14485 DEBUG, DBG => LOG_DEBUG
14487 and as before, if none of those prefixes are present at the
14488 beginning of the string, LOG_ERR is chosen.
14490 On Win32, the `LOG_*` levels are mapped according to this:
14492 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14493 LOG_WARNING => EVENTLOG_WARNING_TYPE
14494 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14498 * Made it possible to reconfigure with just the configuration
14499 argument "reconf" or "reconfigure". The command line arguments
14500 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14501 and are retrieved from there when reconfiguring.
14507 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14509 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14513 * The obj_dat.pl script was messing up the sorting of object
14514 names. The reason was that it compared the quoted version
14515 of strings as a result "OCSP" > "OCSP Signing" because
14516 " > SPACE. Changed script to store unquoted versions of
14517 names and add quotes on output. It was also omitting some
14518 names from the lookup table if they were given a default
14519 value (that is if SN is missing it is given the same
14520 value as LN and vice versa), these are now added on the
14521 grounds that if an object has a name we should be able to
14522 look it up. Finally added warning output when duplicate
14523 short or long names are found.
14527 * Changes needed for Tandem NSK.
14529 *Scott Uroff <scott@xypro.com>*
14531 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14532 RSA_padding_check_SSLv23(), special padding was never detected
14533 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14534 version rollback attacks was not effective.
14536 In s23_clnt.c, don't use special rollback-attack detection padding
14537 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14538 client; similarly, in s23_srvr.c, don't do the rollback check if
14539 SSL 2.0 is the only protocol enabled in the server.
14543 * Make it possible to get hexdumps of unprintable data with 'openssl
14544 asn1parse'. By implication, the functions ASN1_parse_dump() and
14545 BIO_dump_indent() are added.
14549 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14550 these print out strings and name structures based on various
14551 flags including RFC2253 support and proper handling of
14552 multibyte characters. Added options to the 'x509' utility
14553 to allow the various flags to be set.
14557 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14558 Also change the functions X509_cmp_current_time() and
14559 X509_gmtime_adj() work with an ASN1_TIME structure,
14560 this will enable certificates using GeneralizedTime in validity
14561 dates to be checked.
14565 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14566 negative public key encodings) on by default,
14567 NO_NEG_PUBKEY_BUG can be set to disable it.
14571 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14572 content octets. An i2c_ASN1_OBJECT is unnecessary because
14573 the encoding can be trivially obtained from the structure.
14577 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14578 not read locks (`CRYPTO_r_[un]lock`).
14582 * A first attempt at creating official support for shared
14583 libraries through configuration. I've kept it so the
14584 default is static libraries only, and the OpenSSL programs
14585 are always statically linked for now, but there are
14586 preparations for dynamic linking in place.
14587 This has been tested on Linux and Tru64.
14591 * Randomness polling function for Win9x, as described in:
14592 Peter Gutmann, Software Generation of Practically Strong
14597 * Fix so PRNG is seeded in req if using an already existing
14602 * New options to smime application. -inform and -outform
14603 allow alternative formats for the S/MIME message including
14604 PEM and DER. The -content option allows the content to be
14605 specified separately. This should allow things like Netscape
14606 form signing output easier to verify.
14610 * Fix the ASN1 encoding of tags using the 'long form'.
14614 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14615 STRING types. These convert content octets to and from the
14616 underlying type. The actual tag and length octets are
14617 already assumed to have been read in and checked. These
14618 are needed because all other string types have virtually
14619 identical handling apart from the tag. By having versions
14620 of the ASN1 functions that just operate on content octets
14621 IMPLICIT tagging can be handled properly. It also allows
14622 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14623 and ASN1_INTEGER are identical apart from the tag.
14627 * Change the handling of OID objects as follows:
14629 - New object identifiers are inserted in objects.txt, following
14630 the syntax given in objects.README.
14631 - objects.pl is used to process obj_mac.num and create a new
14633 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14636 This is currently kind of a hack, and the perl code in objects.pl
14637 isn't very elegant, but it works as I intended. The simplest way
14638 to check that it worked correctly is to look in obj_dat.h and
14639 check the array nid_objs and make sure the objects haven't moved
14640 around (this is important!). Additions are OK, as well as
14641 consistent name changes.
14645 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14649 * Addition of the command line parameter '-rand file' to 'openssl req'.
14650 The given file adds to whatever has already been seeded into the
14651 random pool through the RANDFILE configuration file option or
14652 environment variable, or the default random state file.
14656 * mkstack.pl now sorts each macro group into lexical order.
14657 Previously the output order depended on the order the files
14658 appeared in the directory, resulting in needless rewriting
14663 * Patches to make OpenSSL compile under Win32 again. Mostly
14664 work arounds for the VC++ problem that it treats func() as
14665 func(void). Also stripped out the parts of mkdef.pl that
14666 added extra typesafe functions: these no longer exist.
14670 * Reorganisation of the stack code. The macros are now all
14671 collected in safestack.h . Each macro is defined in terms of
14672 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14673 DEBUG_SAFESTACK is now handled in terms of function casts,
14674 this has the advantage of retaining type safety without the
14675 use of additional functions. If DEBUG_SAFESTACK is not defined
14676 then the non typesafe macros are used instead. Also modified the
14677 mkstack.pl script to handle the new form. Needs testing to see
14678 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14679 the default if no major problems. Similar behaviour for ASN1_SET_OF
14680 and PKCS12_STACK_OF.
14684 * When some versions of IIS use the 'NET' form of private key the
14685 key derivation algorithm is different. Normally MD5(password) is
14686 used as a 128 bit RC4 key. In the modified case
14687 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14688 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14689 as the old Netscape_RSA functions except they have an additional
14690 'sgckey' parameter which uses the modified algorithm. Also added
14691 an -sgckey command line option to the rsa utility. Thanks to
14692 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14693 algorithm to openssl-dev.
14697 * The evp_local.h macros were using 'c.##kname' which resulted in
14698 invalid expansion on some systems (SCO 5.0.5 for example).
14699 Corrected to 'c.kname'.
14701 *Phillip Porch <root@theporch.com>*
14703 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14704 a STACK of email addresses from a certificate or request, these look
14705 in the subject name and the subject alternative name extensions and
14706 omit any duplicate addresses.
14710 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14711 This makes DSA verification about 2 % faster.
14715 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14716 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14717 plus overhead for 1024 bit moduli).
14718 This makes exponentiations about 0.5 % faster for 1024 bit
14719 exponents (as measured by "openssl speed rsa2048").
14723 * Rename memory handling macros to avoid conflicts with other
14725 Malloc => OPENSSL_malloc
14726 Malloc_locked => OPENSSL_malloc_locked
14727 Realloc => OPENSSL_realloc
14728 Free => OPENSSL_free
14732 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14733 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14737 * CygWin32 support.
14739 *John Jarvie <jjarvie@newsguy.com>*
14741 * The type-safe stack code has been rejigged. It is now only compiled
14742 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14743 by default all type-specific stack functions are "#define"d back to
14744 standard stack functions. This results in more streamlined output
14745 but retains the type-safety checking possibilities of the original
14750 * The STACK code has been cleaned up, and certain type declarations
14751 that didn't make a lot of sense have been brought in line. This has
14752 also involved a cleanup of sorts in safestack.h to more correctly
14753 map type-safe stack functions onto their plain stack counterparts.
14754 This work has also resulted in a variety of "const"ifications of
14755 lots of the code, especially `_cmp` operations which should normally
14756 be prototyped with "const" parameters anyway.
14760 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14761 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14762 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14763 where all of 'md' is used each time the PRNG is used, but 'state'
14764 is used only indexed by a cyclic counter. As entropy may not be
14765 well distributed from the beginning, 'md' is important as a
14766 chaining variable. However, the output function chains only half
14767 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14768 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14769 in all of 'state' being rewritten, with the new values depending
14770 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14774 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14775 the handshake is continued after ssl_verify_cert_chain();
14776 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14777 can lead to 'unexplainable' connection aborts later.
14779 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14781 * Major EVP API cipher revision.
14782 Add hooks for extra EVP features. This allows various cipher
14783 parameters to be set in the EVP interface. Support added for variable
14784 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14785 setting of RC2 and RC5 parameters.
14787 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14790 Remove lots of duplicated code from the EVP library. For example *every*
14791 cipher init() function handles the 'iv' in the same way according to the
14792 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14793 for CFB and OFB modes they zero ctx->num.
14795 New functionality allows removal of S/MIME code RC2 hack.
14797 Most of the routines have the same form and so can be declared in terms
14800 By shifting this to the top level EVP_CipherInit() it can be removed from
14801 all individual ciphers. If the cipher wants to handle IVs or keys
14802 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14805 Change lots of functions like EVP_EncryptUpdate() to now return a
14806 value: although software versions of the algorithms cannot fail
14807 any installed hardware versions can.
14811 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14812 this option is set, tolerate broken clients that send the negotiated
14813 protocol version number instead of the requested protocol version
14818 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14819 i.e. non-zero for export ciphersuites, zero otherwise.
14820 Previous versions had this flag inverted, inconsistent with
14821 rsa_tmp_cb (..._TMP_RSA_CB).
14823 *Bodo Moeller; problem reported by Amit Chopra*
14825 * Add missing DSA library text string. Work around for some IIS
14826 key files with invalid SEQUENCE encoding.
14830 * Add a document (doc/standards.txt) that list all kinds of standards
14831 and so on that are implemented in OpenSSL.
14835 * Enhance c_rehash script. Old version would mishandle certificates
14836 with the same subject name hash and wouldn't handle CRLs at all.
14837 Added -fingerprint option to crl utility, to support new c_rehash
14842 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14846 * Fix for SSL server purpose checking. Server checking was
14847 rejecting certificates which had extended key usage present
14848 but no ssl client purpose.
14850 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14852 * Make PKCS#12 code work with no password. The PKCS#12 spec
14853 is a little unclear about how a blank password is handled.
14854 Since the password in encoded as a BMPString with terminating
14855 double NULL a zero length password would end up as just the
14856 double NULL. However no password at all is different and is
14857 handled differently in the PKCS#12 key generation code. NS
14858 treats a blank password as zero length. MSIE treats it as no
14859 password on export: but it will try both on import. We now do
14860 the same: PKCS12_parse() tries zero length and no password if
14861 the password is set to "" or NULL (NULL is now a valid password:
14862 it wasn't before) as does the pkcs12 application.
14866 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14867 perror when PEM_read_bio_X509_REQ fails, the error message must
14868 be obtained from the error queue.
14872 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14873 it in ERR_remove_state if appropriate, and change ERR_get_state
14874 accordingly to avoid race conditions (this is necessary because
14875 thread_hash is no longer constant once set).
14879 * Bugfix for linux-elf makefile.one.
14883 * RSA_get_default_method() will now cause a default
14884 RSA_METHOD to be chosen if one doesn't exist already.
14885 Previously this was only set during a call to RSA_new()
14886 or RSA_new_method(NULL) meaning it was possible for
14887 RSA_get_default_method() to return NULL.
14891 * Added native name translation to the existing DSO code
14892 that will convert (if the flag to do so is set) filenames
14893 that are sufficiently small and have no path information
14894 into a canonical native form. Eg. "blah" converted to
14895 "libblah.so" or "blah.dll" etc.
14899 * New function ERR_error_string_n(e, buf, len) which is like
14900 ERR_error_string(e, buf), but writes at most 'len' bytes
14901 including the 0 terminator. For ERR_error_string_n, 'buf'
14904 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14906 * CONF library reworked to become more general. A new CONF
14907 configuration file reader "class" is implemented as well as a
14908 new functions (`NCONF_*`, for "New CONF") to handle it. The now
14909 old `CONF_*` functions are still there, but are reimplemented to
14910 work in terms of the new functions. Also, a set of functions
14911 to handle the internal storage of the configuration data is
14912 provided to make it easier to write new configuration file
14913 reader "classes" (I can definitely see something reading a
14914 configuration file in XML format, for example), called `_CONF_*`,
14915 or "the configuration storage API"...
