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/
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx] ###
27 * The test suite is changed to preserve results of each test recipe.
28 A new directory test-runs/ with subdirectories named like the
29 test recipes are created in the build tree for this purpose.
33 * The command line utilities ecparam and ec have been deprecated. Instead
34 use the pkeyparam, pkey and genpkey programs.
38 * All of the low level RSA functions have been deprecated including:
40 RSA_new_method, RSA_bits, RSA_size, RSA_security_bits,
41 RSA_get0_pss_params, RSA_get_version, RSA_get0_engine,
42 RSA_generate_key_ex, RSA_generate_multi_prime_key,
43 RSA_X931_derive_ex, RSA_X931_generate_key_ex, RSA_check_key,
44 RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
45 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
46 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
47 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
48 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING,
49 RSA_blinding_on, RSA_blinding_off, RSA_setup_blinding,
50 RSA_padding_add_PKCS1_type_1, RSA_padding_check_PKCS1_type_1,
51 RSA_padding_add_PKCS1_type_2, RSA_padding_check_PKCS1_type_2,
52 PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP, RSA_padding_check_PKCS1_OAEP,
53 RSA_padding_add_PKCS1_OAEP_mgf1, RSA_padding_check_PKCS1_OAEP_mgf1,
54 RSA_padding_add_SSLv23, RSA_padding_check_SSLv23,
55 RSA_padding_add_none, RSA_padding_check_none, RSA_padding_add_X931,
56 RSA_padding_check_X931, RSA_X931_hash_id, RSA_verify_PKCS1_PSS,
57 RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
58 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
59 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
60 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
61 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
62 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
63 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
64 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
65 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
66 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
67 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
68 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
69 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
71 Use of these low level functions has been informally discouraged for a long
72 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
73 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
74 L<EVP_PKEY_decrypt(3)>.
78 * X509 certificates signed using SHA1 are no longer allowed at security
80 In TLS/SSL the default security level is 1. It can be set either
81 using the cipher string with @SECLEVEL, or calling
82 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
83 a call to SSL_CTX_use_certificate() will fail if the security level is not
85 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
86 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
91 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
92 deprecated. Instead use the pkeyparam, pkey, genpkey and pkeyparam
93 programs respectively.
97 * All of the low level DH functions have been deprecated including:
99 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
100 DH_new_method, DH_bits, DH_size, DH_security_bits, DH_get_ex_new_index,
101 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
102 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
103 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
104 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
105 DH_KDF_X9_42, DH_get0_engine, DH_get_length, DH_set_length, DH_meth_new,
106 DH_meth_free, DH_meth_dup, DH_meth_get0_name, DH_meth_set1_name,
107 DH_meth_get_flags, DH_meth_set_flags, DH_meth_get0_app_data,
108 DH_meth_set0_app_data, DH_meth_get_generate_key,
109 DH_meth_set_generate_key, DH_meth_get_compute_key,
110 DH_meth_set_compute_key, DH_meth_get_bn_mod_exp,
111 DH_meth_set_bn_mod_exp, DH_meth_get_init, DH_meth_set_init,
112 DH_meth_get_finish, DH_meth_set_finish, DH_meth_get_generate_params
113 and DH_meth_set_generate_params.
115 Use of these low level functions has been informally discouraged for a long
116 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
117 and L<EVP_PKEY_derive(3)>.
121 * All of the low level DSA functions have been deprecated including:
123 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
124 DSA_get_default_method, DSA_set_method, DSA_get_method, DSA_new_method,
125 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
126 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
127 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
128 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
129 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
130 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
131 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
132 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
133 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
134 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
135 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
137 Use of these low level functions has been informally discouraged for a long
138 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
139 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
143 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
144 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
145 This means that applications don't have to look at the curve NID and
146 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
147 However, they still can, that EVP_PKEY_set_alias_type() call acts as
148 a no-op when the EVP_PKEY is already of the given type.
150 Parameter and key generation is also reworked to make it possible
151 to generate EVP_PKEY_SM2 parameters and keys without having to go
152 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
153 However, code that does the latter will still work as before.
157 * Deprecated low level ECDH and ECDSA functions. These include:
159 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
160 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
163 Use of these low level functions has been informally discouraged for a long
164 time. Instead applications should use the EVP_PKEY_derive(3),
165 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
169 * Deprecated the EC_KEY_METHOD functions. These include:
171 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
172 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
173 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
174 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
175 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
176 EC_KEY_METHOD_get_verify.
178 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
182 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
183 and EVP_PKEY_decrypt() instead.
184 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
185 and EVP_PKEY_encrypt() instead.
189 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
190 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
191 a new formulation to include all the things it can be used for,
192 as well as words of caution.
196 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
197 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
201 * All of the low level HMAC functions have been deprecated including:
203 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
204 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
207 Use of these low level functions has been informally discouraged for a long
208 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
209 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
210 and L<EVP_MAC_final(3)>.
214 * Over two thousand fixes were made to the documentation, including:
215 - Common options (such as -rand/-writerand, TLS version control, etc)
216 were refactored and point to newly-enhanced descriptions in openssl.pod.
217 - Added style conformance for all options (with help from Richard Levitte),
218 documented all reported missing options, added a CI build to check
219 that all options are documented and that no unimplemented options
221 - Documented some internals, such as all use of environment variables.
222 - Addressed all internal broken L<> references.
226 * All of the low level CMAC functions have been deprecated including:
228 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
229 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
231 Use of these low level functions has been informally discouraged for a long
232 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
233 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
234 and L<EVP_MAC_final(3)>.
238 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
239 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
242 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
243 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
244 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
245 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
246 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
247 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
248 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
249 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
250 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
251 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
253 Use of these low level functions has been informally discouraged
254 for a long time. Applications should use the EVP_DigestInit_ex(3),
255 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
259 * Corrected the documentation of the return values from the EVP_DigestSign*
260 set of functions. The documentation mentioned negative values for some
261 errors, but this was never the case, so the mention of negative values
264 Code that followed the documentation and thereby check with something
265 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
269 * All of the low level cipher functions have been deprecated including:
271 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
272 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
273 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
274 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
275 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
276 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
277 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
278 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
279 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
280 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
281 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
282 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
283 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
284 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
285 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
286 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
287 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
288 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
289 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
290 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
291 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
292 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
293 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
294 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
295 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
296 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
297 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
298 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
299 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
301 Use of these low level functions has been informally discouraged for
302 a long time. Applications should use the high level EVP APIs, e.g.
303 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
304 equivalently named decrypt functions instead.
306 *Matt Caswell and Paul Dale*
308 * Removed include/openssl/opensslconf.h.in and replaced it with
309 include/openssl/configuration.h.in, which differs in not including
310 <openssl/macros.h>. A short header include/openssl/opensslconf.h
311 was added to include both.
313 This allows internal hacks where one might need to modify the set
314 of configured macros, for example this if deprecated symbols are
315 still supposed to be available internally:
317 #include <openssl/configuration.h>
319 #undef OPENSSL_NO_DEPRECATED
320 #define OPENSSL_SUPPRESS_DEPRECATED
322 #include <openssl/macros.h>
324 This should not be used by applications that use the exported
325 symbols, as that will lead to linking errors.
329 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
330 used in exponentiation with 512-bit moduli. No EC algorithms are
331 affected. Analysis suggests that attacks against 2-prime RSA1024,
332 3-prime RSA1536, and DSA1024 as a result of this defect would be very
333 difficult to perform and are not believed likely. Attacks against DH512
334 are considered just feasible. However, for an attack the target would
335 have to re-use the DH512 private key, which is not recommended anyway.
336 Also applications directly using the low level API BN_mod_exp may be
337 affected if they use BN_FLG_CONSTTIME.
342 * Most memory-debug features have been deprecated, and the functionality
343 replaced with no-ops.
347 * Introduced a new method type and API, OSSL_SERIALIZER, to
348 represent generic serializers. An implementation is expected to
349 be able to serialize an object associated with a given name (such
350 as an algorithm name for an asymmetric key) into forms given by
351 implementation properties.
353 Serializers are primarily used from inside libcrypto, through
354 calls to functions like EVP_PKEY_print_private(),
355 PEM_write_bio_PrivateKey() and similar.
357 Serializers are specified in such a way that they can be made to
358 directly handle the provider side portion of an object, if this
359 provider side part comes from the same provider as the serializer
360 itself, but can also be made to handle objects in parametrized
361 form (as an OSSL_PARAM array of data). This allows a provider to
362 offer generic serializers as a service for any other provider.
366 * Added a .pragma directive to the syntax of configuration files, to
367 allow varying behavior in a supported and predictable manner.
368 Currently added pragma:
372 This allows dollar signs to be a keyword character unless it's
373 followed by a opening brace or parenthesis. This is useful for
374 platforms where dollar signs are commonly used in names, such as
375 volume names and system directory names on VMS.
379 * Added functionality to create an EVP_PKEY from user data. This
380 is effectively the same as creating a RSA, DH or DSA object and
381 then assigning them to an EVP_PKEY, but directly using algorithm
382 agnostic EVP functions. A benefit is that this should be future
383 proof for public key algorithms to come.
387 * Change the interpretation of the '--api' configuration option to
388 mean that this is a desired API compatibility level with no
389 further meaning. The previous interpretation, that this would
390 also mean to remove all deprecated symbols up to and including
391 the given version, no requires that 'no-deprecated' is also used
392 in the configuration.