14917 The new configuration file reading functions are:
14919 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14920 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14922 NCONF_default, NCONF_WIN32
14924 NCONF_dump_fp, NCONF_dump_bio
14926 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14927 NCONF_new creates a new CONF object. This works in the same way
14928 as other interfaces in OpenSSL, like the BIO interface.
14929 `NCONF_dump_*` dump the internal storage of the configuration file,
14930 which is useful for debugging. All other functions take the same
14931 arguments as the old `CONF_*` functions with the exception of the
14932 first that must be a `CONF *` instead of a `LHASH *`.
14934 To make it easier to use the new classes with the old `CONF_*` functions,
14935 the function CONF_set_default_method is provided.
14939 * Add '-tls1' option to 'openssl ciphers', which was already
14940 mentioned in the documentation but had not been implemented.
14941 (This option is not yet really useful because even the additional
14942 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14946 * Initial DSO code added into libcrypto for letting OpenSSL (and
14947 OpenSSL-based applications) load shared libraries and bind to
14948 them in a portable way.
14950 *Geoff Thorpe, with contributions from Richard Levitte*
14952 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
14954 * Make sure _lrotl and _lrotr are only used with MSVC.
14956 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14957 (the default implementation of RAND_status).
14959 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14960 to '-clrext' (= clear extensions), as intended and documented.
14961 *Bodo Moeller; inconsistency pointed out by Michael Attili
14962 <attili@amaxo.com>*
14964 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14965 was larger than the MD block size.
14967 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14969 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14970 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14971 using the passed key: if the passed key was a private key the result
14972 of X509_print(), for example, would be to print out all the private key
14977 * des_quad_cksum() byte order bug fix.
14978 *Ulf Möller, using the problem description in krb4-0.9.7, where
14979 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14981 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14984 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14986 * For easily testing in shell scripts whether some command
14987 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14988 returns with exit code 0 iff no command of the given name is available.
14989 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14990 the output goes to stdout and nothing is printed to stderr.
14991 Additional arguments are always ignored.
14993 Since for each cipher there is a command of the same name,
14994 the 'no-cipher' compilation switches can be tested this way.
14996 ('openssl no-XXX' is not able to detect pseudo-commands such
14997 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15001 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15005 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15006 is set; it will be thrown away anyway because each handshake creates
15008 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15009 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15010 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15011 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15015 * New s_client option -ign_eof: EOF at stdin is ignored, and
15016 'Q' and 'R' lose their special meanings (quit/renegotiate).
15017 This is part of what -quiet does; unlike -quiet, -ign_eof
15018 does not suppress any output.
15022 * Add compatibility options to the purpose and trust code. The
15023 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15024 accepts a certificate or CA, this was the previous behaviour,
15025 with all the associated security issues.
15027 X509_TRUST_COMPAT is the old trust behaviour: only and
15028 automatically trust self signed roots in certificate store. A
15029 new trust setting X509_TRUST_DEFAULT is used to specify that
15030 a purpose has no associated trust setting and it should instead
15031 use the value in the default purpose.
15035 * Fix the PKCS#8 DSA private key code so it decodes keys again
15036 and fix a memory leak.
15040 * In util/mkerr.pl (which implements 'make errors'), preserve
15041 reason strings from the previous version of the .c file, as
15042 the default to have only downcase letters (and digits) in
15043 automatically generated reasons codes is not always appropriate.
15047 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15048 using strerror. Previously, ERR_reason_error_string() returned
15049 library names as reason strings for SYSerr; but SYSerr is a special
15050 case where small numbers are errno values, not library numbers.
15054 * Add '-dsaparam' option to 'openssl dhparam' application. This
15055 converts DSA parameters into DH parameters. (When creating parameters,
15056 DSA_generate_parameters is used.)
15060 * Include 'length' (recommended exponent length) in C code generated
15061 by 'openssl dhparam -C'.
15065 * The second argument to set_label in perlasm was already being used
15066 so couldn't be used as a "file scope" flag. Moved to third argument
15071 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15072 instead of RAND_bytes for encryption IVs and salts.
15076 * Include RAND_status() into RAND_METHOD instead of implementing
15077 it only for md_rand.c Otherwise replacing the PRNG by calling
15078 RAND_set_rand_method would be impossible.
15082 * Don't let DSA_generate_key() enter an infinite loop if the random
15083 number generation fails.
15087 * New 'rand' application for creating pseudo-random output.
15091 * Added configuration support for Linux/IA64
15093 *Rolf Haberrecker <rolf@suse.de>*
15095 * Assembler module support for Mingw32.
15099 * Shared library support for HPUX (in shlib/).
15101 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15103 * Shared library support for Solaris gcc.
15105 *Lutz Behnke <behnke@trustcenter.de>*
15107 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15109 * PKCS7_encrypt() was adding text MIME headers twice because they
15110 were added manually and by SMIME_crlf_copy().
15114 * In bntest.c don't call BN_rand with zero bits argument.
15116 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15118 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15119 case was implemented. This caused BN_div_recp() to fail occasionally.
15123 * Add an optional second argument to the set_label() in the perl
15124 assembly language builder. If this argument exists and is set
15125 to 1 it signals that the assembler should use a symbol whose
15126 scope is the entire file, not just the current function. This
15127 is needed with MASM which uses the format label:: for this scope.
15129 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15131 * Change the ASN1 types so they are typedefs by default. Before
15132 almost all types were #define'd to ASN1_STRING which was causing
15133 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15138 * Change names of new functions to the new get1/get0 naming
15139 convention: After 'get1', the caller owns a reference count
15140 and has to call `..._free`; 'get0' returns a pointer to some
15141 data structure without incrementing reference counters.
15142 (Some of the existing 'get' functions increment a reference
15143 counter, some don't.)
15144 Similarly, 'set1' and 'add1' functions increase reference
15145 counters or duplicate objects.
15149 * Allow for the possibility of temp RSA key generation failure:
15150 the code used to assume it always worked and crashed on failure.
15154 * Fix potential buffer overrun problem in BIO_printf().
15155 *Ulf Möller, using public domain code by Patrick Powell; problem
15156 pointed out by David Sacerdote <das33@cornell.edu>*
15158 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15159 RAND_egd() and RAND_status(). In the command line application,
15160 the EGD socket can be specified like a seed file using RANDFILE
15165 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15166 Some CAs (e.g. Verisign) distribute certificates in this form.
15170 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15171 list to exclude them. This means that no special compilation option
15172 is needed to use anonymous DH: it just needs to be included in the
15177 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15178 EVP_MD_type. The old functionality is available in a new macro called
15179 EVP_MD_md(). Change code that uses it and update docs.
15183 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15184 where the `void *` argument is replaced by a function pointer argument.
15185 Previously `void *` was abused to point to functions, which works on
15186 many platforms, but is not correct. As these functions are usually
15187 called by macros defined in OpenSSL header files, most source code
15188 should work without changes.
15192 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15193 sections with information on -D... compiler switches used for
15194 compiling the library so that applications can see them. To enable
15195 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15196 must be defined. E.g.,
15197 #define OPENSSL_ALGORITHM_DEFINES
15198 #include <openssl/opensslconf.h>
15199 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15201 *Richard Levitte, Ulf and Bodo Möller*
15203 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15208 * Change the 'other' type in certificate aux info to a STACK_OF
15209 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15210 the required ASN1 format: arbitrary types determined by an OID.
15214 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15215 argument to 'req'. This is not because the function is newer or
15216 better than others it just uses the work 'NEW' in the certificate
15217 request header lines. Some software needs this.
15221 * Reorganise password command line arguments: now passwords can be
15222 obtained from various sources. Delete the PEM_cb function and make
15223 it the default behaviour: i.e. if the callback is NULL and the
15224 usrdata argument is not NULL interpret it as a null terminated pass
15225 phrase. If usrdata and the callback are NULL then the pass phrase
15226 is prompted for as usual.
15230 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15231 the support is automatically enabled. The resulting binaries will
15232 autodetect the card and use it if present.
15234 *Ben Laurie and Compaq Inc.*
15236 * Work around for Netscape hang bug. This sends certificate request
15237 and server done in one record. Since this is perfectly legal in the
15238 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15239 the bugs/SSLv3 entry for more info.
15243 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15247 * Add -rand argument to smime and pkcs12 applications and read/write
15252 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15256 * Add command line password options to the remaining applications.
15260 * Bug fix for BN_div_recp() for numerators with an even number of
15265 * More tests in bntest.c, and changed test_bn output.
15269 * ./config recognizes MacOS X now.
15273 * Bug fix for BN_div() when the first words of num and divisor are
15274 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15278 * Add support for various broken PKCS#8 formats, and command line
15279 options to produce them.
15283 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15284 get temporary BIGNUMs from a BN_CTX.
15288 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15293 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15294 include a #define from the old name to the new. The original intent
15295 was that statically linked binaries could for example just call
15296 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15297 link with digests. This never worked because SSLeay_add_all_digests()
15298 and SSLeay_add_all_ciphers() were in the same source file so calling
15299 one would link with the other. They are now in separate source files.
15303 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15307 * Use a less unusual form of the Miller-Rabin primality test (it used
15308 a binary algorithm for exponentiation integrated into the Miller-Rabin
15309 loop, our standard modexp algorithms are faster).
15313 * Support for the EBCDIC character set completed.
15315 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15317 * Source code cleanups: use const where appropriate, eliminate casts,
15318 use `void *` instead of `char *` in lhash.
15322 * Bugfix: ssl3_send_server_key_exchange was not restartable
15323 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15324 this the server could overwrite ephemeral keys that the client
15329 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15330 using 50 iterations of the Rabin-Miller test.
15332 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15333 iterations of the Rabin-Miller test as required by the appendix
15334 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15335 As BN_is_prime_fasttest includes trial division, DSA parameter
15336 generation becomes much faster.
15338 This implies a change for the callback functions in DSA_is_prime
15339 and DSA_generate_parameters: The callback function is called once
15340 for each positive witness in the Rabin-Miller test, not just
15341 occasionally in the inner loop; and the parameters to the
15342 callback function now provide an iteration count for the outer
15343 loop rather than for the current invocation of the inner loop.
15344 DSA_generate_parameters additionally can call the callback
15345 function with an 'iteration count' of -1, meaning that a
15346 candidate has passed the trial division test (when q is generated
15347 from an application-provided seed, trial division is skipped).
15351 * New function BN_is_prime_fasttest that optionally does trial
15352 division before starting the Rabin-Miller test and has
15353 an additional BN_CTX * argument (whereas BN_is_prime always
15354 has to allocate at least one BN_CTX).
15355 'callback(1, -1, cb_arg)' is called when a number has passed the
15356 trial division stage.
15360 * Fix for bug in CRL encoding. The validity dates weren't being handled
15365 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15369 * New function BN_pseudo_rand().
15373 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15374 bignum version of BN_from_montgomery() with the working code from
15375 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15380 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15381 made it impossible to use the same SSL_SESSION data structure in
15382 SSL2 clients in multiple threads.
15386 * The return value of RAND_load_file() no longer counts bytes obtained
15387 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15388 to seed the PRNG (previously an explicit byte count was required).
15390 *Ulf Möller, Bodo Möller*
15392 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15393 used `char *` instead of `void *` and had casts all over the place.
15397 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15401 * Retain source code compatibility for BN_prime_checks macro:
15402 BN_is_prime(..., BN_prime_checks, ...) now uses
15403 BN_prime_checks_for_size to determine the appropriate number of
15404 Rabin-Miller iterations.
15408 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15409 DH_CHECK_P_NOT_SAFE_PRIME.
15410 (Check if this is true? OpenPGP calls them "strong".)