394 When building applications, the desired API compatibility level
395 can be set with the OPENSSL_API_COMPAT macro like before. For
396 API compatibility version below 3.0, the old style numerical
397 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
398 For version 3.0 and on, the value is expected to be the decimal
399 value calculated from the major and minor version like this:
401 MAJOR * 10000 + MINOR * 100
405 -DOPENSSL_API_COMPAT=30000 For 3.0
406 -DOPENSSL_API_COMPAT=30200 For 3.2
408 To hide declarations that are deprecated up to and including the
409 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
410 given when building the application as well.
414 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
415 access to certificate and CRL stores via URIs and OSSL_STORE
418 This adds the following functions:
420 - X509_LOOKUP_store()
421 - X509_STORE_load_file()
422 - X509_STORE_load_path()
423 - X509_STORE_load_store()
424 - SSL_add_store_cert_subjects_to_stack()
425 - SSL_CTX_set_default_verify_store()
426 - SSL_CTX_load_verify_file()
427 - SSL_CTX_load_verify_dir()
428 - SSL_CTX_load_verify_store()
430 Also, the following functions are now deprecated:
432 - X509_STORE_load_locations() (use X509_STORE_load_file(),
433 X509_STORE_load_path() or X509_STORE_load_store() instead)
434 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
435 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
439 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
440 The presence of this system service is determined at run-time.
444 * Added functionality to create an EVP_PKEY context based on data
445 for methods from providers. This takes an algorithm name and a
446 property query string and simply stores them, with the intent
447 that any operation that uses this context will use those strings
448 to fetch the needed methods implicitly, thereby making the port
449 of application written for pre-3.0 OpenSSL easier.
453 * The undocumented function NCONF_WIN32() has been deprecated; for
454 conversion details see the HISTORY section of doc/man5/config.pod
458 * Introduced the new functions EVP_DigestSignInit_ex() and
459 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
460 EVP_DigestVerifyUpdate() have been converted to functions. See the man
461 pages for further details.
465 * Over two thousand fixes were made to the documentation, including:
466 adding missing command flags, better style conformance, documentation
469 *Rich Salz, Richard Levitte*
471 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
472 X25519, X448, Ed25519 and Ed448.
476 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
481 * Deprecated the public definition of ERR_STATE as well as the function
482 ERR_get_state(). This is done in preparation of making ERR_STATE an
487 * Added ERR functionality to give callers access to the stored function
488 names that have replaced the older function code based functions.
490 New functions are ERR_get_error_func(), ERR_peek_error_func(),
491 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
492 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
493 and ERR_peek_last_error_all().
495 These functions have become deprecated: ERR_get_error_line_data(),
496 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
497 ERR_func_error_string().
501 * Extended testing to be verbose for failing tests only. The make variables
502 VERBOSE_FAILURE or VF can be used to enable this:
504 $ make VF=1 test # Unix
505 $ mms /macro=(VF=1) test ! OpenVMS
506 $ nmake VF=1 test # Windows
511 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
512 used even when parsing explicit parameters, when loading a serialized key
513 or calling `EC_GROUP_new_from_ecpkparameters()`/
514 `EC_GROUP_new_from_ecparameters()`.
515 This prevents bypass of security hardening and performance gains,
516 especially for curves with specialized EC_METHODs.
517 By default, if a key encoded with explicit parameters is loaded and later
518 serialized, the output is still encoded with explicit parameters, even if
519 internally a "named" EC_GROUP is used for computation.
523 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
524 this change, EC_GROUP_set_generator would accept order and/or cofactor as
525 NULL. After this change, only the cofactor parameter can be NULL. It also
526 does some minimal sanity checks on the passed order.
531 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
532 An attack is simple, if the first CMS_recipientInfo is valid but the
533 second CMS_recipientInfo is chosen ciphertext. If the second
534 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
535 encryption key will be replaced by garbage, and the message cannot be
536 decoded, but if the RSA decryption fails, the correct encryption key is
537 used and the recipient will not notice the attack.
538 As a work around for this potential attack the length of the decrypted
539 key must be equal to the cipher default key length, in case the
540 certifiate is not given and all recipientInfo are tried out.
541 The old behaviour can be re-enabled in the CMS code by setting the
542 CMS_DEBUG_DECRYPT flag.
546 * Early start up entropy quality from the DEVRANDOM seed source has been
547 improved for older Linux systems. The RAND subsystem will wait for
548 /dev/random to be producing output before seeding from /dev/urandom.
549 The seeded state is stored for future library initialisations using
550 a system global shared memory segment. The shared memory identifier
551 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
552 the desired value. The default identifier is 114.
556 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
557 when primes for RSA keys are computed.
558 Since we previously always generated primes == 2 (mod 3) for RSA keys,
559 the 2-prime and 3-prime RSA modules were easy to distinguish, since
560 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
561 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
562 This avoids possible fingerprinting of newly generated RSA modules.
566 * Correct the extended master secret constant on EBCDIC systems. Without this
567 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
568 negotiate EMS will fail. Unfortunately this also means that TLS connections
569 between EBCDIC systems with this fix, and EBCDIC systems without this
570 fix will fail if they negotiate EMS.
574 * Changed the library initialisation so that the config file is now loaded
575 by default. This was already the case for libssl. It now occurs for both
576 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
577 OPENSSL_init_crypto() to suppress automatic loading of a config file.
581 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
582 where the former acts as a replacement for ERR_put_error(), and the
583 latter replaces the combination ERR_put_error()+ERR_add_error_data().
584 ERR_raise_data() adds more flexibility by taking a format string and
585 an arbitrary number of arguments following it, to be processed with
590 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
591 to check if a named provider is loaded and available. When called, it
592 will also activate all fallback providers if such are still present.
596 * Enforce a minimum DH modulus size of 512 bits.
600 * Changed DH parameters to generate the order q subgroup instead of 2q.
601 Previously generated DH parameters are still accepted by DH_check
602 but DH_generate_key works around that by clearing bit 0 of the
603 private key for those. This avoids leaking bit 0 of the private key.
607 * Significantly reduce secure memory usage by the randomness pools.
611 * {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
616 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
617 algorithms. An implementation of a key exchange algorithm can be obtained
618 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
619 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
620 the older EVP_PKEY_derive_init() function. See the man pages for the new
621 functions for further details.
625 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
629 * Removed the function names from error messages and deprecated the
632 * Removed NextStep support and the macro OPENSSL_UNISTD
636 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
637 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
638 Also removed "export var as function" capability; we do not export
639 variables, only functions.
643 * RC5_32_set_key has been changed to return an int type, with 0 indicating
644 an error and 1 indicating success. In previous versions of OpenSSL this
645 was a void type. If a key was set longer than the maximum possible this
650 * Support SM2 signing and verification schemes with X509 certificate.
654 * Use SHA256 as the default digest for TS query in the ts app.
658 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
659 This checks that the salt length is at least 128 bits, the derived key
660 length is at least 112 bits, and that the iteration count is at least 1000.
661 For backwards compatibility these checks are disabled by default in the
662 default provider, but are enabled by default in the fips provider.
663 To enable or disable these checks use the control
664 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
668 * Default cipher lists/suites are now available via a function, the
669 #defines are deprecated.
673 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
674 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
675 for Windows Store apps easier. Also, the "no-uplink" option has been added.
679 * Join the directories crypto/x509 and crypto/x509v3
683 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
684 This changes the size when using the genpkey app when no size is given. It
685 fixes an omission in earlier changes that changed all RSA, DSA and DH
686 generation apps to use 2048 bits by default.
690 * Added command 'openssl kdf' that uses the EVP_KDF API.
694 * Added command 'openssl mac' that uses the EVP_MAC API.
698 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
699 as default directories. Also added the command 'openssl info'
700 for scripting purposes.
704 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
705 deprecated. These undocumented functions were never integrated into the EVP
706 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
707 Bi-directional IGE mode. These modes were never formally standardised and
708 usage of these functions is believed to be very small. In particular
709 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
710 is ever used. The security implications are believed to be minimal, but
711 this issue was never fixed for backwards compatibility reasons. New code
712 should not use these modes.
716 * Add prediction resistance to the DRBG reseeding process.
720 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
721 mandated by IEEE Std 1619-2018.
725 * Added newline escaping functionality to a filename when using openssl dgst.
726 This output format is to replicate the output format found in the '*sum'
727 checksum programs. This aims to preserve backward compatibility.
729 *Matt Eaton, Richard Levitte, and Paul Dale*
731 * Removed the heartbeat message in DTLS feature, as it has very
732 little usage and doesn't seem to fulfill a valuable purpose.
733 The configuration option is now deprecated.
737 * Changed the output of 'openssl {digestname} < file' to display the
738 digest name in its output.
742 * Added a new generic trace API which provides support for enabling
743 instrumentation through trace output. This feature is mainly intended
744 as an aid for developers and is disabled by default. To utilize it,
745 OpenSSL needs to be configured with the `enable-trace` option.
747 If the tracing API is enabled, the application can activate trace output
748 by registering BIOs as trace channels for a number of tracing and debugging
751 The 'openssl' application has been expanded to enable any of the types
752 available via environment variables defined by the user, and serves as
753 one possible example on how to use this functionality.
755 *Richard Levitte & Matthias St. Pierre*
757 * Added build tests for C++. These are generated files that only do one
758 thing, to include one public OpenSSL head file each. This tests that
759 the public header files can be usefully included in a C++ application.
761 This test isn't enabled by default. It can be enabled with the option
762 'enable-buildtest-c++'.
766 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
770 * Add KMAC to EVP_MAC.
774 * Added property based algorithm implementation selection framework to
779 * Added SCA hardening for modular field inversion in EC_GROUP through
780 a new dedicated field_inv() pointer in EC_METHOD.
781 This also addresses a leakage affecting conversions from projective
782 to affine coordinates.