15414 * Merge the functionality of "dh" and "gendh" programs into a new program
15415 "dhparam". The old programs are retained for now but will handle DH keys
15416 (instead of parameters) in future.
15420 * Make the ciphers, s_server and s_client programs check the return values
15421 when a new cipher list is set.
15425 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15426 ciphers. Before when the 56bit ciphers were enabled the sorting was
15429 The syntax for the cipher sorting has been extended to support sorting by
15430 cipher-strength (using the strength_bits hard coded in the tables).
15431 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15433 Fix a bug in the cipher-command parser: when supplying a cipher command
15434 string with an "undefined" symbol (neither command nor alphanumeric
15435 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15436 an error is flagged.
15438 Due to the strength-sorting extension, the code of the
15439 ssl_create_cipher_list() function was completely rearranged. I hope that
15440 the readability was also increased :-)
15442 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15444 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15445 for the first serial number and places 2 in the serial number file. This
15446 avoids problems when the root CA is created with serial number zero and
15447 the first user certificate has the same issuer name and serial number
15452 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15453 the new code. Add documentation for this stuff.
15457 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15458 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15459 structures and behave in an analogous way to the X509v3 functions:
15460 they shouldn't be called directly but wrapper functions should be used
15463 So we also now have some wrapper functions that call the X509at functions
15464 when passed certificate requests. (TO DO: similar things can be done with
15465 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15466 things. Some of these need some d2i or i2d and print functionality
15467 because they handle more complex structures.)
15471 * Add missing #ifndefs that caused missing symbols when building libssl
15472 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15473 NO_RSA in `ssl/s2*.c`.
15475 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15477 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15478 has a return value which indicates the quality of the random data
15479 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15480 error queue. New function RAND_pseudo_bytes() generates output that is
15481 guaranteed to be unique but not unpredictable. RAND_add is like
15482 RAND_seed, but takes an extra argument for an entropy estimate
15483 (RAND_seed always assumes full entropy).
15487 * Do more iterations of Rabin-Miller probable prime test (specifically,
15488 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15489 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15490 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15491 false-positive rate of at most 2^-80 for random input.
15495 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15499 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15500 in the 0.9.5 release), this returns the chain
15501 from an X509_CTX structure with a dup of the stack and all
15502 the X509 reference counts upped: so the stack will exist
15503 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15506 Also make SSL_SESSION_print() print out the verify return
15511 * Add manpage for the pkcs12 command. Also change the default
15512 behaviour so MAC iteration counts are used unless the new
15513 -nomaciter option is used. This improves file security and
15514 only older versions of MSIE (4.0 for example) need it.
15518 * Honor the no-xxx Configure options when creating .DEF files.
15522 * Add PKCS#10 attributes to field table: challengePassword,
15523 unstructuredName and unstructuredAddress. These are taken from
15524 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15525 international characters are used.
15527 More changes to X509_ATTRIBUTE code: allow the setting of types
15528 based on strings. Remove the 'loc' parameter when adding
15529 attributes because these will be a SET OF encoding which is sorted
15534 * Initial changes to the 'req' utility to allow request generation
15535 automation. This will allow an application to just generate a template
15536 file containing all the field values and have req construct the
15539 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15540 used all over the place including certificate requests and PKCS#7
15541 structures. They are currently handled manually where necessary with
15542 some primitive wrappers for PKCS#7. The new functions behave in a
15543 manner analogous to the X509 extension functions: they allow
15544 attributes to be looked up by NID and added.
15546 Later something similar to the X509V3 code would be desirable to
15547 automatically handle the encoding, decoding and printing of the
15548 more complex types. The string types like challengePassword can
15549 be handled by the string table functions.
15551 Also modified the multi byte string table handling. Now there is
15552 a 'global mask' which masks out certain types. The table itself
15553 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15554 is useful when for example there is only one permissible type
15555 (as in countryName) and using the mask might result in no valid
15560 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15561 SSL_get_peer_finished to allow applications to obtain the latest
15562 Finished messages sent to the peer or expected from the peer,
15563 respectively. (SSL_get_peer_finished is usually the Finished message
15564 actually received from the peer, otherwise the protocol will be aborted.)
15566 As the Finished message are message digests of the complete handshake
15567 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15568 be used for external authentication procedures when the authentication
15569 provided by SSL/TLS is not desired or is not enough.
15573 * Enhanced support for Alpha Linux is added. Now ./config checks if
15574 the host supports BWX extension and if Compaq C is present on the
15575 $PATH. Just exploiting of the BWX extension results in 20-30%
15576 performance kick for some algorithms, e.g. DES and RC4 to mention
15577 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15582 * Add support for MS "fast SGC". This is arguably a violation of the
15583 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15584 weak crypto and after checking the certificate is SGC a second one
15585 with strong crypto. MS SGC stops the first handshake after receiving
15586 the server certificate message and sends a second client hello. Since
15587 a server will typically do all the time consuming operations before
15588 expecting any further messages from the client (server key exchange
15589 is the most expensive) there is little difference between the two.
15591 To get OpenSSL to support MS SGC we have to permit a second client
15592 hello message after we have sent server done. In addition we have to
15593 reset the MAC if we do get this second client hello.
15597 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15598 if a DER encoded private key is RSA or DSA traditional format. Changed
15599 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15600 format DER encoded private key. Newer code should use PKCS#8 format which
15601 has the key type encoded in the ASN1 structure. Added DER private key
15602 support to pkcs8 application.
15606 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15607 ciphersuites has been selected (as required by the SSL 3/TLS 1
15608 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15609 is set, we interpret this as a request to violate the specification
15610 (the worst that can happen is a handshake failure, and 'correct'
15611 behaviour would result in a handshake failure anyway).
15615 * In SSL_CTX_add_session, take into account that there might be multiple
15616 SSL_SESSION structures with the same session ID (e.g. when two threads
15617 concurrently obtain them from an external cache).
15618 The internal cache can handle only one SSL_SESSION with a given ID,
15619 so if there's a conflict, we now throw out the old one to achieve
15624 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15625 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15626 some routines that use cipher OIDs: some ciphers do not have OIDs
15627 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15632 * Simplify the trust setting structure and code. Now we just have
15633 two sequences of OIDs for trusted and rejected settings. These will
15634 typically have values the same as the extended key usage extension
15635 and any application specific purposes.
15637 The trust checking code now has a default behaviour: it will just
15638 check for an object with the same NID as the passed id. Functions can
15639 be provided to override either the default behaviour or the behaviour
15640 for a given id. SSL client, server and email already have functions
15641 in place for compatibility: they check the NID and also return "trusted"
15642 if the certificate is self signed.
15646 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15647 traditional format into an EVP_PKEY structure.
15651 * Add a password callback function PEM_cb() which either prompts for
15652 a password if usr_data is NULL or otherwise assumes it is a null
15653 terminated password. Allow passwords to be passed on command line
15654 environment or config files in a few more utilities.
15658 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15659 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15660 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15661 Update documentation.
15665 * Support for ASN1 "NULL" type. This could be handled before by using
15666 ASN1_TYPE but there wasn't any function that would try to read a NULL
15667 and produce an error if it couldn't. For compatibility we also have
15668 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15669 don't allocate anything because they don't need to.
15673 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15676 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15678 * Rebuild of the memory allocation routines used by OpenSSL code and
15679 possibly others as well. The purpose is to make an interface that
15680 provide hooks so anyone can build a separate set of allocation and
15681 deallocation routines to be used by OpenSSL, for example memory
15682 pool implementations, or something else, which was previously hard
15683 since Malloc(), Realloc() and Free() were defined as macros having
15684 the values malloc, realloc and free, respectively (except for Win32
15685 compilations). The same is provided for memory debugging code.
15686 OpenSSL already comes with functionality to find memory leaks, but
15687 this gives people a chance to debug other memory problems.
15689 With these changes, a new set of functions and macros have appeared:
15691 CRYPTO_set_mem_debug_functions() [F]
15692 CRYPTO_get_mem_debug_functions() [F]
15693 CRYPTO_dbg_set_options() [F]
15694 CRYPTO_dbg_get_options() [F]
15695 CRYPTO_malloc_debug_init() [M]
15697 The memory debug functions are NULL by default, unless the library
15698 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15699 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15700 gives the standard debugging functions that come with OpenSSL) or
15701 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15702 provided by the library user) must be used. When the standard
15703 debugging functions are used, CRYPTO_dbg_set_options can be used to
15704 request additional information:
15705 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15706 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15708 Also, things like CRYPTO_set_mem_functions will always give the
15709 expected result (the new set of functions is used for allocation
15710 and deallocation) at all times, regardless of platform and compiler
15713 To finish it up, some functions that were never use in any other
15714 way than through macros have a new API and new semantic:
15716 CRYPTO_dbg_malloc()
15717 CRYPTO_dbg_realloc()
15720 All macros of value have retained their old syntax.
15722 *Richard Levitte and Bodo Moeller*
15724 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15725 ordering of SMIMECapabilities wasn't in "strength order" and there
15726 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15731 * Some ASN1 types with illegal zero length encoding (INTEGER,
15732 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15734 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15736 * Merge in my S/MIME library for OpenSSL. This provides a simple
15737 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15738 functionality to handle multipart/signed properly) and a utility
15739 called 'smime' to call all this stuff. This is based on code I
15740 originally wrote for Celo who have kindly allowed it to be
15741 included in OpenSSL.
15745 * Add variants des_set_key_checked and des_set_key_unchecked of
15746 des_set_key (aka des_key_sched). Global variable des_check_key
15747 decides which of these is called by des_set_key; this way
15748 des_check_key behaves as it always did, but applications and
15749 the library itself, which was buggy for des_check_key == 1,
15750 have a cleaner way to pick the version they need.
15754 * New function PKCS12_newpass() which changes the password of a
15759 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15760 dynamic mix. In both cases the ids can be used as an index into the
15761 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15762 functions so they accept a list of the field values and the
15763 application doesn't need to directly manipulate the X509_TRUST
15768 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15773 * Modify the way the V3 extension code looks up extensions. This now
15774 works in a similar way to the object code: we have some "standard"
15775 extensions in a static table which is searched with OBJ_bsearch()
15776 and the application can add dynamic ones if needed. The file
15777 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15778 updated whenever a new extension is added to the core code and kept
15779 in ext_nid order. There is a simple program 'tabtest.c' which checks
15780 this. New extensions are not added too often so this file can readily
15781 be maintained manually.
15783 There are two big advantages in doing things this way. The extensions
15784 can be looked up immediately and no longer need to be "added" using
15785 X509V3_add_standard_extensions(): this function now does nothing.
15786 Side note: I get *lots* of email saying the extension code doesn't
15787 work because people forget to call this function.
15788 Also no dynamic allocation is done unless new extensions are added:
15789 so if we don't add custom extensions there is no need to call
15790 X509V3_EXT_cleanup().
15794 * Modify enc utility's salting as follows: make salting the default. Add a
15795 magic header, so unsalted files fail gracefully instead of just decrypting
15796 to garbage. This is because not salting is a big security hole, so people
15797 should be discouraged from doing it.
15801 * Fixes and enhancements to the 'x509' utility. It allowed a message
15802 digest to be passed on the command line but it only used this
15803 parameter when signing a certificate. Modified so all relevant
15804 operations are affected by the digest parameter including the
15805 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15806 DSA key was used because it didn't fix the digest.
15810 * Initial certificate chain verify code. Currently tests the untrusted
15811 certificates for consistency with the verify purpose (which is set
15812 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15814 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15815 this is because it will reject chains with invalid extensions whereas
15816 every previous version of OpenSSL and SSLeay made no checks at all.
15818 Trust code: checks the root CA for the relevant trust settings. Trust
15819 settings have an initial value consistent with the verify purpose: e.g.
15820 if the verify purpose is for SSL client use it expects the CA to be
15821 trusted for SSL client use. However the default value can be changed to
15822 permit custom trust settings: one example of this would be to only trust
15823 certificates from a specific "secure" set of CAs.
15825 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15826 which should be used for version portability: especially since the
15827 verify structure is likely to change more often now.
15829 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15830 to set them. If not set then assume SSL clients will verify SSL servers
15833 Two new options to the verify program: -untrusted allows a set of
15834 untrusted certificates to be passed in and -purpose which sets the
15835 intended purpose of the certificate. If a purpose is set then the
15836 new chain verify code is used to check extension consistency.
15840 * Support for the authority information access extension.
15844 * Modify RSA and DSA PEM read routines to transparently handle
15845 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15846 public keys in a format compatible with certificate
15847 SubjectPublicKeyInfo structures. Unfortunately there were already
15848 functions called *_PublicKey_* which used various odd formats so
15849 these are retained for compatibility: however the DSA variants were
15850 never in a public release so they have been deleted. Changed dsa/rsa
15851 utilities to handle the new format: note no releases ever handled public
15852 keys so we should be OK.
15854 The primary motivation for this change is to avoid the same fiasco
15855 that dogs private keys: there are several incompatible private key
15856 formats some of which are standard and some OpenSSL specific and
15857 require various evil hacks to allow partial transparent handling and
15858 even then it doesn't work with DER formats. Given the option anything
15859 other than PKCS#8 should be dumped: but the other formats have to
15860 stay in the name of compatibility.
15862 With public keys and the benefit of hindsight one standard format
15863 is used which works with EVP_PKEY, RSA or DSA structures: though
15864 it clearly returns an error if you try to read the wrong kind of key.
15866 Added a -pubkey option to the 'x509' utility to output the public key.
15867 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15868 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15869 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15870 that do the same as the `EVP_PKEY_assign_*()` except they up the
15871 reference count of the added key (they don't "swallow" the
15876 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15877 CRLs would fail if the file contained no certificates or no CRLs:
15878 added a new function to read in both types and return the number
15879 read: this means that if none are read it will be an error. The
15880 DER versions of the certificate and CRL reader would always fail
15881 because it isn't possible to mix certificates and CRLs in DER format
15882 without choking one or the other routine. Changed this to just read
15883 a certificate: this is the best we can do. Also modified the code
15884 in apps/verify.c to take notice of return codes: it was previously
15885 attempting to read in certificates from NULL pointers and ignoring
15886 any errors: this is one reason why the cert and CRL reader seemed
15887 to work. It doesn't check return codes from the default certificate
15888 routines: these may well fail if the certificates aren't installed.
15892 * Code to support otherName option in GeneralName.
15896 * First update to verify code. Change the verify utility
15897 so it warns if it is passed a self signed certificate:
15898 for consistency with the normal behaviour. X509_verify
15899 has been modified to it will now verify a self signed
15900 certificate if *exactly* the same certificate appears
15901 in the store: it was previously impossible to trust a
15902 single self signed certificate. This means that:
15903 openssl verify ss.pem
15904 now gives a warning about a self signed certificate but
15905 openssl verify -CAfile ss.pem ss.pem
15910 * For servers, store verify_result in SSL_SESSION data structure
15911 (and add it to external session representation).
15912 This is needed when client certificate verifications fails,
15913 but an application-provided verification callback (set by
15914 SSL_CTX_set_cert_verify_callback) allows accepting the session
15915 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15916 but returns 1): When the session is reused, we have to set
15917 ssl->verify_result to the appropriate error code to avoid
15920 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15922 * Fix a bug in the new PKCS#7 code: it didn't consider the
15923 case in PKCS7_dataInit() where the signed PKCS7 structure
15924 didn't contain any existing data because it was being created.
15926 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15928 * Add a salt to the key derivation routines in enc.c. This
15929 forms the first 8 bytes of the encrypted file. Also add a
15930 -S option to allow a salt to be input on the command line.
15934 * New function X509_cmp(). Oddly enough there wasn't a function
15935 to compare two certificates. We do this by working out the SHA1
15936 hash and comparing that. X509_cmp() will be needed by the trust
15941 * SSL_get1_session() is like SSL_get_session(), but increments
15942 the reference count in the SSL_SESSION returned.
15944 *Geoff Thorpe <geoff@eu.c2.net>*
15946 * Fix for 'req': it was adding a null to request attributes.
15947 Also change the X509_LOOKUP and X509_INFO code to handle
15948 certificate auxiliary information.
15952 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15957 * Add the possibility to add extra information to the memory leak
15958 detecting output, to form tracebacks, showing from where each
15959 allocation was originated: CRYPTO_push_info("constant string") adds
15960 the string plus current file name and line number to a per-thread
15961 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15962 is like calling CYRPTO_pop_info() until the stack is empty.
15963 Also updated memory leak detection code to be multi-thread-safe.
15967 * Add options -text and -noout to pkcs7 utility and delete the
15968 encryption options which never did anything. Update docs.
15972 * Add options to some of the utilities to allow the pass phrase
15973 to be included on either the command line (not recommended on
15974 OSes like Unix) or read from the environment. Update the
15975 manpages and fix a few bugs.
15979 * Add a few manpages for some of the openssl commands.
15983 * Fix the -revoke option in ca. It was freeing up memory twice,
15984 leaking and not finding already revoked certificates.
15988 * Extensive changes to support certificate auxiliary information.
15989 This involves the use of X509_CERT_AUX structure and X509_AUX
15990 functions. An X509_AUX function such as PEM_read_X509_AUX()
15991 can still read in a certificate file in the usual way but it
15992 will also read in any additional "auxiliary information". By
15993 doing things this way a fair degree of compatibility can be
15994 retained: existing certificates can have this information added
15995 using the new 'x509' options.
15997 Current auxiliary information includes an "alias" and some trust
15998 settings. The trust settings will ultimately be used in enhanced
15999 certificate chain verification routines: currently a certificate
16000 can only be trusted if it is self signed and then it is trusted
16005 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16006 The problem was that one of the replacement routines had not been working
16007 since SSLeay releases. For now the offending routine has been replaced
16008 with non-optimised assembler. Even so, this now gives around 95%
16009 performance improvement for 1024 bit RSA signs.
16013 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16014 handling. Most clients have the effective key size in bits equal to
16015 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16016 A few however don't do this and instead use the size of the decrypted key
16017 to determine the RC2 key length and the AlgorithmIdentifier to determine
16018 the effective key length. In this case the effective key length can still
16019 be 40 bits but the key length can be 168 bits for example. This is fixed
16020 by manually forcing an RC2 key into the EVP_PKEY structure because the
16021 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16022 the key length and effective key length are equal.
16026 * Add a bunch of functions that should simplify the creation of
16027 X509_NAME structures. Now you should be able to do:
16028 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16029 and have it automatically work out the correct field type and fill in
16030 the structures. The more adventurous can try:
16031 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16032 and it will (hopefully) work out the correct multibyte encoding.
16036 * Change the 'req' utility to use the new field handling and multibyte
16037 copy routines. Before the DN field creation was handled in an ad hoc
16038 way in req, ca, and x509 which was rather broken and didn't support
16039 BMPStrings or UTF8Strings. Since some software doesn't implement
16040 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16041 using the dirstring_type option. See the new comment in the default
16042 openssl.cnf for more info.
16046 * Make crypto/rand/md_rand.c more robust:
16047 - Assure unique random numbers after fork().
16048 - Make sure that concurrent threads access the global counter and
16049 md serializably so that we never lose entropy in them
16050 or use exactly the same state in multiple threads.
16051 Access to the large state is not always serializable because
16052 the additional locking could be a performance killer, and
16053 md should be large enough anyway.
16057 * New file apps/app_rand.c with commonly needed functionality
16058 for handling the random seed file.
16060 Use the random seed file in some applications that previously did not:
16062 dsaparam -genkey (which also ignored its '-rand' option),
16065 x509 (when signing).
16066 Except on systems with /dev/urandom, it is crucial to have a random
16067 seed file at least for key creation, DSA signing, and for DH exchanges;
16068 for RSA signatures we could do without one.
16070 gendh and gendsa (unlike genrsa) used to read only the first byte
16071 of each file listed in the '-rand' option. The function as previously
16072 found in genrsa is now in app_rand.c and is used by all programs
16073 that support '-rand'.
16077 * In RAND_write_file, use mode 0600 for creating files;
16078 don't just chmod when it may be too late.
16082 * Report an error from X509_STORE_load_locations
16083 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16087 * New function ASN1_mbstring_copy() this copies a string in either
16088 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16089 into an ASN1_STRING type. A mask of permissible types is passed
16090 and it chooses the "minimal" type to use or an error if not type
16095 * Add function equivalents to the various macros in asn1.h. The old
16096 macros are retained with an `M_` prefix. Code inside the library can
16097 use the `M_` macros. External code (including the openssl utility)
16098 should *NOT* in order to be "shared library friendly".
16102 * Add various functions that can check a certificate's extensions
16103 to see if it usable for various purposes such as SSL client,
16104 server or S/MIME and CAs of these types. This is currently
16105 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16106 verification. Also added a -purpose flag to x509 utility to
16107 print out all the purposes.
16111 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16116 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16117 for, obtain and decode and extension and obtain its critical flag.
16118 This allows all the necessary extension code to be handled in a
16119 single function call.
16123 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16124 platforms. See crypto/rc4/rc4_enc.c for further details.
16128 * New -noout option to asn1parse. This causes no output to be produced
16129 its main use is when combined with -strparse and -out to extract data
16130 from a file (which may not be in ASN.1 format).
16134 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16135 when producing the local key id.
16137 *Richard Levitte <levitte@stacken.kth.se>*
16139 * New option -dhparam in s_server. This allows a DH parameter file to be
16140 stated explicitly. If it is not stated then it tries the first server
16141 certificate file. The previous behaviour hard coded the filename
16146 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16147 a public key to be input or output. For example:
16148 openssl rsa -in key.pem -pubout -out pubkey.pem
16149 Also added necessary DSA public key functions to handle this.
16153 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16154 in the message. This was handled by allowing
16155 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16157 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16159 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16160 to the end of the strings whereas this didn't. This would cause problems
16161 if strings read with d2i_ASN1_bytes() were later modified.
16163 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16165 * Fix for base64 decode bug. When a base64 bio reads only one line of
16166 data and it contains EOF it will end up returning an error. This is
16167 caused by input 46 bytes long. The cause is due to the way base64
16168 BIOs find the start of base64 encoded data. They do this by trying a
16169 trial decode on each line until they find one that works. When they
16170 do a flag is set and it starts again knowing it can pass all the
16171 data directly through the decoder. Unfortunately it doesn't reset
16172 the context it uses. This means that if EOF is reached an attempt
16173 is made to pass two EOFs through the context and this causes the
16174 resulting error. This can also cause other problems as well. As is
16175 usual with these problems it takes *ages* to find and the fix is
16176 trivial: move one line.
16178 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16180 * Ugly workaround to get s_client and s_server working under Windows. The
16181 old code wouldn't work because it needed to select() on sockets and the
16182 tty (for keypresses and to see if data could be written). Win32 only
16183 supports select() on sockets so we select() with a 1s timeout on the
16184 sockets and then see if any characters are waiting to be read, if none
16185 are present then we retry, we also assume we can always write data to
16186 the tty. This isn't nice because the code then blocks until we've
16187 received a complete line of data and it is effectively polling the
16188 keyboard at 1s intervals: however it's quite a bit better than not
16189 working at all :-) A dedicated Windows application might handle this
16190 with an event loop for example.
16194 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16195 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16196 will be called when RSA_sign() and RSA_verify() are used. This is useful
16197 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16198 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16199 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16200 This necessitated the support of an extra signature type NID_md5_sha1
16201 for SSL signatures and modifications to the SSL library to use it instead
16202 of calling RSA_public_decrypt() and RSA_private_encrypt().