784 *Billy Bob Brumley, Nicola Tuveri*
786 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
787 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
788 those algorithms that were already supported through the EVP_PKEY API
789 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
790 and scrypt are now wrappers that call EVP_KDF.
794 * Build devcrypto engine as a dynamic engine.
798 * Add keyed BLAKE2 to EVP_MAC.
802 * Fix a bug in the computation of the endpoint-pair shared secret used
803 by DTLS over SCTP. This breaks interoperability with older versions
804 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
805 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
806 interoperability with such broken implementations. However, enabling
807 this switch breaks interoperability with correct implementations.
809 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
810 re-used X509_PUBKEY object if the second PUBKEY is malformed.
814 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
818 * Change the license to the Apache License v2.0.
822 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
824 - Major releases (indicated by incrementing the MAJOR release number)
825 may introduce incompatible API/ABI changes.
826 - Minor releases (indicated by incrementing the MINOR release number)
827 may introduce new features but retain API/ABI compatibility.
828 - Patch releases (indicated by incrementing the PATCH number)
829 are intended for bug fixes and other improvements of existing
830 features only (like improving performance or adding documentation)
831 and retain API/ABI compatibility.
835 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
839 * Remove the 'dist' target and add a tarball building script. The
840 'dist' target has fallen out of use, and it shouldn't be
841 necessary to configure just to create a source distribution.
845 * Recreate the OS390-Unix config target. It no longer relies on a
846 special script like it did for OpenSSL pre-1.1.0.
850 * Instead of having the source directories listed in Configure, add
851 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
856 * Add GMAC to EVP_MAC.
860 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
864 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
865 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
866 to facilitate the continued use of MACs through raw private keys in
867 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
871 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
872 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
876 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
877 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
878 are retained for backwards compatibility.
882 * AES-XTS mode now enforces that its two keys are different to mitigate
883 the attacked described in "Efficient Instantiations of Tweakable
884 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
885 Details of this attack can be obtained from:
886 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
890 * Rename the object files, i.e. give them other names than in previous
891 versions. Their names now include the name of the final product, as
892 well as its type mnemonic (bin, lib, shlib).
896 * Added new option for 'openssl list', '-objects', which will display the
897 list of built in objects, i.e. OIDs with names.
901 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
902 improves application performance by removing data copies and providing
903 applications with zero-copy system calls such as sendfile and splice.
911 ### Changes between 1.1.1d and 1.1.1e [xx XXX xxxx] ###
913 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
914 The presence of this system service is determined at run-time.
918 * Added newline escaping functionality to a filename when using openssl dgst.
919 This output format is to replicate the output format found in the '*sum'
920 checksum programs. This aims to preserve backward compatibility.
922 *Matt Eaton, Richard Levitte, and Paul Dale*
924 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
929 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019] ###
931 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
932 number generator (RNG). This was intended to include protection in the
933 event of a fork() system call in order to ensure that the parent and child
934 processes did not share the same RNG state. However this protection was not
935 being used in the default case.
937 A partial mitigation for this issue is that the output from a high
938 precision timer is mixed into the RNG state so the likelihood of a parent
939 and child process sharing state is significantly reduced.
941 If an application already calls OPENSSL_init_crypto() explicitly using
942 OPENSSL_INIT_ATFORK then this problem does not occur at all.
945 *Matthias St. Pierre*
947 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
948 used even when parsing explicit parameters, when loading a serialized key
949 or calling `EC_GROUP_new_from_ecpkparameters()`/
950 `EC_GROUP_new_from_ecparameters()`.
951 This prevents bypass of security hardening and performance gains,
952 especially for curves with specialized EC_METHODs.
953 By default, if a key encoded with explicit parameters is loaded and later
954 serialized, the output is still encoded with explicit parameters, even if
955 internally a "named" EC_GROUP is used for computation.
959 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
960 this change, EC_GROUP_set_generator would accept order and/or cofactor as
961 NULL. After this change, only the cofactor parameter can be NULL. It also
962 does some minimal sanity checks on the passed order.
967 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
968 An attack is simple, if the first CMS_recipientInfo is valid but the
969 second CMS_recipientInfo is chosen ciphertext. If the second
970 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
971 encryption key will be replaced by garbage, and the message cannot be
972 decoded, but if the RSA decryption fails, the correct encryption key is
973 used and the recipient will not notice the attack.
974 As a work around for this potential attack the length of the decrypted
975 key must be equal to the cipher default key length, in case the
976 certifiate is not given and all recipientInfo are tried out.
977 The old behaviour can be re-enabled in the CMS code by setting the
978 CMS_DEBUG_DECRYPT flag.
983 * Early start up entropy quality from the DEVRANDOM seed source has been
984 improved for older Linux systems. The RAND subsystem will wait for
985 /dev/random to be producing output before seeding from /dev/urandom.
986 The seeded state is stored for future library initialisations using
987 a system global shared memory segment. The shared memory identifier
988 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
989 the desired value. The default identifier is 114.
993 * Correct the extended master secret constant on EBCDIC systems. Without this
994 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
995 negotiate EMS will fail. Unfortunately this also means that TLS connections
996 between EBCDIC systems with this fix, and EBCDIC systems without this
997 fix will fail if they negotiate EMS.
1001 * Use Windows installation paths in the mingw builds
1003 Mingw isn't a POSIX environment per se, which means that Windows
1004 paths should be used for installation.
1009 * Changed DH_check to accept parameters with order q and 2q subgroups.
1010 With order 2q subgroups the bit 0 of the private key is not secret
1011 but DH_generate_key works around that by clearing bit 0 of the
1012 private key for those. This avoids leaking bit 0 of the private key.
1016 * Significantly reduce secure memory usage by the randomness pools.
1020 * Revert the DEVRANDOM_WAIT feature for Linux systems
1022 The DEVRANDOM_WAIT feature added a select() call to wait for the
1023 /dev/random device to become readable before reading from the
1024 /dev/urandom device.
1026 It turned out that this change had negative side effects on
1027 performance which were not acceptable. After some discussion it
1028 was decided to revert this feature and leave it up to the OS
1029 resp. the platform maintainer to ensure a proper initialization
1030 during early boot time.
1032 *Matthias St. Pierre*
1034 ### Changes between 1.1.1b and 1.1.1c [28 May 2019] ###
1036 * Add build tests for C++. These are generated files that only do one
1037 thing, to include one public OpenSSL head file each. This tests that
1038 the public header files can be usefully included in a C++ application.
1040 This test isn't enabled by default. It can be enabled with the option
1041 'enable-buildtest-c++'.
1045 * Enable SHA3 pre-hashing for ECDSA and DSA.
1049 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1050 This changes the size when using the genpkey app when no size is given. It
1051 fixes an omission in earlier changes that changed all RSA, DSA and DH
1052 generation apps to use 2048 bits by default.
1056 * Reorganize the manual pages to consistently have RETURN VALUES,
1057 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1058 util/fix-doc-nits accordingly.
1060 *Paul Yang, Joshua Lock*
1062 * Add the missing accessor EVP_PKEY_get0_engine()
1066 * Have apps like 's_client' and 's_server' output the signature scheme
1067 along with other cipher suite parameters when debugging.
1071 * Make OPENSSL_config() error agnostic again.
1075 * Do the error handling in RSA decryption constant time.
1079 * Prevent over long nonces in ChaCha20-Poly1305.
1081 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1082 for every encryption operation. RFC 7539 specifies that the nonce value
1083 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1084 and front pads the nonce with 0 bytes if it is less than 12
1085 bytes. However it also incorrectly allows a nonce to be set of up to 16
1086 bytes. In this case only the last 12 bytes are significant and any
1087 additional leading bytes are ignored.
1089 It is a requirement of using this cipher that nonce values are
1090 unique. Messages encrypted using a reused nonce value are susceptible to
1091 serious confidentiality and integrity attacks. If an application changes
1092 the default nonce length to be longer than 12 bytes and then makes a
1093 change to the leading bytes of the nonce expecting the new value to be a
1094 new unique nonce then such an application could inadvertently encrypt
1095 messages with a reused nonce.
1097 Additionally the ignored bytes in a long nonce are not covered by the
1098 integrity guarantee of this cipher. Any application that relies on the
1099 integrity of these ignored leading bytes of a long nonce may be further
1100 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1101 is safe because no such use sets such a long nonce value. However user
1102 applications that use this cipher directly and set a non-default nonce
1103 length to be longer than 12 bytes may be vulnerable.
1105 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1111 * Add DEVRANDOM_WAIT feature for Linux systems
1113 On older Linux systems where the getrandom() system call is not available,
1114 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1115 Contrary to getrandom(), the /dev/urandom device will not block during
1116 early boot when the kernel CSPRNG has not been seeded yet.
1118 To mitigate this known weakness, use select() to wait for /dev/random to
1119 become readable before reading from /dev/urandom.
1121 * Ensure that SM2 only uses SM3 as digest algorithm
1125 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019] ###
1127 * Change the info callback signals for the start and end of a post-handshake
1128 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1129 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1130 confused by this and assume that a TLSv1.2 renegotiation has started. This
1131 can break KeyUpdate handling. Instead we no longer signal the start and end
1132 of a post handshake message exchange (although the messages themselves are
1133 still signalled). This could break some applications that were expecting
1134 the old signals. However without this KeyUpdate is not usable for many
1139 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018] ###
1141 * Timing vulnerability in DSA signature generation
1143 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1144 timing side channel attack. An attacker could use variations in the signing
1145 algorithm to recover the private key.
1147 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1152 * Timing vulnerability in ECDSA signature generation
1154 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1155 timing side channel attack. An attacker could use variations in the signing
1156 algorithm to recover the private key.
1158 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1163 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1164 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1165 of two gigabytes and the error handling improved.
1167 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1168 categorized as a normal bug, not a security issue, because the DRBG reseeds
1169 automatically and is fully functional even without additional randomness
1170 provided by the application.
1172 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018] ###
1174 * Add a new ClientHello callback. Provides a callback interface that gives
1175 the application the ability to adjust the nascent SSL object at the
1176 earliest stage of ClientHello processing, immediately after extensions have
1177 been collected but before they have been processed. In particular, this
1178 callback can adjust the supported TLS versions in response to the contents
1183 * Add SM2 base algorithm support.
1187 * s390x assembly pack: add (improved) hardware-support for the following
1188 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1189 aes-cfb/cfb8, aes-ecb.
1193 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1194 parameter is no longer accepted, as it leads to a corrupt table. NULL
1195 pem_str is reserved for alias entries only.
1199 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1200 step for prime curves. The new implementation is based on formulae from
1201 differential addition-and-doubling in homogeneous projective coordinates
1202 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1203 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1204 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1205 to work in projective coordinates.
1207 *Billy Bob Brumley, Nicola Tuveri*
1209 * Change generating and checking of primes so that the error rate of not
1210 being prime depends on the intended use based on the size of the input.
1211 For larger primes this will result in more rounds of Miller-Rabin.
1212 The maximal error rate for primes with more than 1080 bits is lowered
1215 *Kurt Roeckx, Annie Yousar*
1217 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1221 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1222 moving between systems, and to avoid confusion when a Windows build is
1223 done with mingw vs with MSVC. For POSIX installs, there's still a
1224 symlink or copy named 'tsget' to avoid that confusion as well.
1228 * Revert blinding in ECDSA sign and instead make problematic addition
1229 length-invariant. Switch even to fixed-length Montgomery multiplication.
1233 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1234 step for binary curves. The new implementation is based on formulae from
1235 differential addition-and-doubling in mixed Lopez-Dahab projective
1236 coordinates, modified to independently blind the operands.
1238 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1240 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1241 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1242 EC_METHODs to implement their own specialized "ladder step", to take
1243 advantage of more favorable coordinate systems or more efficient
1244 differential addition-and-doubling algorithms.
1246 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1248 * Modified the random device based seed sources to keep the relevant
1249 file descriptors open rather than reopening them on each access.
1250 This allows such sources to operate in a chroot() jail without
1251 the associated device nodes being available. This behaviour can be
1252 controlled using RAND_keep_random_devices_open().
1256 * Numerous side-channel attack mitigations have been applied. This may have
1257 performance impacts for some algorithms for the benefit of improved
1258 security. Specific changes are noted in this change log by their respective
1263 * AIX shared library support overhaul. Switch to AIX "natural" way of
1264 handling shared libraries, which means collecting shared objects of
1265 different versions and bitnesses in one common archive. This allows to
1266 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1267 doesn't affect the way 3rd party applications are linked, only how
1268 multi-version installation is managed.
1272 * Make ec_group_do_inverse_ord() more robust and available to other
1273 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1274 mitigations are applied to the fallback BN_mod_inverse().
1275 When using this function rather than BN_mod_inverse() directly, new
1276 EC cryptosystem implementations are then safer-by-default.
1280 * Add coordinate blinding for EC_POINT and implement projective
1281 coordinate blinding for generic prime curves as a countermeasure to
1282 chosen point SCA attacks.
1284 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1286 * Add blinding to ECDSA and DSA signatures to protect against side channel
1287 attacks discovered by Keegan Ryan (NCC Group).
1291 * Enforce checking in the pkeyutl command line app to ensure that the input
1292 length does not exceed the maximum supported digest length when performing
1293 a sign, verify or verifyrecover operation.
1297 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1298 I/O in combination with something like select() or poll() will hang. This
1299 can be turned off again using SSL_CTX_clear_mode().
1300 Many applications do not properly handle non-application data records, and
1301 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1302 around the problems in those applications, but can also break some.
1303 It's recommended to read the manpages about SSL_read(), SSL_write(),
1304 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1305 SSL_CTX_set_read_ahead() again.
1309 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1310 now allow empty (zero character) pass phrases.
1314 * Apply blinding to binary field modular inversion and remove patent
1315 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1319 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1320 binary and prime elliptic curves.
1324 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1325 constant time fixed point multiplication.
1329 * Revise elliptic curve scalar multiplication with timing attack
1330 defenses: ec_wNAF_mul redirects to a constant time implementation
1331 when computing fixed point and variable point multiplication (which
1332 in OpenSSL are mostly used with secret scalars in keygen, sign,
1333 ECDH derive operations).
1334 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1337 * Updated CONTRIBUTING
1341 * Updated DRBG / RAND to request nonce and additional low entropy
1342 randomness from the system.
1344 *Matthias St. Pierre*
1346 * Updated 'openssl rehash' to use OpenSSL consistent default.
1350 * Moved the load of the ssl_conf module to libcrypto, which helps
1351 loading engines that libssl uses before libssl is initialised.
1355 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1359 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1361 *Ingo Schwarze, Rich Salz*
1363 * Added output of accepting IP address and port for 'openssl s_server'
1367 * Added a new API for TLSv1.3 ciphersuites:
1368 SSL_CTX_set_ciphersuites()
1369 SSL_set_ciphersuites()
1373 * Memory allocation failures consistently add an error to the error
1378 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1379 in libcrypto when run as setuid/setgid.
1383 * Load any config file by default when libssl is used.
1387 * Added new public header file <openssl/rand_drbg.h> and documentation
1388 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1390 *Matthias St. Pierre*
1392 * QNX support removed (cannot find contributors to get their approval
1393 for the license change).
1397 * TLSv1.3 replay protection for early data has been implemented. See the
1398 SSL_read_early_data() man page for further details.
1402 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1403 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1404 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1405 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1406 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1407 configuration has been separated out. See the ciphers man page or the
1408 SSL_CTX_set_ciphersuites() man page for more information.
1412 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1413 in responder mode now supports the new "-multi" option, which
1414 spawns the specified number of child processes to handle OCSP
1415 requests. The "-timeout" option now also limits the OCSP
1416 responder's patience to wait to receive the full client request
1417 on a newly accepted connection. Child processes are respawned
1418 as needed, and the CA index file is automatically reloaded
1419 when changed. This makes it possible to run the "ocsp" responder
1420 as a long-running service, making the OpenSSL CA somewhat more
1421 feature-complete. In this mode, most diagnostic messages logged
1422 after entering the event loop are logged via syslog(3) rather than
1427 * Added support for X448 and Ed448. Heavily based on original work by
1432 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1433 objects loaded. This adds the functions OSSL_STORE_expect() and
1434 OSSL_STORE_find() as well as needed tools to construct searches and
1435 get the search data out of them.
1439 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1440 version of OpenSSL should review their configuration settings to ensure
1441 that they are still appropriate for TLSv1.3. For further information see:
1442 https://wiki.openssl.org/index.php/TLS1.3
1446 * Grand redesign of the OpenSSL random generator
1448 The default RAND method now utilizes an AES-CTR DRBG according to
1449 NIST standard SP 800-90Ar1. The new random generator is essentially
1450 a port of the default random generator from the OpenSSL FIPS 2.0
1451 object module. It is a hybrid deterministic random bit generator
1452 using an AES-CTR bit stream and which seeds and reseeds itself
1453 automatically using trusted system entropy sources.
1455 Some of its new features are:
1456 - Support for multiple DRBG instances with seed chaining.
1457 - The default RAND method makes use of a DRBG.
1458 - There is a public and private DRBG instance.
1459 - The DRBG instances are fork-safe.
1460 - Keep all global DRBG instances on the secure heap if it is enabled.
1461 - The public and private DRBG instance are per thread for lock free
1464 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1466 * Changed Configure so it only says what it does and doesn't dump
1467 so much data. Instead, ./configdata.pm should be used as a script
1468 to display all sorts of configuration data.
1472 * Added processing of "make variables" to Configure.
1476 * Added SHA512/224 and SHA512/256 algorithm support.
1480 * The last traces of Netware support, first removed in 1.1.0, have
1485 * Get rid of Makefile.shared, and in the process, make the processing
1486 of certain files (rc.obj, or the .def/.map/.opt files produced from
1487 the ordinal files) more visible and hopefully easier to trace and
1488 debug (or make silent).
1492 * Make it possible to have environment variable assignments as
1493 arguments to config / Configure.
1497 * Add multi-prime RSA (RFC 8017) support.
1501 * Add SM3 implemented according to GB/T 32905-2016
1502 * Jack Lloyd <jack.lloyd@ribose.com>,
1503 Ronald Tse <ronald.tse@ribose.com>,
1504 Erick Borsboom <erick.borsboom@ribose.com> *
1506 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1507 as documented in RFC6066.
1508 Based on a patch from Tomasz Moń
1510 *Filipe Raimundo da Silva*
1512 * Add SM4 implemented according to GB/T 32907-2016.
1513 * Jack Lloyd <jack.lloyd@ribose.com>,
1514 Ronald Tse <ronald.tse@ribose.com>,
1515 Erick Borsboom <erick.borsboom@ribose.com> *
1517 * Reimplement -newreq-nodes and ERR_error_string_n; the
1518 original author does not agree with the license change.
1522 * Add ARIA AEAD TLS support.
1526 * Some macro definitions to support VS6 have been removed. Visual
1527 Studio 6 has not worked since 1.1.0
1531 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1532 without clearing the errors.
1536 * Add "atfork" functions. If building on a system that without
1537 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1538 requirements. The RAND facility now uses/requires this.