16206 * Add new -verify -CAfile and -CApath options to the crl program, these
16207 will lookup a CRL issuers certificate and verify the signature in a
16208 similar way to the verify program. Tidy up the crl program so it
16209 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16210 less strict. It will now permit CRL extensions even if it is not
16211 a V2 CRL: this will allow it to tolerate some broken CRLs.
16215 * Initialize all non-automatic variables each time one of the openssl
16216 sub-programs is started (this is necessary as they may be started
16217 multiple times from the "OpenSSL>" prompt).
16219 *Lennart Bang, Bodo Moeller*
16221 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16222 removing all other RSA functionality (this is what NO_RSA does). This
16223 is so (for example) those in the US can disable those operations covered
16224 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16229 * Non-copying interface to BIO pairs.
16230 (still largely untested)
16234 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16235 ASCII string. This was handled independently in various places before.
16239 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16240 UTF8 strings a character at a time.
16244 * Use client_version from client hello to select the protocol
16245 (s23_srvr.c) and for RSA client key exchange verification
16246 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16250 * Add various utility functions to handle SPKACs, these were previously
16251 handled by poking round in the structure internals. Added new function
16252 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16253 print, verify and generate SPKACs. Based on an original idea from
16254 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16258 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16262 * Allow the config file extension section to be overwritten on the
16263 command line. Based on an original idea from Massimiliano Pala
16264 <madwolf@comune.modena.it>. The new option is called -extensions
16265 and can be applied to ca, req and x509. Also -reqexts to override
16266 the request extensions in req and -crlexts to override the crl extensions
16271 * Add new feature to the SPKAC handling in ca. Now you can include
16272 the same field multiple times by preceding it by "XXXX." for example:
16275 this is the same syntax as used in the req config file.
16279 * Allow certificate extensions to be added to certificate requests. These
16280 are specified in a 'req_extensions' option of the req section of the
16281 config file. They can be printed out with the -text option to req but
16282 are otherwise ignored at present.
16286 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16287 data read consists of only the final block it would not decrypted because
16288 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16289 A misplaced 'break' also meant the decrypted final block might not be
16290 copied until the next read.
16294 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16295 a few extra parameters to the DH structure: these will be useful if
16296 for example we want the value of 'q' or implement X9.42 DH.
16300 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16301 provides hooks that allow the default DSA functions or functions on a
16302 "per key" basis to be replaced. This allows hardware acceleration and
16303 hardware key storage to be handled without major modification to the
16304 library. Also added low level modexp hooks and CRYPTO_EX structure and
16305 associated functions.
16309 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16310 as "read only": it can't be written to and the buffer it points to will
16311 not be freed. Reading from a read only BIO is much more efficient than
16312 a normal memory BIO. This was added because there are several times when
16313 an area of memory needs to be read from a BIO. The previous method was
16314 to create a memory BIO and write the data to it, this results in two
16315 copies of the data and an O(n^2) reading algorithm. There is a new
16316 function BIO_new_mem_buf() which creates a read only memory BIO from
16317 an area of memory. Also modified the PKCS#7 routines to use read only
16322 * Bugfix: ssl23_get_client_hello did not work properly when called in
16323 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16324 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16325 but a retry condition occurred while trying to read the rest.
16329 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16330 NID_pkcs7_encrypted by default: this was wrong since this should almost
16331 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16332 the encrypted data type: this is a more sensible place to put it and it
16333 allows the PKCS#12 code to be tidied up that duplicated this
16338 * Changed obj_dat.pl script so it takes its input and output files on
16339 the command line. This should avoid shell escape redirection problems
16344 * Initial support for certificate extension requests, these are included
16345 in things like Xenroll certificate requests. Included functions to allow
16346 extensions to be obtained and added.
16350 * -crlf option to s_client and s_server for sending newlines as
16351 CRLF (as required by many protocols).
16355 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16357 * Install libRSAglue.a when OpenSSL is built with RSAref.
16359 *Ralf S. Engelschall*
16361 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16363 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16365 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16370 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16371 DH parameters/keys (q is lost during that conversion, but the resulting
16372 DH parameters contain its length).
16374 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16375 much faster than DH_generate_parameters (which creates parameters
16376 where `p = 2*q + 1`), and also the smaller q makes DH computations
16377 much more efficient (160-bit exponentiation instead of 1024-bit
16378 exponentiation); so this provides a convenient way to support DHE
16379 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16380 utter importance to use
16381 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16383 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16384 when such DH parameters are used, because otherwise small subgroup
16385 attacks may become possible!
16389 * Avoid memory leak in i2d_DHparams.
16393 * Allow the -k option to be used more than once in the enc program:
16394 this allows the same encrypted message to be read by multiple recipients.
16398 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16399 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16400 it will always use the numerical form of the OID, even if it has a short
16405 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16406 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16407 otherwise bn_mod_exp was called. In the case of hardware keys for example
16408 no private key components need be present and it might store extra data
16409 in the RSA structure, which cannot be accessed from bn_mod_exp.
16410 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16411 private key operations.
16415 * Added support for SPARC Linux.
16419 * pem_password_cb function type incompatibly changed from
16420 typedef int pem_password_cb(char *buf, int size, int rwflag);
16422 ....(char *buf, int size, int rwflag, void *userdata);
16423 so that applications can pass data to their callbacks:
16424 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16425 additional void * argument, which is just handed through whenever
16426 the password callback is called.
16428 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16430 New function SSL_CTX_set_default_passwd_cb_userdata.
16432 Compatibility note: As many C implementations push function arguments
16433 onto the stack in reverse order, the new library version is likely to
16434 interoperate with programs that have been compiled with the old
16435 pem_password_cb definition (PEM_whatever takes some data that
16436 happens to be on the stack as its last argument, and the callback
16437 just ignores this garbage); but there is no guarantee whatsoever that
16440 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16441 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16442 problems not only on Windows, but also on some Unix platforms.
16443 To avoid problematic command lines, these definitions are now in an
16444 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16445 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16449 * MIPS III/IV assembler module is reimplemented.
16453 * More DES library cleanups: remove references to srand/rand and
16454 delete an unused file.
16458 * Add support for the free Netwide assembler (NASM) under Win32,
16459 since not many people have MASM (ml) and it can be hard to obtain.
16460 This is currently experimental but it seems to work OK and pass all
16461 the tests. Check out INSTALL.W32 for info.
16465 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16466 without temporary keys kept an extra copy of the server key,
16467 and connections with temporary keys did not free everything in case
16472 * New function RSA_check_key and new openssl rsa option -check
16473 for verifying the consistency of RSA keys.
16475 *Ulf Moeller, Bodo Moeller*
16477 * Various changes to make Win32 compile work:
16478 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16479 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16480 comparison" warnings.
16481 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16485 * Add a debugging option to PKCS#5 v2 key generation function: when
16486 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16487 derived keys are printed to stderr.
16491 * Copy the flags in ASN1_STRING_dup().
16493 *Roman E. Pavlov <pre@mo.msk.ru>*
16495 * The x509 application mishandled signing requests containing DSA
16496 keys when the signing key was also DSA and the parameters didn't match.
16498 It was supposed to omit the parameters when they matched the signing key:
16499 the verifying software was then supposed to automatically use the CA's
16500 parameters if they were absent from the end user certificate.
16502 Omitting parameters is no longer recommended. The test was also
16503 the wrong way round! This was probably due to unusual behaviour in
16504 EVP_cmp_parameters() which returns 1 if the parameters match.
16505 This meant that parameters were omitted when they *didn't* match and
16506 the certificate was useless. Certificates signed with 'ca' didn't have
16509 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16511 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16512 The interface is as follows:
16513 Applications can use
16514 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16515 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16516 "off" is now the default.
16517 The library internally uses
16518 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16519 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16520 to disable memory-checking temporarily.
16522 Some inconsistent states that previously were possible (and were
16523 even the default) are now avoided.
16525 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16526 with each memory chunk allocated; this is occasionally more helpful
16527 than just having a counter.
16529 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16531 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16536 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16537 which largely parallels "options", but is for changing API behaviour,
16538 whereas "options" are about protocol behaviour.
16539 Initial "mode" flags are:
16541 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16542 a single record has been written.
16543 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16544 retries use the same buffer location.
16545 (But all of the contents must be
16550 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16553 * Fix problems with no-hmac etc.
16555 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16557 * New functions RSA_get_default_method(), RSA_set_method() and
16558 RSA_get_method(). These allows replacement of RSA_METHODs without having
16559 to mess around with the internals of an RSA structure.
16563 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16564 Also really enable memory leak checks in openssl.c and in some
16567 *Chad C. Mulligan, Bodo Moeller*
16569 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16570 up the length of negative integers. This has now been simplified to just
16571 store the length when it is first determined and use it later, rather
16572 than trying to keep track of where data is copied and updating it to
16574 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16576 * Add a new function PKCS7_signatureVerify. This allows the verification
16577 of a PKCS#7 signature but with the signing certificate passed to the
16578 function itself. This contrasts with PKCS7_dataVerify which assumes the
16579 certificate is present in the PKCS#7 structure. This isn't always the
16580 case: certificates can be omitted from a PKCS#7 structure and be
16581 distributed by "out of band" means (such as a certificate database).
16585 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16586 function prototypes in pem.h, also change util/mkdef.pl to add the
16587 necessary function names.
16591 * mk1mf.pl (used by Windows builds) did not properly read the
16592 options set by Configure in the top level Makefile, and Configure
16593 was not even able to write more than one option correctly.
16594 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16598 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16599 file to be loaded from a BIO or FILE pointer. The BIO version will
16600 for example allow memory BIOs to contain config info.
16604 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16605 Whoever hopes to achieve shared-library compatibility across versions
16606 must use this, not the compile-time macro.
16607 (Exercise 0.9.4: Which is the minimum library version required by
16609 Note: All this applies only to multi-threaded programs, others don't
16614 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16615 through a BIO pair triggered the default case, i.e.
16616 SSLerr(...,SSL_R_UNKNOWN_STATE).
16620 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16621 can use the SSL library even if none of the specific BIOs is
16626 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16627 for the encoded length.
16629 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16631 * Add initial documentation of the X509V3 functions.
16635 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16636 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16637 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16638 secure PKCS#8 private key format with a high iteration count.
16642 * Fix determination of Perl interpreter: A perl or perl5
16643 *directory* in $PATH was also accepted as the interpreter.
16645 *Ralf S. Engelschall*
16647 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16648 wrong with it but it was very old and did things like calling
16649 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16650 unusual formatting.
16654 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16655 to use the new extension code.
16659 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16660 with macros. This should make it easier to change their form, add extra
16661 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16666 * Add to configuration table a new entry that can specify an alternative
16667 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16668 according to Mark Crispin <MRC@Panda.COM>.
16672 * DES CBC did not update the IV. Weird.
16676 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16677 Changing the behaviour of the former might break existing programs --
16678 where IV updating is needed, des_ncbc_encrypt can be used.
16681 * When bntest is run from "make test" it drives bc to check its
16682 calculations, as well as internally checking them. If an internal check
16683 fails, it needs to cause bc to give a non-zero result or make test carries
16684 on without noticing the failure. Fixed.
16688 * DES library cleanups.
16692 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16693 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16694 ciphers. NOTE: although the key derivation function has been verified
16695 against some published test vectors it has not been extensively tested
16696 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16701 * Instead of "mkdir -p", which is not fully portable, use new
16702 Perl script "util/mkdir-p.pl".
16706 * Rewrite the way password based encryption (PBE) is handled. It used to
16707 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16708 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16709 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16710 the 'parameter' field of the AlgorithmIdentifier is passed to the
16711 underlying key generation function so it must do its own ASN1 parsing.