1546 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1547 not possible to disable entirely. However, it's still possible to
1548 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1551 To disable, configure with 'no-ui-console'. 'no-ui' is still
1552 possible to use as an alias. Check at compile time with the
1553 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1554 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1558 * Add a STORE module, which implements a uniform and URI based reader of
1559 stores that can contain keys, certificates, CRLs and numerous other
1560 objects. The main API is loosely based on a few stdio functions,
1561 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1562 OSSL_STORE_error and OSSL_STORE_close.
1563 The implementation uses backends called "loaders" to implement arbitrary
1564 URI schemes. There is one built in "loader" for the 'file' scheme.
1568 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1569 then adjusted to work on FreeBSD 8.4 as well.
1570 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1571 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1575 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1576 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1577 error code calls like this:
1579 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1581 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1582 that can be encoded in C. For the foreseeable future, this will only
1585 *Richard Levitte and Tim Hudson*
1587 * Removed BSD cryptodev engine.
1591 * Add a build target 'build_all_generated', to build all generated files
1592 and only that. This can be used to prepare everything that requires
1593 things like perl for a system that lacks perl and then move everything
1594 to that system and do the rest of the build there.
1598 * In the UI interface, make it possible to duplicate the user data. This
1599 can be used by engines that need to retain the data for a longer time
1600 than just the call where this user data is passed.
1604 * Ignore the '-named_curve auto' value for compatibility of applications
1607 *Tomas Mraz <tmraz@fedoraproject.org>*
1609 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1610 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1611 alerts across multiple records (some of which could be empty). In practice
1612 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1613 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1614 support this at all. Supporting it adds significant complexity to the
1615 record layer, and its removal is unlikely to cause interoperability
1620 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1621 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1622 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1627 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1628 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1630 *Richard Levitte, Andy Polyakov*
1632 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1637 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1638 platform rather than 'mingw'.
1642 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1643 success if they are asked to add an object which already exists
1644 in the store. This change cascades to other functions which load
1645 certificates and CRLs.
1649 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1650 facilitate stack unwinding even from assembly subroutines.
1654 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1655 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1659 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1660 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1661 which is the minimum version we support.
1665 * Certificate time validation (X509_cmp_time) enforces stricter
1666 compliance with RFC 5280. Fractional seconds and timezone offsets
1667 are no longer allowed.
1671 * Add support for ARIA
1675 * s_client will now send the Server Name Indication (SNI) extension by
1676 default unless the new "-noservername" option is used. The server name is
1677 based on the host provided to the "-connect" option unless overridden by
1678 using "-servername".
1682 * Add support for SipHash
1686 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1687 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1688 prevent issues where no progress is being made and the peer continually
1689 sends unrecognised record types, using up resources processing them.
1693 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1694 using the algorithm defined in
1695 https://www.akkadia.org/drepper/SHA-crypt.txt
1699 * Heartbeat support has been removed; the ABI is changed for now.
1701 *Richard Levitte, Rich Salz*
1703 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1707 * The RSA "null" method, which was partially supported to avoid patent
1708 issues, has been replaced to always returns NULL.
1716 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019] ###
1718 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1719 used even when parsing explicit parameters, when loading a serialized key
1720 or calling `EC_GROUP_new_from_ecpkparameters()`/
1721 `EC_GROUP_new_from_ecparameters()`.
1722 This prevents bypass of security hardening and performance gains,
1723 especially for curves with specialized EC_METHODs.
1724 By default, if a key encoded with explicit parameters is loaded and later
1725 serialized, the output is still encoded with explicit parameters, even if
1726 internally a "named" EC_GROUP is used for computation.
1730 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1731 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1732 NULL. After this change, only the cofactor parameter can be NULL. It also
1733 does some minimal sanity checks on the passed order.
1738 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1739 An attack is simple, if the first CMS_recipientInfo is valid but the
1740 second CMS_recipientInfo is chosen ciphertext. If the second
1741 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1742 encryption key will be replaced by garbage, and the message cannot be
1743 decoded, but if the RSA decryption fails, the correct encryption key is
1744 used and the recipient will not notice the attack.
1745 As a work around for this potential attack the length of the decrypted
1746 key must be equal to the cipher default key length, in case the
1747 certifiate is not given and all recipientInfo are tried out.
1748 The old behaviour can be re-enabled in the CMS code by setting the
1749 CMS_DEBUG_DECRYPT flag.
1754 * Use Windows installation paths in the mingw builds
1756 Mingw isn't a POSIX environment per se, which means that Windows
1757 paths should be used for installation.
1762 ### Changes between 1.1.0j and 1.1.0k [28 May 2019] ###
1764 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1765 This changes the size when using the genpkey app when no size is given. It
1766 fixes an omission in earlier changes that changed all RSA, DSA and DH
1767 generation apps to use 2048 bits by default.
1771 * Prevent over long nonces in ChaCha20-Poly1305.
1773 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1774 for every encryption operation. RFC 7539 specifies that the nonce value
1775 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1776 and front pads the nonce with 0 bytes if it is less than 12
1777 bytes. However it also incorrectly allows a nonce to be set of up to 16
1778 bytes. In this case only the last 12 bytes are significant and any
1779 additional leading bytes are ignored.
1781 It is a requirement of using this cipher that nonce values are
1782 unique. Messages encrypted using a reused nonce value are susceptible to
1783 serious confidentiality and integrity attacks. If an application changes
1784 the default nonce length to be longer than 12 bytes and then makes a
1785 change to the leading bytes of the nonce expecting the new value to be a
1786 new unique nonce then such an application could inadvertently encrypt
1787 messages with a reused nonce.
1789 Additionally the ignored bytes in a long nonce are not covered by the
1790 integrity guarantee of this cipher. Any application that relies on the
1791 integrity of these ignored leading bytes of a long nonce may be further
1792 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1793 is safe because no such use sets such a long nonce value. However user
1794 applications that use this cipher directly and set a non-default nonce
1795 length to be longer than 12 bytes may be vulnerable.
1797 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1803 * Added SCA hardening for modular field inversion in EC_GROUP through
1804 a new dedicated field_inv() pointer in EC_METHOD.
1805 This also addresses a leakage affecting conversions from projective
1806 to affine coordinates.
1808 *Billy Bob Brumley, Nicola Tuveri*
1810 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1811 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1815 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1819 * Remove the 'dist' target and add a tarball building script. The
1820 'dist' target has fallen out of use, and it shouldn't be
1821 necessary to configure just to create a source distribution.
1825 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018] ###
1827 * Timing vulnerability in DSA signature generation
1829 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1830 timing side channel attack. An attacker could use variations in the signing
1831 algorithm to recover the private key.
1833 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1838 * Timing vulnerability in ECDSA signature generation
1840 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1841 timing side channel attack. An attacker could use variations in the signing
1842 algorithm to recover the private key.
1844 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1849 * Add coordinate blinding for EC_POINT and implement projective
1850 coordinate blinding for generic prime curves as a countermeasure to
1851 chosen point SCA attacks.
1853 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1855 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018] ###
1857 * Client DoS due to large DH parameter
1859 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1860 malicious server can send a very large prime value to the client. This will
1861 cause the client to spend an unreasonably long period of time generating a
1862 key for this prime resulting in a hang until the client has finished. This
1863 could be exploited in a Denial Of Service attack.
1865 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1870 * Cache timing vulnerability in RSA Key Generation
1872 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1873 a cache timing side channel attack. An attacker with sufficient access to
1874 mount cache timing attacks during the RSA key generation process could
1875 recover the private key.
1877 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1878 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1883 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1884 parameter is no longer accepted, as it leads to a corrupt table. NULL
1885 pem_str is reserved for alias entries only.
1889 * Revert blinding in ECDSA sign and instead make problematic addition
1890 length-invariant. Switch even to fixed-length Montgomery multiplication.
1894 * Change generating and checking of primes so that the error rate of not
1895 being prime depends on the intended use based on the size of the input.
1896 For larger primes this will result in more rounds of Miller-Rabin.
1897 The maximal error rate for primes with more than 1080 bits is lowered
1900 *Kurt Roeckx, Annie Yousar*
1902 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1906 * Add blinding to ECDSA and DSA signatures to protect against side channel
1907 attacks discovered by Keegan Ryan (NCC Group).
1911 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1912 now allow empty (zero character) pass phrases.
1916 * Certificate time validation (X509_cmp_time) enforces stricter
1917 compliance with RFC 5280. Fractional seconds and timezone offsets
1918 are no longer allowed.
1922 * Fixed a text canonicalisation bug in CMS
1924 Where a CMS detached signature is used with text content the text goes
1925 through a canonicalisation process first prior to signing or verifying a
1926 signature. This process strips trailing space at the end of lines, converts
1927 line terminators to CRLF and removes additional trailing line terminators
1928 at the end of a file. A bug in the canonicalisation process meant that
1929 some characters, such as form-feed, were incorrectly treated as whitespace
1930 and removed. This is contrary to the specification (RFC5485). This fix
1931 could mean that detached text data signed with an earlier version of
1932 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1933 signed with a fixed OpenSSL may fail to verify with an earlier version of
1934 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1935 and use the "-binary" flag (for the "cms" command line application) or set
1936 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1940 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018] ###
1942 * Constructed ASN.1 types with a recursive definition could exceed the stack
1944 Constructed ASN.1 types with a recursive definition (such as can be found
1945 in PKCS7) could eventually exceed the stack given malicious input with
1946 excessive recursion. This could result in a Denial Of Service attack. There
1947 are no such structures used within SSL/TLS that come from untrusted sources
1948 so this is considered safe.
1950 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1956 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1958 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1959 effectively reduced to only comparing the least significant bit of each
1960 byte. This allows an attacker to forge messages that would be considered as
1961 authenticated in an amount of tries lower than that guaranteed by the
1962 security claims of the scheme. The module can only be compiled by the
1963 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1965 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1971 * Add a build target 'build_all_generated', to build all generated files
1972 and only that. This can be used to prepare everything that requires
1973 things like perl for a system that lacks perl and then move everything
1974 to that system and do the rest of the build there.
1978 * Backport SSL_OP_NO_RENGOTIATION
1980 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1981 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1982 changes this is no longer possible in 1.1.0. Therefore the new
1983 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1984 1.1.0 to provide equivalent functionality.
1986 Note that if an application built against 1.1.0h headers (or above) is run
1987 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1988 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1992 * Removed the OS390-Unix config target. It relied on a script that doesn't
1997 * rsaz_1024_mul_avx2 overflow bug on x86_64
1999 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2000 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2001 Analysis suggests that attacks against RSA and DSA as a result of this
2002 defect would be very difficult to perform and are not believed likely.
2003 Attacks against DH1024 are considered just feasible, because most of the
2004 work necessary to deduce information about a private key may be performed
2005 offline. The amount of resources required for such an attack would be
2006 significant. However, for an attack on TLS to be meaningful, the server
2007 would have to share the DH1024 private key among multiple clients, which is
2008 no longer an option since CVE-2016-0701.
2010 This only affects processors that support the AVX2 but not ADX extensions
2011 like Intel Haswell (4th generation).
2013 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2014 was originally found via the OSS-Fuzz project.
2019 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017] ###
2021 * bn_sqrx8x_internal carry bug on x86_64
2023 There is a carry propagating bug in the x86_64 Montgomery squaring
2024 procedure. No EC algorithms are affected. Analysis suggests that attacks
2025 against RSA and DSA as a result of this defect would be very difficult to
2026 perform and are not believed likely. Attacks against DH are considered just
2027 feasible (although very difficult) because most of the work necessary to
2028 deduce information about a private key may be performed offline. The amount
2029 of resources required for such an attack would be very significant and
2030 likely only accessible to a limited number of attackers. An attacker would
2031 additionally need online access to an unpatched system using the target
2032 private key in a scenario with persistent DH parameters and a private
2033 key that is shared between multiple clients.
2035 This only affects processors that support the BMI1, BMI2 and ADX extensions
2036 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2038 This issue was reported to OpenSSL by the OSS-Fuzz project.
2043 * Malformed X.509 IPAddressFamily could cause OOB read
2045 If an X.509 certificate has a malformed IPAddressFamily extension,
2046 OpenSSL could do a one-byte buffer overread. The most likely result
2047 would be an erroneous display of the certificate in text format.
2049 This issue was reported to OpenSSL by the OSS-Fuzz project.
2054 ### Changes between 1.1.0e and 1.1.0f [25 May 2017] ###
2056 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2057 platform rather than 'mingw'.
2061 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2062 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2063 which is the minimum version we support.
2067 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017] ###
2069 * Encrypt-Then-Mac renegotiation crash
2071 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2072 negotiated where it was not in the original handshake (or vice-versa) then
2073 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2074 and servers are affected.
2076 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2081 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017] ###
2083 * Truncated packet could crash via OOB read
2085 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2086 cipher is being used, then a truncated packet can cause that host to
2087 perform an out-of-bounds read, usually resulting in a crash.
2089 This issue was reported to OpenSSL by Robert Święcki of Google.
2094 * Bad (EC)DHE parameters cause a client crash
2096 If a malicious server supplies bad parameters for a DHE or ECDHE key
2097 exchange then this can result in the client attempting to dereference a
2098 NULL pointer leading to a client crash. This could be exploited in a Denial
2101 This issue was reported to OpenSSL by Guido Vranken.
2106 * BN_mod_exp may produce incorrect results on x86_64
2108 There is a carry propagating bug in the x86_64 Montgomery squaring
2109 procedure. No EC algorithms are affected. Analysis suggests that attacks
2110 against RSA and DSA as a result of this defect would be very difficult to
2111 perform and are not believed likely. Attacks against DH are considered just
2112 feasible (although very difficult) because most of the work necessary to
2113 deduce information about a private key may be performed offline. The amount
2114 of resources required for such an attack would be very significant and
2115 likely only accessible to a limited number of attackers. An attacker would
2116 additionally need online access to an unpatched system using the target
2117 private key in a scenario with persistent DH parameters and a private
2118 key that is shared between multiple clients. For example this can occur by
2119 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2120 similar to CVE-2015-3193 but must be treated as a separate problem.
2122 This issue was reported to OpenSSL by the OSS-Fuzz project.
2127 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016] ###
2129 * ChaCha20/Poly1305 heap-buffer-overflow
2131 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
2132 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2133 crash. This issue is not considered to be exploitable beyond a DoS.
2135 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2140 * CMS Null dereference
2142 Applications parsing invalid CMS structures can crash with a NULL pointer
2143 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2144 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2145 structure callback if an attempt is made to free certain invalid encodings.
2146 Only CHOICE structures using a callback which do not handle NULL value are
2149 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2154 * Montgomery multiplication may produce incorrect results
2156 There is a carry propagating bug in the Broadwell-specific Montgomery
2157 multiplication procedure that handles input lengths divisible by, but
2158 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2159 and DH private keys are impossible. This is because the subroutine in
2160 question is not used in operations with the private key itself and an input
2161 of the attacker's direct choice. Otherwise the bug can manifest itself as
2162 transient authentication and key negotiation failures or reproducible
2163 erroneous outcome of public-key operations with specially crafted input.
2164 Among EC algorithms only Brainpool P-512 curves are affected and one
2165 presumably can attack ECDH key negotiation. Impact was not analyzed in
2166 detail, because pre-requisites for attack are considered unlikely. Namely
2167 multiple clients have to choose the curve in question and the server has to
2168 share the private key among them, neither of which is default behaviour.
2169 Even then only clients that chose the curve will be affected.
2171 This issue was publicly reported as transient failures and was not
2172 initially recognized as a security issue. Thanks to Richard Morgan for
2173 providing reproducible case.
2178 * Removed automatic addition of RPATH in shared libraries and executables,
2179 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2183 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016] ###
2185 * Fix Use After Free for large message sizes
2187 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2188 message larger than approx 16k is received then the underlying buffer to
2189 store the incoming message is reallocated and moved. Unfortunately a
2190 dangling pointer to the old location is left which results in an attempt to
2191 write to the previously freed location. This is likely to result in a
2192 crash, however it could potentially lead to execution of arbitrary code.
2194 This issue only affects OpenSSL 1.1.0a.
2196 This issue was reported to OpenSSL by Robert Święcki.
2201 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016] ###
2203 * OCSP Status Request extension unbounded memory growth
2205 A malicious client can send an excessively large OCSP Status Request
2206 extension. If that client continually requests renegotiation, sending a
2207 large OCSP Status Request extension each time, then there will be unbounded
2208 memory growth on the server. This will eventually lead to a Denial Of
2209 Service attack through memory exhaustion. Servers with a default
2210 configuration are vulnerable even if they do not support OCSP. Builds using
2211 the "no-ocsp" build time option are not affected.
2213 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2218 * SSL_peek() hang on empty record
2220 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2221 sends an empty record. This could be exploited by a malicious peer in a
2222 Denial Of Service attack.
2224 This issue was reported to OpenSSL by Alex Gaynor.
2229 * Excessive allocation of memory in tls_get_message_header() and
2230 dtls1_preprocess_fragment()
2232 A (D)TLS message includes 3 bytes for its length in the header for the
2233 message. This would allow for messages up to 16Mb in length. Messages of
2234 this length are excessive and OpenSSL includes a check to ensure that a
2235 peer is sending reasonably sized messages in order to avoid too much memory
2236 being consumed to service a connection. A flaw in the logic of version
2237 1.1.0 means that memory for the message is allocated too early, prior to
2238 the excessive message length check. Due to way memory is allocated in
2239 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2240 to service a connection. This could lead to a Denial of Service through
2241 memory exhaustion. However, the excessive message length check still takes
2242 place, and this would cause the connection to immediately fail. Assuming
2243 that the application calls SSL_free() on the failed connection in a timely
2244 manner then the 21Mb of allocated memory will then be immediately freed
2245 again. Therefore the excessive memory allocation will be transitory in
2246 nature. This then means that there is only a security impact if:
2248 1) The application does not call SSL_free() in a timely manner in the event
2249 that the connection fails
2251 2) The application is working in a constrained environment where there is
2252 very little free memory
2254 3) The attacker initiates multiple connection attempts such that there are
2255 multiple connections in a state where memory has been allocated for the
2256 connection; SSL_free() has not yet been called; and there is insufficient
2257 memory to service the multiple requests.
2259 Except in the instance of (1) above any Denial Of Service is likely to be
2260 transitory because as soon as the connection fails the memory is
2261 subsequently freed again in the SSL_free() call. However there is an
2262 increased risk during this period of application crashes due to the lack of
2263 memory - which would then mean a more serious Denial of Service.
2265 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2266 (CVE-2016-6307 and CVE-2016-6308)
2270 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2271 had to be removed. Primary reason is that vendor assembler can't
2272 assemble our modules with -KPIC flag. As result it, assembly
2273 support, was not even available as option. But its lack means
2274 lack of side-channel resistant code, which is incompatible with
2275 security by todays standards. Fortunately gcc is readily available
2276 prepackaged option, which we firmly point at...