16712 This has also changed the EVP_PBE_CipherInit() function which now has a
16713 'parameter' argument instead of literal salt and iteration count values
16714 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16718 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16719 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16720 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16721 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16722 value was just used as a "magic string" and not used directly its
16723 value doesn't matter.
16727 * Introduce some semblance of const correctness to BN. Shame C doesn't
16732 * "linux-sparc64" configuration (ultrapenguin).
16734 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16735 "linux-sparc" configuration.
16737 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16739 * config now generates no-xxx options for missing ciphers.
16743 * Support the EBCDIC character set (work in progress).
16744 File ebcdic.c not yet included because it has a different license.
16746 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16748 * Support BS2000/OSD-POSIX.
16750 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16752 * Make callbacks for key generation use `void *` instead of `char *`.
16756 * Make S/MIME samples compile (not yet tested).
16760 * Additional typesafe stacks.
16764 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16768 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16770 * New configuration variant "sco5-gcc".
16772 * Updated some demos.
16774 *Sean O Riordain, Wade Scholine*
16776 * Add missing BIO_free at exit of pkcs12 application.
16780 * Fix memory leak in conf.c.
16784 * Updates for Win32 to assembler version of MD5.
16788 * Set #! path to perl in apps/der_chop to where we found it
16789 instead of using a fixed path.
16793 * SHA library changes for irix64-mips4-cc.
16797 * Improvements for VMS support.
16801 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16803 * Bignum library bug fix. IRIX 6 passes "make test" now!
16804 This also avoids the problems with SC4.2 and unpatched SC5.
16806 *Andy Polyakov <appro@fy.chalmers.se>*
16808 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16809 These are required because of the typesafe stack would otherwise break
16810 existing code. If old code used a structure member which used to be STACK
16811 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16812 sk_num or sk_value it would produce an error because the num, data members
16813 are not present in STACK_OF. Now it just produces a warning. sk_set
16814 replaces the old method of assigning a value to sk_value
16815 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16816 that does this will no longer work (and should use sk_set instead) but
16817 this could be regarded as a "questionable" behaviour anyway.
16821 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16822 correctly handle encrypted S/MIME data.
16826 * Change type of various DES function arguments from des_cblock
16827 (which means, in function argument declarations, pointer to char)
16828 to des_cblock * (meaning pointer to array with 8 char elements),
16829 which allows the compiler to do more typechecking; it was like
16830 that back in SSLeay, but with lots of ugly casts.
16832 Introduce new type const_des_cblock.
16836 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16837 problems: find RecipientInfo structure that matches recipient certificate
16838 and initialise the ASN1 structures properly based on passed cipher.
16842 * Belatedly make the BN tests actually check the results.
16846 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16847 to and from BNs: it was completely broken. New compilation option
16848 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16849 key elements as negative integers.
16853 * Reorganize and speed up MD5.
16855 *Andy Polyakov <appro@fy.chalmers.se>*
16859 *Richard Levitte <richard@levitte.org>*
16861 * New option -out to asn1parse to allow the parsed structure to be
16862 output to a file. This is most useful when combined with the -strparse
16863 option to examine the output of things like OCTET STRINGS.
16867 * Make SSL library a little more fool-proof by not requiring any longer
16868 that `SSL_set_{accept,connect}_state` be called before
16869 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16870 in many applications because usually everything *appeared* to work as
16871 intended anyway -- now it really works as intended).
16875 * Move openssl.cnf out of lib/.
16879 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16880 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16881 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16883 *Ralf S. Engelschall*
16885 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16886 handle PKCS#7 enveloped data properly.
16888 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16890 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16891 copying pointers. The cert_st handling is changed by this in
16892 various ways (and thus what used to be known as ctx->default_cert
16893 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
16894 any longer when s->cert does not give us what we need).
16895 ssl_cert_instantiate becomes obsolete by this change.
16896 As soon as we've got the new code right (possibly it already is?),
16897 we have solved a couple of bugs of the earlier code where s->cert
16898 was used as if it could not have been shared with other SSL structures.
16900 Note that using the SSL API in certain dirty ways now will result
16901 in different behaviour than observed with earlier library versions:
16902 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
16903 does not influence s as it used to.
16905 In order to clean up things more thoroughly, inside SSL_SESSION
16906 we don't use CERT any longer, but a new structure SESS_CERT
16907 that holds per-session data (if available); currently, this is
16908 the peer's certificate chain and, for clients, the server's certificate
16909 and temporary key. CERT holds only those values that can have
16910 meaningful defaults in an SSL_CTX.
16914 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16915 from the internal representation. Various PKCS#7 fixes: remove some
16916 evil casts and set the enc_dig_alg field properly based on the signing
16921 * Allow PKCS#12 password to be set from the command line or the
16922 environment. Let 'ca' get its config file name from the environment
16923 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16928 * Allow certificate policies extension to use an IA5STRING for the
16929 organization field. This is contrary to the PKIX definition but
16930 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16935 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16936 without disallowing inline assembler and the like for non-pedantic builds.
16940 * Support Borland C++ builder.
16942 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16948 * SHA-1 cleanups and performance enhancements.
16950 *Andy Polyakov <appro@fy.chalmers.se>*
16952 * Sparc v8plus assembler for the bignum library.
16954 *Andy Polyakov <appro@fy.chalmers.se>*
16956 * Accept any -xxx and +xxx compiler options in Configure.
16960 * Update HPUX configuration.
16964 * Add missing `sk_<type>_unshift()` function to safestack.h
16966 *Ralf S. Engelschall*
16968 * New function SSL_CTX_use_certificate_chain_file that sets the
16969 "extra_cert"s in addition to the certificate. (This makes sense
16970 only for "PEM" format files, as chains as a whole are not
16975 * Support verify_depth from the SSL API.
16976 x509_vfy.c had what can be considered an off-by-one-error:
16977 Its depth (which was not part of the external interface)
16978 was actually counting the number of certificates in a chain;
16979 now it really counts the depth.
16983 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16984 instead of X509err, which often resulted in confusing error
16985 messages since the error codes are not globally unique
16986 (e.g. an alleged error in ssl3_accept when a certificate
16987 didn't match the private key).
16989 * New function SSL_CTX_set_session_id_context that allows to set a default
16990 value (so that you don't need SSL_set_session_id_context for each
16991 connection using the SSL_CTX).
16995 * OAEP decoding bug fix.
16999 * Support INSTALL_PREFIX for package builders, as proposed by
17004 * New Configure options "threads" and "no-threads". For systems
17005 where the proper compiler options are known (currently Solaris
17006 and Linux), "threads" is the default.
17010 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17014 * Install various scripts to $(OPENSSLDIR)/misc, not to
17015 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17016 such as /usr/local/bin.
17020 * "make linux-shared" to build shared libraries.
17022 *Niels Poppe <niels@netbox.org>*
17024 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17028 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17029 extension adding in x509 utility.
17033 * Remove NOPROTO sections and error code comments.
17037 * Partial rewrite of the DEF file generator to now parse the ANSI
17042 * New Configure options --prefix=DIR and --openssldir=DIR.
17046 * Complete rewrite of the error code script(s). It is all now handled
17047 by one script at the top level which handles error code gathering,
17048 header rewriting and C source file generation. It should be much better
17049 than the old method: it now uses a modified version of Ulf's parser to
17050 read the ANSI prototypes in all header files (thus the old K&R definitions
17051 aren't needed for error creation any more) and do a better job of
17052 translating function codes into names. The old 'ASN1 error code embedded
17053 in a comment' is no longer necessary and it doesn't use .err files which
17054 have now been deleted. Also the error code call doesn't have to appear all
17055 on one line (which resulted in some large lines...).
17059 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17063 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17064 0 (which usually indicates a closed connection), but continue reading.
17068 * Fix some race conditions.
17072 * Add support for CRL distribution points extension. Add Certificate
17073 Policies and CRL distribution points documentation.
17077 * Move the autogenerated header file parts to crypto/opensslconf.h.
17081 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17082 8 of keying material. Merlin has also confirmed interop with this fix
17083 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17085 *Merlin Hughes <merlin@baltimore.ie>*
17087 * Fix lots of warnings.
17089 *Richard Levitte <levitte@stacken.kth.se>*
17091 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17092 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17094 *Richard Levitte <levitte@stacken.kth.se>*
17096 * Fix problems with sizeof(long) == 8.
17098 *Andy Polyakov <appro@fy.chalmers.se>*
17100 * Change functions to ANSI C.
17104 * Fix typos in error codes.
17106 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17108 * Remove defunct assembler files from Configure.
17112 * SPARC v8 assembler BIGNUM implementation.
17114 *Andy Polyakov <appro@fy.chalmers.se>*
17116 * Support for Certificate Policies extension: both print and set.
17117 Various additions to support the r2i method this uses.
17121 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17122 return a const string when you are expecting an allocated buffer.
17126 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17127 types DirectoryString and DisplayText.
17131 * Add code to allow r2i extensions to access the configuration database,
17132 add an LHASH database driver and add several ctx helper functions.
17136 * Fix an evil bug in bn_expand2() which caused various BN functions to
17137 fail when they extended the size of a BIGNUM.
17141 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17142 support typesafe stack.
17146 * Fix typo in SSL_[gs]et_options().
17148 *Nils Frostberg <nils@medcom.se>*
17150 * Delete various functions and files that belonged to the (now obsolete)
17151 old X509V3 handling code.
17155 * New Configure option "rsaref".
17159 * Don't auto-generate pem.h.
17163 * Introduce type-safe ASN.1 SETs.
17167 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17169 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17171 * Introduce type-safe STACKs. This will almost certainly break lots of code
17172 that links with OpenSSL (well at least cause lots of warnings), but fear
17173 not: the conversion is trivial, and it eliminates loads of evil casts. A
17174 few STACKed things have been converted already. Feel free to convert more.
17175 In the fullness of time, I'll do away with the STACK type altogether.
17179 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17180 specified in `<certfile>` by updating the entry in the index.txt file.
17181 This way one no longer has to edit the index.txt file manually for
17182 revoking a certificate. The -revoke option does the gory details now.
17184 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17186 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17187 `-text` option at all and this way the `-noout -text` combination was
17188 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17190 *Ralf S. Engelschall*
17192 * Make sure a corresponding plain text error message exists for the
17193 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17194 verify callback function determined that a certificate was revoked.
17196 *Ralf S. Engelschall*
17198 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17199 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17200 all available ciphers including rc5, which was forgotten until now.
17201 In order to let the testing shell script know which algorithms
17202 are available, a new (up to now undocumented) command
17203 `openssl list-cipher-commands` is used.
17207 * Bugfix: s_client occasionally would sleep in select() when
17208 it should have checked SSL_pending() first.
17212 * New functions DSA_do_sign and DSA_do_verify to provide access to
17213 the raw DSA values prior to ASN.1 encoding.
17217 * Tweaks to Configure
17219 *Niels Poppe <niels@netbox.org>*
17221 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17226 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17230 * New config option to avoid instructions that are illegal on the 80386.
17231 The default code is faster, but requires at least a 486.
17235 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17236 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17237 same as SSL2_VERSION anyway.
17241 * New "-showcerts" option for s_client.
17245 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17246 application. Various cleanups and fixes.
17250 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17251 modify error routines to work internally. Add error codes and PBE init
17252 to library startup routines.
17256 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17257 packing functions to asn1 and evp. Changed function names and error
17258 codes along the way.
17262 * PKCS12 integration: and so it begins... First of several patches to
17263 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17264 objects to objects.h
17268 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17269 and display support for Thawte strong extranet extension.
17273 * Add LinuxPPC support.
17275 *Jeff Dubrule <igor@pobox.org>*
17277 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17278 bn_div_words in alpha.s.
17280 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17282 * Make sure the RSA OAEP test is skipped under -DRSAref because
17283 OAEP isn't supported when OpenSSL is built with RSAref.
17285 *Ulf Moeller <ulf@fitug.de>*
17287 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17288 so they no longer are missing under -DNOPROTO.
17290 *Soren S. Jorvang <soren@t.dk>*
17292 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17294 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17295 doesn't work when the session is reused. Coming soon!
17299 * Fix a security hole, that allows sessions to be reused in the wrong
17300 context thus bypassing client cert protection! All software that uses
17301 client certs and session caches in multiple contexts NEEDS PATCHING to
17302 allow session reuse! A fuller solution is in the works.
17304 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17306 * Some more source tree cleanups (removed obsolete files
17307 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17308 permission on "config" script to be executable) and a fix for the INSTALL
17311 *Ulf Moeller <ulf@fitug.de>*
17313 * Remove some legacy and erroneous uses of malloc, free instead of
17316 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17318 * Make rsa_oaep_test return non-zero on error.
17320 *Ulf Moeller <ulf@fitug.de>*
17322 * Add support for native Solaris shared libraries. Configure
17323 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17324 if someone would make that last step automatic.
17326 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17328 * ctx_size was not built with the right compiler during "make links". Fixed.
17332 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17333 except NULL ciphers". This means the default cipher list will no longer
17334 enable NULL ciphers. They need to be specifically enabled e.g. with
17335 the string "DEFAULT:eNULL".
17339 * Fix to RSA private encryption routines: if p < q then it would
17340 occasionally produce an invalid result. This will only happen with
17341 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17345 * Be less restrictive and allow also `perl util/perlpath.pl
17346 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17347 because this way one can also use an interpreter named `perl5' (which is
17348 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17349 installed as `perl').
17351 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17353 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17355 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17357 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17358 advapi32.lib to Win32 build and change the pem test comparison
17359 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17360 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17361 and crypto/des/ede_cbcm_enc.c.
17365 * DES quad checksum was broken on big-endian architectures. Fixed.
17369 * Comment out two functions in bio.h that aren't implemented. Fix up the
17370 Win32 test batch file so it (might) work again. The Win32 test batch file
17371 is horrible: I feel ill....
17375 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17376 in e_os.h. Audit of header files to check ANSI and non ANSI
17377 sections: 10 functions were absent from non ANSI section and not exported
17378 from Windows DLLs. Fixed up libeay.num for new functions.
17382 * Make `openssl version' output lines consistent.
17384 *Ralf S. Engelschall*
17386 * Fix Win32 symbol export lists for BIO functions: Added
17387 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17388 to ms/libeay{16,32}.def.
17390 *Ralf S. Engelschall*
17392 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17393 fine under Unix and passes some trivial tests I've now added. But the
17394 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17395 added to make sure no one expects that this stuff really works in the
17396 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17397 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17400 *Ralf S. Engelschall*
17402 * Fix the generation of two part addresses in perl.
17404 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17406 * Add config entry for Linux on MIPS.
17408 *John Tobey <jtobey@channel1.com>*
17410 * Make links whenever Configure is run, unless we are on Windoze.
17414 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17415 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17420 * Add a useful kludge to allow package maintainers to specify compiler and
17421 other platforms details on the command line without having to patch the
17422 Configure script every time: One now can use
17423 `perl Configure <id>:<details>`,
17424 i.e. platform ids are allowed to have details appended
17425 to them (separated by colons). This is treated as there would be a static
17426 pre-configured entry in Configure's %table under key `<id>` with value
17427 `<details>` and `perl Configure <id>` is called. So, when you want to
17428 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17429 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17430 now, which overrides the FreeBSD-elf entry on-the-fly.
17432 *Ralf S. Engelschall*
17434 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17438 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17439 on the `perl Configure ...' command line. This way one can compile
17440 OpenSSL libraries with Position Independent Code (PIC) which is needed
17441 for linking it into DSOs.
17443 *Ralf S. Engelschall*
17445 * Remarkably, export ciphers were totally broken and no-one had noticed!
17450 * Cleaned up the LICENSE document: The official contact for any license
17451 questions now is the OpenSSL core team under openssl-core@openssl.org.
17452 And add a paragraph about the dual-license situation to make sure people
17453 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17454 to the OpenSSL toolkit.
17456 *Ralf S. Engelschall*
17458 * General source tree makefile cleanups: Made `making xxx in yyy...'
17459 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17460 Additionally cleaned up the `make links' target: Remove unnecessary
17461 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17462 to speed processing and no longer clutter the display with confusing
17463 stuff. Instead only the actually done links are displayed.
17465 *Ralf S. Engelschall*
17467 * Permit null encryption ciphersuites, used for authentication only. It used
17468 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17469 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17474 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17475 signed attributes when verifying signatures (this would break them),
17476 the detached data encoding was wrong and public keys obtained using
17477 X509_get_pubkey() weren't freed.
17481 * Add text documentation for the BUFFER functions. Also added a work around
17482 to a Win95 console bug. This was triggered by the password read stuff: the
17483 last character typed gets carried over to the next fread(). If you were
17484 generating a new cert request using 'req' for example then the last
17485 character of the passphrase would be CR which would then enter the first
17490 * Added the new 'Includes OpenSSL Cryptography Software' button as
17491 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17492 button and can be used by applications based on OpenSSL to show the
17493 relationship to the OpenSSL project.
17495 *Ralf S. Engelschall*
17497 * Remove confusing variables in function signatures in files
17498 ssl/ssl_lib.c and ssl/ssl.h.
17500 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17502 * Don't install bss_file.c under PREFIX/include/
17504 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17506 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17507 functions that return function pointers and has support for NT specific
17508 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17509 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17510 unsigned to signed types: this was killing the Win32 compile.
17514 * Add new certificate file to stack functions,
17515 SSL_add_dir_cert_subjects_to_stack() and
17516 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17517 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17518 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17519 This means that Apache-SSL and similar packages don't have to mess around
17520 to add as many CAs as they want to the preferred list.
17524 * Experiment with doxygen documentation. Currently only partially applied to
17526 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17527 openssl.doxy as the configuration file.
17531 * Get rid of remaining C++-style comments which strict C compilers hate.
17533 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17535 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17536 compiled in by default: it has problems with large keys.
17540 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17541 DH private keys and/or callback functions which directly correspond to
17542 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17543 is needed for applications which have to configure certificates on a
17544 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17546 For the RSA certificate situation is makes no difference, but
17547 for the DSA certificate situation this fixes the "no shared cipher"
17548 problem where the OpenSSL cipher selection procedure failed because the
17549 temporary keys were not overtaken from the context and the API provided
17550 no way to reconfigure them.
17551 The new functions now let applications reconfigure the stuff and they
17552 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17553 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17554 non-public-API function ssl_cert_instantiate() is used as a helper
17555 function and also to reduce code redundancy inside ssl_rsa.c.
17557 *Ralf S. Engelschall*
17559 * Move s_server -dcert and -dkey options out of the undocumented feature
17560 area because they are useful for the DSA situation and should be
17561 recognized by the users.
17563 *Ralf S. Engelschall*
17565 * Fix the cipher decision scheme for export ciphers: the export bits are
17566 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17567 SSL_EXP_MASK. So, the original variable has to be used instead of the
17568 already masked variable.
17570 *Richard Levitte <levitte@stacken.kth.se>*
17572 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17574 *Richard Levitte <levitte@stacken.kth.se>*
17576 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17577 from `int` to `unsigned int` because it is a length and initialized by
17578 EVP_DigestFinal() which expects an `unsigned int *`.
17580 *Richard Levitte <levitte@stacken.kth.se>*
17582 * Don't hard-code path to Perl interpreter on shebang line of Configure
17583 script. Instead use the usual Shell->Perl transition trick.
17585 *Ralf S. Engelschall*
17587 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17588 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17589 -noout -modulus' as it's already the case for `openssl rsa -noout
17590 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17591 currently the public key is printed (a decision which was already done by
17592 `openssl dsa -modulus' in the past) which serves a similar purpose.
17593 Additionally the NO_RSA no longer completely removes the whole -modulus
17594 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17597 *Ralf S. Engelschall*
17599 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17600 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17602 *Arne Ansper <arne@ats.cyber.ee>*
17604 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17605 to be added. Now both 'req' and 'ca' can use new objects defined in the
17610 * Add cool BIO that does syslog (or event log on NT).
17612 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17614 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17615 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17616 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17617 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17621 * Add preliminary config info for new extension code.
17625 * Make RSA_NO_PADDING really use no padding.
17627 *Ulf Moeller <ulf@fitug.de>*
17629 * Generate errors when private/public key check is done.
17633 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17634 for some CRL extensions and new objects added.
17638 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17639 key usage extension and fuller support for authority key id.
17643 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17644 padding method for RSA, which is recommended for new applications in PKCS
17645 #1 v2.0 (RFC 2437, October 1998).
17646 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17647 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17648 against Bleichbacher's attack on RSA.
17649 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17652 * Updates to the new SSL compression code
17654 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17656 * Fix so that the version number in the master secret, when passed
17657 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17658 (because the server will not accept higher), that the version number
17659 is 0x03,0x01, not 0x03,0x00
17661 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17663 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17664 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17665 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17669 * Support for RAW extensions where an arbitrary extension can be
17670 created by including its DER encoding. See apps/openssl.cnf for
17675 * Make sure latest Perl versions don't interpret some generated C array
17676 code as Perl array code in the crypto/err/err_genc.pl script.
17678 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17680 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17681 not many people have the assembler. Various Win32 compilation fixes and
17682 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17683 build instructions.
17687 * Modify configure script 'Configure' to automatically create crypto/date.h
17688 file under Win32 and also build pem.h from pem.org. New script
17689 util/mkfiles.pl to create the MINFO file on environments that can't do a
17690 'make files': perl util/mkfiles.pl >MINFO should work.
17694 * Major rework of DES function declarations, in the pursuit of correctness
17695 and purity. As a result, many evil casts evaporated, and some weirdness,
17696 too. You may find this causes warnings in your code. Zapping your evil
17697 casts will probably fix them. Mostly.
17701 * Fix for a typo in asn1.h. Bug fix to object creation script
17702 obj_dat.pl. It considered a zero in an object definition to mean
17703 "end of object": none of the objects in objects.h have any zeros
17704 so it wasn't spotted.
17706 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17708 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17709 Masking (CBCM). In the absence of test vectors, the best I have been able
17710 to do is check that the decrypt undoes the encrypt, so far. Send me test
17711 vectors if you have them.
17715 * Correct calculation of key length for export ciphers (too much space was
17716 allocated for null ciphers). This has not been tested!
17720 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17721 message is now correct (it understands "crypto" and "ssl" on its
17722 command line). There is also now an "update" option. This will update
17723 the util/ssleay.num and util/libeay.num files with any new functions.
17725 perl util/mkdef.pl crypto ssl update
17726 it will update them.
17730 * Overhauled the Perl interface:
17731 - ported BN stuff to OpenSSL's different BN library
17732 - made the perl/ source tree CVS-aware
17733 - renamed the package from SSLeay to OpenSSL (the files still contain
17734 their history because I've copied them in the repository)
17735 - removed obsolete files (the test scripts will be replaced
17736 by better Test::Harness variants in the future)
17738 *Ralf S. Engelschall*
17740 * First cut for a very conservative source tree cleanup:
17741 1. merge various obsolete readme texts into doc/ssleay.txt
17742 where we collect the old documents and readme texts.
17743 2. remove the first part of files where I'm already sure that we no
17744 longer need them because of three reasons: either they are just temporary
17745 files which were left by Eric or they are preserved original files where
17746 I've verified that the diff is also available in the CVS via "cvs diff
17747 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17748 the crypto/md/ stuff).
17750 *Ralf S. Engelschall*
17752 * More extension code. Incomplete support for subject and issuer alt
17753 name, issuer and authority key id. Change the i2v function parameters
17754 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17755 what that's for :-) Fix to ASN1 macro which messed up
17756 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17760 * Preliminary support for ENUMERATED type. This is largely copied from the
17765 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17767 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17769 * Make sure `make rehash` target really finds the `openssl` program.