2280 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016] ###
2282 * Windows command-line tool supports UTF-8 opt-in option for arguments
2283 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2284 (to any value) allows Windows user to access PKCS#12 file generated
2285 with Windows CryptoAPI and protected with non-ASCII password, as well
2286 as files generated under UTF-8 locale on Linux also protected with
2291 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2292 have been disabled by default and removed from DEFAULT, just like RC4.
2293 See the RC4 item below to re-enable both.
2297 * The method for finding the storage location for the Windows RAND seed file
2298 has changed. First we check %RANDFILE%. If that is not set then we check
2299 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2300 all else fails we fall back to C:\.
2304 * The EVP_EncryptUpdate() function has had its return type changed from void
2305 to int. A return of 0 indicates and error while a return of 1 indicates
2310 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2311 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2312 off the constant time implementation for RSA, DSA and DH have been made
2313 no-ops and deprecated.
2317 * Windows RAND implementation was simplified to only get entropy by
2318 calling CryptGenRandom(). Various other RAND-related tickets
2321 *Joseph Wylie Yandle, Rich Salz*
2323 * The stack and lhash API's were renamed to start with OPENSSL_SK_
2324 and OPENSSL_LH_, respectively. The old names are available
2325 with API compatibility. They new names are now completely documented.
2329 * Unify TYPE_up_ref(obj) methods signature.
2330 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2331 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2332 int (instead of void) like all others TYPE_up_ref() methods.
2333 So now these methods also check the return value of CRYPTO_atomic_add(),
2334 and the validity of object reference counter.
2336 *fdasilvayy@gmail.com*
2338 * With Windows Visual Studio builds, the .pdb files are installed
2339 alongside the installed libraries and executables. For a static
2340 library installation, ossl_static.pdb is the associate compiler
2341 generated .pdb file to be used when linking programs.
2345 * Remove openssl.spec. Packaging files belong with the packagers.
2349 * Automatic Darwin/OSX configuration has had a refresh, it will now
2350 recognise x86_64 architectures automatically. You can still decide
2351 to build for a different bitness with the environment variable
2352 KERNEL_BITS (can be 32 or 64), for example:
2354 KERNEL_BITS=32 ./config
2358 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2359 256 bit AES and HMAC with SHA256.
2363 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2367 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2371 * To enable users to have their own config files and build file templates,
2372 Configure looks in the directory indicated by the environment variable
2373 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2374 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2375 name and is used as is.
2379 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2380 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2381 X509_CERT_FILE_CTX was removed.
2385 * "shared" builds are now the default. To create only static libraries use
2386 the "no-shared" Configure option.
2390 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2391 All of these option have not worked for some while and are fundamental
2396 * Make various cleanup routines no-ops and mark them as deprecated. Most
2397 global cleanup functions are no longer required because they are handled
2398 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2399 Explicitly de-initing can cause problems (e.g. where a library that uses
2400 OpenSSL de-inits, but an application is still using it). The affected
2401 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2402 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2403 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2404 COMP_zlib_cleanup().
2408 * --strict-warnings no longer enables runtime debugging options
2409 such as REF_DEBUG. Instead, debug options are automatically
2410 enabled with '--debug' builds.
2412 *Andy Polyakov, Emilia Käsper*
2414 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2415 have been moved out of the public header files. New functions for managing
2416 these have been added.
2420 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2421 objects have been moved out of the public header files. New
2422 functions for managing these have been added.
2426 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2427 have been moved out of the public header files. New functions for managing
2428 these have been added.
2432 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2433 moved out of the public header files. New functions for managing these
2438 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2442 * Removed the mk1mf build scripts.
2446 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2447 it is always safe to #include a header now.
2451 * Removed the aged BC-32 config and all its supporting scripts
2455 * Removed support for Ultrix, Netware, and OS/2.
2459 * Add support for HKDF.
2461 *Alessandro Ghedini*
2463 * Add support for blake2b and blake2s
2467 * Added support for "pipelining". Ciphers that have the
2468 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2469 encryptions/decryptions simultaneously. There are currently no built-in
2470 ciphers with this property but the expectation is that engines will be able
2471 to offer it to significantly improve throughput. Support has been extended
2472 into libssl so that multiple records for a single connection can be
2473 processed in one go (for >=TLS 1.1).
2477 * Added the AFALG engine. This is an async capable engine which is able to
2478 offload work to the Linux kernel. In this initial version it only supports
2479 AES128-CBC. The kernel must be version 4.1.0 or greater.
2483 * OpenSSL now uses a new threading API. It is no longer necessary to
2484 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2485 are two supported threading models: pthreads and windows threads. It is
2486 also possible to configure OpenSSL at compile time for "no-threads". The
2487 old threading API should no longer be used. The functions have been
2488 replaced with "no-op" compatibility macros.
2490 *Alessandro Ghedini, Matt Caswell*
2492 * Modify behavior of ALPN to invoke callback after SNI/servername
2493 callback, such that updates to the SSL_CTX affect ALPN.
2497 * Add SSL_CIPHER queries for authentication and key-exchange.
2501 * Changes to the DEFAULT cipherlist:
2502 - Prefer (EC)DHE handshakes over plain RSA.
2503 - Prefer AEAD ciphers over legacy ciphers.
2504 - Prefer ECDSA over RSA when both certificates are available.
2505 - Prefer TLSv1.2 ciphers/PRF.
2506 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2511 * Change the ECC default curve list to be this, in order: x25519,
2512 secp256r1, secp521r1, secp384r1.
2516 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2517 disabled by default. They can be re-enabled using the
2518 enable-weak-ssl-ciphers option to Configure.
2522 * If the server has ALPN configured, but supports no protocols that the
2523 client advertises, send a fatal "no_application_protocol" alert.
2524 This behaviour is SHALL in RFC 7301, though it isn't universally
2525 implemented by other servers.
2529 * Add X25519 support.
2530 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2531 for public and private key encoding using the format documented in
2532 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2533 key generation and key derivation.
2535 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2540 * Deprecate SRP_VBASE_get_by_user.
2541 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2542 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2543 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2544 seed, even if the seed is configured.
2546 Users should use SRP_VBASE_get1_by_user instead. Note that in
2547 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2548 also that even though configuring the SRP seed attempts to hide
2549 invalid usernames by continuing the handshake with fake
2550 credentials, this behaviour is not constant time and no strong
2551 guarantees are made that the handshake is indistinguishable from
2552 that of a valid user.
2556 * Configuration change; it's now possible to build dynamic engines
2557 without having to build shared libraries and vice versa. This
2558 only applies to the engines in engines/, those in crypto/engine/
2559 will always be built into libcrypto (i.e. "static").
2561 Building dynamic engines is enabled by default; to disable, use
2562 the configuration option "disable-dynamic-engine".
2564 The only requirements for building dynamic engines are the
2565 presence of the DSO module and building with position independent
2566 code, so they will also automatically be disabled if configuring
2567 with "disable-dso" or "disable-pic".
2569 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2570 are also taken away from openssl/opensslconf.h, as they are
2575 * Configuration change; if there is a known flag to compile
2576 position independent code, it will always be applied on the
2577 libcrypto and libssl object files, and never on the application
2578 object files. This means other libraries that use routines from
2579 libcrypto / libssl can be made into shared libraries regardless
2580 of how OpenSSL was configured.
2582 If this isn't desirable, the configuration options "disable-pic"
2583 or "no-pic" can be used to disable the use of PIC. This will
2584 also disable building shared libraries and dynamic engines.
2588 * Removed JPAKE code. It was experimental and has no wide use.
2592 * The INSTALL_PREFIX Makefile variable has been renamed to
2593 DESTDIR. That makes for less confusion on what this variable
2594 is for. Also, the configuration option --install_prefix is
2599 * Heartbeat for TLS has been removed and is disabled by default
2600 for DTLS; configure with enable-heartbeats. Code that uses the
2601 old #define's might need to be updated.
2603 *Emilia Käsper, Rich Salz*
2605 * Rename REF_CHECK to REF_DEBUG.
2609 * New "unified" build system
2611 The "unified" build system is aimed to be a common system for all
2612 platforms we support. With it comes new support for VMS.
2614 This system builds supports building in a different directory tree
2615 than the source tree. It produces one Makefile (for unix family
2616 or lookalikes), or one descrip.mms (for VMS).
2618 The source of information to make the Makefile / descrip.mms is
2619 small files called 'build.info', holding the necessary
2620 information for each directory with source to compile, and a
2621 template in Configurations, like unix-Makefile.tmpl or
2624 With this change, the library names were also renamed on Windows
2625 and on VMS. They now have names that are closer to the standard
2626 on Unix, and include the major version number, and in certain
2627 cases, the architecture they are built for. See "Notes on shared
2628 libraries" in INSTALL.
2630 We rely heavily on the perl module Text::Template.
2634 * Added support for auto-initialisation and de-initialisation of the library.
2635 OpenSSL no longer requires explicit init or deinit routines to be called,
2636 except in certain circumstances. See the OPENSSL_init_crypto() and
2637 OPENSSL_init_ssl() man pages for further information.
2641 * The arguments to the DTLSv1_listen function have changed. Specifically the
2642 "peer" argument is now expected to be a BIO_ADDR object.
2644 * Rewrite of BIO networking library. The BIO library lacked consistent
2645 support of IPv6, and adding it required some more extensive
2646 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2647 which hold all types of addresses and chains of address information.
2648 It also introduces a new API, with functions like BIO_socket,
2649 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2650 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2651 have been adapted accordingly.
2655 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2660 * CRIME protection: disable compression by default, even if OpenSSL is
2661 compiled with zlib enabled. Applications can still enable compression
2662 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2663 using the SSL_CONF library to configure compression.