17771 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17773 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17774 like to hear about it if this slows down other processors.
17778 * Add CygWin32 platform information to Configure script.
17780 *Alan Batie <batie@aahz.jf.intel.com>*
17782 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17784 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17786 * New program nseq to manipulate netscape certificate sequences
17790 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17795 * Fixes to BN code. Previously the default was to define BN_RECURSION
17796 but the BN code had some problems that would cause failures when
17797 doing certificate verification and some other functions.
17799 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17801 * Add ASN1 and PEM code to support netscape certificate sequences.
17805 * Add ASN1 and PEM code to support netscape certificate sequences.
17809 * Add several PKIX and private extended key usage OIDs.
17813 * Modify the 'ca' program to handle the new extension code. Modify
17814 openssl.cnf for new extension format, add comments.
17818 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17819 and add a sample to openssl.cnf so req -x509 now adds appropriate
17824 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17825 error code, add initial support to X509_print() and x509 application.
17829 * Takes a deep breath and start adding X509 V3 extension support code. Add
17830 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17831 stuff is currently isolated and isn't even compiled yet.
17835 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17836 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17837 Removed the versions check from X509 routines when loading extensions:
17838 this allows certain broken certificates that don't set the version
17839 properly to be processed.
17843 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17844 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17845 can still be regenerated with "make depend".
17849 * Spelling mistake in C version of CAST-128.
17851 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17853 * Changes to the error generation code. The perl script err-code.pl
17854 now reads in the old error codes and retains the old numbers, only
17855 adding new ones if necessary. It also only changes the .err files if new
17856 codes are added. The makefiles have been modified to only insert errors
17857 when needed (to avoid needlessly modifying header files). This is done
17858 by only inserting errors if the .err file is newer than the auto generated
17859 C file. To rebuild all the error codes from scratch (the old behaviour)
17860 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17861 or delete all the .err files.
17865 * CAST-128 was incorrectly implemented for short keys. The C version has
17866 been fixed, but is untested. The assembler versions are also fixed, but
17867 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17868 to regenerate it if needed.
17869 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17870 Hagino <itojun@kame.net>*
17872 * File was opened incorrectly in randfile.c.
17874 *Ulf Möller <ulf@fitug.de>*
17876 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17877 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17878 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17879 al: it's just almost always a UTCTime. Note this patch adds new error
17880 codes so do a "make errors" if there are problems.
17884 * Correct Linux 1 recognition in config.
17886 *Ulf Möller <ulf@fitug.de>*
17888 * Remove pointless MD5 hash when using DSA keys in ca.
17890 *Anonymous <nobody@replay.com>*
17892 * Generate an error if given an empty string as a cert directory. Also
17893 generate an error if handed NULL (previously returned 0 to indicate an
17894 error, but didn't set one).
17896 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17898 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17902 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17903 parameters. This was causing a warning which killed off the Win32 compile.
17907 * Remove C++ style comments from crypto/bn/bn_local.h.
17909 *Neil Costigan <neil.costigan@celocom.com>*
17911 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17912 based on a text string, looking up short and long names and finally
17913 "dot" format. The "dot" format stuff didn't work. Added new function
17914 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17915 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17916 OID is not part of the table.
17920 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17921 X509_LOOKUP_by_alias().
17925 * Sort openssl functions by name.
17929 * Get the gendsa program working (hopefully) and add it to app list. Remove
17930 encryption from sample DSA keys (in case anyone is interested the password
17935 * Make *all* `*_free` functions accept a NULL pointer.
17937 *Frans Heymans <fheymans@isaserver.be>*
17939 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17942 *Anonymous <nobody@replay.com>*
17944 * s_server should send the CAfile as acceptable CAs, not its own cert.
17946 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17948 * Don't blow it for numeric -newkey arguments to apps/req.
17950 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17952 * Temp key "for export" tests were wrong in s3_srvr.c.
17954 *Anonymous <nobody@replay.com>*
17956 * Add prototype for temp key callback functions
17957 SSL_CTX_set_tmp_{rsa,dh}_callback().
17961 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17962 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17966 * X509_name_add_entry() freed the wrong thing after an error.
17968 *Arne Ansper <arne@ats.cyber.ee>*
17970 * rsa_eay.c would attempt to free a NULL context.
17972 *Arne Ansper <arne@ats.cyber.ee>*
17974 * BIO_s_socket() had a broken should_retry() on Windoze.
17976 *Arne Ansper <arne@ats.cyber.ee>*
17978 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17980 *Arne Ansper <arne@ats.cyber.ee>*
17982 * Make sure the already existing X509_STORE->depth variable is initialized
17983 in X509_STORE_new(), but document the fact that this variable is still
17984 unused in the certificate verification process.
17986 *Ralf S. Engelschall*
17988 * Fix the various library and apps files to free up pkeys obtained from
17989 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17993 * Fix reference counting in X509_PUBKEY_get(). This makes
17994 demos/maurice/example2.c work, amongst others, probably.
17996 *Steve Henson and Ben Laurie*
17998 * First cut of a cleanup for apps/. First the `ssleay` program is now named
17999 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18000 are no longer created. This way we have a single and consistent command
18001 line interface `openssl <command>`, similar to `cvs <command>`.
18003 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18005 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18006 BIT STRING wrapper always have zero unused bits.
18010 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18014 * Make the top-level INSTALL documentation easier to understand.
18018 * Makefiles updated to exit if an error occurs in a sub-directory
18019 make (including if user presses ^C) [Paul Sutton]
18021 * Make Montgomery context stuff explicit in RSA data structure.
18025 * Fix build order of pem and err to allow for generated pem.h.
18029 * Fix renumbering bug in X509_NAME_delete_entry().
18033 * Enhanced the err-ins.pl script so it makes the error library number
18034 global and can add a library name. This is needed for external ASN1 and
18035 other error libraries.
18039 * Fixed sk_insert which never worked properly.
18043 * Fix ASN1 macros so they can handle indefinite length constructed
18044 EXPLICIT tags. Some non standard certificates use these: they can now
18049 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18050 into a single doc/ssleay.txt bundle. This way the information is still
18051 preserved but no longer messes up this directory. Now it's new room for
18052 the new set of documentation files.
18054 *Ralf S. Engelschall*
18056 * SETs were incorrectly DER encoded. This was a major pain, because they
18057 shared code with SEQUENCEs, which aren't coded the same. This means that
18058 almost everything to do with SETs or SEQUENCEs has either changed name or
18059 number of arguments.
18061 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18063 * Fix test data to work with the above.
18067 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18068 was already fixed by Eric for 0.9.1 it seems.
18070 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18072 * Autodetect FreeBSD3.
18076 * Fix various bugs in Configure. This affects the following platforms:
18080 unixware-2.0-pentium
18085 * Eliminate generated files from CVS. Reorder tests to regenerate files
18086 before they are needed.
18090 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18094 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18096 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18097 changed SSLeay to OpenSSL in version strings.
18099 *Ralf S. Engelschall*
18101 * Some fixups to the top-level documents.
18105 * Fixed the nasty bug where rsaref.h was not found under compile-time
18106 because the symlink to include/ was missing.
18108 *Ralf S. Engelschall*
18110 * Incorporated the popular no-RSA/DSA-only patches
18111 which allow to compile a RSA-free SSLeay.
18113 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18115 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18116 when "ssleay" is still not found.
18118 *Ralf S. Engelschall*
18120 * Added more platforms to Configure: Cray T3E, HPUX 11,
18122 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18124 * Updated the README file.
18126 *Ralf S. Engelschall*
18128 * Added various .cvsignore files in the CVS repository subdirs
18129 to make a "cvs update" really silent.
18131 *Ralf S. Engelschall*
18133 * Recompiled the error-definition header files and added
18134 missing symbols to the Win32 linker tables.
18136 *Ralf S. Engelschall*
18138 * Cleaned up the top-level documents;
18139 o new files: CHANGES and LICENSE
18140 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18141 o merged COPYRIGHT into LICENSE
18142 o removed obsolete TODO file
18143 o renamed MICROSOFT to INSTALL.W32
18145 *Ralf S. Engelschall*
18147 * Removed dummy files from the 0.9.1b source tree:
18148 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18149 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18150 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18151 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18152 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18154 *Ralf S. Engelschall*
18156 * Added various platform portability fixes.
18160 * The Genesis of the OpenSSL rpject:
18161 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18162 Young and Tim J. Hudson created while they were working for C2Net until
18165 *The OpenSSL Project*
18167 ### Changes between 0.9.0b and 0.9.1b [not released]
18169 * Updated a few CA certificates under certs/
18173 * Changed some BIGNUM api stuff.
18177 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18178 DGUX x86, Linux Alpha, etc.
18182 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18183 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18188 * Add -strparse option to asn1pars program which parses nested
18191 *Dr Stephen Henson <shenson@bigfoot.com>*
18193 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18197 * DSA fix for "ca" program.
18201 * Added "-genkey" option to "dsaparam" program.
18205 * Added RIPE MD160 (rmd160) message digest.
18209 * Added -a (all) option to "ssleay version" command.
18213 * Added PLATFORM define which is the id given to Configure.
18217 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18221 * Extended the ASN.1 parser routines.
18225 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18229 * Added a BN_CTX to the BN library.
18233 * Fixed the weak key values in DES library
18237 * Changed API in EVP library for cipher aliases.
18241 * Added support for RC2/64bit cipher.
18245 * Converted the lhash library to the crypto/mem.c functions.
18249 * Added more recognized ASN.1 object ids.
18253 * Added more RSA padding checks for SSL/TLS.
18257 * Added BIO proxy/filter functionality.
18261 * Added extra_certs to SSL_CTX which can be used
18262 send extra CA certificates to the client in the CA cert chain sending
18263 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18267 * Now Fortezza is denied in the authentication phase because
18268 this is key exchange mechanism is not supported by SSLeay at all.
18272 * Additional PKCS1 checks.
18276 * Support the string "TLSv1" for all TLS v1 ciphers.
18280 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18281 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18285 * Fixed a few memory leaks.
18289 * Fixed various code and comment typos.
18293 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18294 bytes sent in the client random.
18296 *Edward Bishop <ebishop@spyglass.com>*
18300 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18301 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18302 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18303 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18304 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18305 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18306 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18307 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18308 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18309 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18310 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18311 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18312 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18313 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18314 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18315 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18316 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18317 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18318 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18319 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18320 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18321 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18322 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18323 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18324 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18325 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18326 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18327 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18328 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18329 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18330 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18331 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18332 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18333 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18334 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18335 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18336 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18337 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18338 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18339 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18340 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18341 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18342 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18343 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18344 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18345 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18346 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18347 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18348 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18349 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18350 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18351 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18352 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18353 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18354 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18355 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18356 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18357 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18358 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18359 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18360 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18361 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18362 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18363 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18364 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18365 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18366 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18367 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18368 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18369 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18370 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18371 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18372 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18373 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18374 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18375 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18376 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18377 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18378 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18379 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18380 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18381 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18382 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18383 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18384 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18385 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18386 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18387 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18388 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18389 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18390 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18391 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18392 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18393 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18394 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18395 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18396 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18397 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18398 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18399 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18400 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18401 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18402 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18403 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18404 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18405 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18406 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18407 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18408 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18409 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18410 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18411 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18412 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18413 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18414 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18415 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18416 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18417 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18418 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18419 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18420 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18421 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18422 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18423 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18424 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18425 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18426 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18427 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18428 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18429 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18430 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18431 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18432 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18433 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18434 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18435 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18436 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18437 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18438 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18439 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18440 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18441 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18442 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18443 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18444 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18445 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18446 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18447 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18448 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18449 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18450 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18451 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18452 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18453 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18454 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18455 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18456 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18457 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18458 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18459 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655