2667 * The signature of the session callback configured with
2668 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2669 was explicitly marked as 'const unsigned char*' instead of
2674 * Always DPURIFY. Remove the use of uninitialized memory in the
2675 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2679 * Removed many obsolete configuration items, including
2680 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2681 MD2_CHAR, MD2_INT, MD2_LONG
2683 IDEA_SHORT, IDEA_LONG
2684 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2686 *Rich Salz, with advice from Andy Polyakov*
2688 * Many BN internals have been moved to an internal header file.
2690 *Rich Salz with help from Andy Polyakov*
2692 * Configuration and writing out the results from it has changed.
2693 Files such as Makefile include/openssl/opensslconf.h and are now
2694 produced through general templates, such as Makefile.in and
2695 crypto/opensslconf.h.in and some help from the perl module
2698 Also, the center of configuration information is no longer
2699 Makefile. Instead, Configure produces a perl module in
2700 configdata.pm which holds most of the config data (in the hash
2701 table %config), the target data that comes from the target
2702 configuration in one of the Configurations/*.conf files (in
2707 * To clarify their intended purposes, the Configure options
2708 --prefix and --openssldir change their semantics, and become more
2709 straightforward and less interdependent.
2711 --prefix shall be used exclusively to give the location INSTALLTOP
2712 where programs, scripts, libraries, include files and manuals are
2713 going to be installed. The default is now /usr/local.
2715 --openssldir shall be used exclusively to give the default
2716 location OPENSSLDIR where certificates, private keys, CRLs are
2717 managed. This is also where the default openssl.cnf gets
2719 If the directory given with this option is a relative path, the
2720 values of both the --prefix value and the --openssldir value will
2721 be combined to become OPENSSLDIR.
2722 The default for --openssldir is INSTALLTOP/ssl.
2724 Anyone who uses --openssldir to specify where OpenSSL is to be
2725 installed MUST change to use --prefix instead.
2729 * The GOST engine was out of date and therefore it has been removed. An up
2730 to date GOST engine is now being maintained in an external repository.
2731 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2732 support for GOST ciphersuites (these are only activated if a GOST engine
2737 * EGD is no longer supported by default; use enable-egd when
2740 *Ben Kaduk and Rich Salz*
2742 * The distribution now has Makefile.in files, which are used to
2743 create Makefile's when Configure is run. *Configure must be run
2744 before trying to build now.*
2748 * The return value for SSL_CIPHER_description() for error conditions
2753 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2755 Obtaining and performing DNSSEC validation of TLSA records is
2756 the application's responsibility. The application provides
2757 the TLSA records of its choice to OpenSSL, and these are then
2758 used to authenticate the peer.
2760 The TLSA records need not even come from DNS. They can, for
2761 example, be used to implement local end-entity certificate or
2762 trust-anchor "pinning", where the "pin" data takes the form
2763 of TLSA records, which can augment or replace verification
2764 based on the usual WebPKI public certification authorities.
2768 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2769 continues to support deprecated interfaces in default builds.
2770 However, applications are strongly advised to compile their
2771 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2772 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2773 or the 1.1.0 releases.
2775 In environments in which all applications have been ported to
2776 not use any deprecated interfaces OpenSSL's Configure script
2777 should be used with the --api=1.1.0 option to entirely remove
2778 support for the deprecated features from the library and
2779 unconditionally disable them in the installed headers.
2780 Essentially the same effect can be achieved with the "no-deprecated"
2781 argument to Configure, except that this will always restrict
2782 the build to just the latest API, rather than a fixed API
2785 As applications are ported to future revisions of the API,
2786 they should update their compile-time OPENSSL_API_COMPAT define
2787 accordingly, but in most cases should be able to continue to
2788 compile with later releases.
2790 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2791 0x10000000L and 0x00908000L, respectively. However those
2792 versions did not support the OPENSSL_API_COMPAT feature, and
2793 so applications are not typically tested for explicit support
2794 of just the undeprecated features of either release.
2798 * Add support for setting the minimum and maximum supported protocol.
2799 It can bet set via the SSL_set_min_proto_version() and
2800 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2801 MaxProtocol. It's recommended to use the new APIs to disable
2802 protocols instead of disabling individual protocols using
2803 SSL_set_options() or SSL_CONF's Protocol. This change also
2804 removes support for disabling TLS 1.2 in the OpenSSL TLS
2805 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2809 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2813 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2814 and integrates ECDSA and ECDH functionality into EC. Implementations can
2815 now redirect key generation and no longer need to convert to or from
2818 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2819 include the ec.h header file instead.
2823 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2824 ciphers who are no longer supported and drops support the ephemeral RSA key
2825 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2829 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2830 opaque. For HMAC_CTX, the following constructors and destructors
2833 HMAC_CTX *HMAC_CTX_new(void);
2834 void HMAC_CTX_free(HMAC_CTX *ctx);
2836 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2837 destroy such methods has been added. See EVP_MD_meth_new(3) and
2838 EVP_CIPHER_meth_new(3) for documentation.
2841 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2842 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2843 EVP_MD_CTX_reset() should be called instead to reinitialise
2844 an already created structure.
2845 2) For consistency with the majority of our object creators and
2846 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2847 EVP_MD_CTX_(new|free). The old names are retained as macros
2848 for deprecated builds.
2852 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2853 cryptographic operations to be performed asynchronously as long as an
2854 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2855 further details. Libssl has also had this capability integrated with the
2856 introduction of the new mode SSL_MODE_ASYNC and associated error
2857 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2858 pages. This work was developed in partnership with Intel Corp.
2862 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2863 always enabled now. If you want to disable the support you should
2864 exclude it using the list of supported ciphers. This also means that the
2865 "-no_ecdhe" option has been removed from s_server.
2869 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2870 SSL_{CTX_}set1_curves() which can set a list.
2874 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2875 curve you want to support using SSL_{CTX_}set1_curves().
2879 * State machine rewrite. The state machine code has been significantly
2880 refactored in order to remove much duplication of code and solve issues
2881 with the old code (see ssl/statem/README for further details). This change
2882 does have some associated API changes. Notably the SSL_state() function
2883 has been removed and replaced by SSL_get_state which now returns an
2884 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2885 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2890 * All instances of the string "ssleay" in the public API were replaced
2891 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2892 Some error codes related to internal RSA_eay API's were renamed.
2896 * The demo files in crypto/threads were moved to demo/threads.
2900 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2903 *Matt Caswell, Rich Salz*
2905 * New ASN.1 embed macro.
2907 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2908 structure is not allocated: it is part of the parent. That is instead of
2916 This reduces memory fragmentation and make it impossible to accidentally
2917 set a mandatory field to NULL.
2919 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2920 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2921 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2926 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2930 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2931 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2932 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2933 DES and RC4 ciphersuites.
2937 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2938 This changes the decoding behaviour for some invalid messages,
2939 though the change is mostly in the more lenient direction, and
2940 legacy behaviour is preserved as much as possible.
2944 * Fix no-stdio build.
2945 * David Woodhouse <David.Woodhouse@intel.com> and also
2946 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
2948 * New testing framework
2949 The testing framework has been largely rewritten and is now using
2950 perl and the perl modules Test::Harness and an extended variant of
2951 Test::More called OpenSSL::Test to do its work. All test scripts in
2952 test/ have been rewritten into test recipes, and all direct calls to
2953 executables in test/Makefile have become individual recipes using the
2954 simplified testing OpenSSL::Test::Simple.
2956 For documentation on our testing modules, do:
2958 perldoc test/testlib/OpenSSL/Test/Simple.pm
2959 perldoc test/testlib/OpenSSL/Test.pm
2963 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
2964 are used; the latter aborts on memory leaks (usually checked on exit).
2965 Some undocumented "set malloc, etc., hooks" functions were removed
2966 and others were changed. All are now documented.
2970 * In DSA_generate_parameters_ex, if the provided seed is too short,
2973 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
2975 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
2976 from RFC4279, RFC4785, RFC5487, RFC5489.
2978 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
2979 original RSA_PSK patch.
2983 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
2984 era flag was never set throughout the codebase (only read). Also removed
2985 SSL3_FLAGS_POP_BUFFER which was only used if
2986 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
2990 * Changed the default name options in the "ca", "crl", "req" and "x509"
2991 to be "oneline" instead of "compat".
2995 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
2996 not aware of clients that still exhibit this bug, and the workaround
2997 hasn't been working properly for a while.
3001 * The return type of BIO_number_read() and BIO_number_written() as well as
3002 the corresponding num_read and num_write members in the BIO structure has
3003 changed from unsigned long to uint64_t. On platforms where an unsigned
3004 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3009 * Given the pervasive nature of TLS extensions it is inadvisable to run
3010 OpenSSL without support for them. It also means that maintaining
3011 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3012 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3016 * Removed support for the two export grade static DH ciphersuites
3017 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3018 were newly added (along with a number of other static DH ciphersuites) to
3019 1.0.2. However the two export ones have *never* worked since they were
3020 introduced. It seems strange in any case to be adding new export
3021 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3025 * Version negotiation has been rewritten. In particular SSLv23_method(),
3026 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3027 and turned into macros which simply call the new preferred function names
3028 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3029 should use the new names instead. Also as part of this change the ssl23.h
3030 header file has been removed.
3034 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3035 code and the associated standard is no longer considered fit-for-purpose.
3039 * RT2547 was closed. When generating a private key, try to make the
3040 output file readable only by the owner. This behavior change might
3041 be noticeable when interacting with other software.
3043 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3048 * Added HTTP GET support to the ocsp command.
3052 * Changed default digest for the dgst and enc commands from MD5 to
3057 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3061 * Added support for TLS extended master secret from
3062 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3063 initial patch which was a great help during development.
3067 * All libssl internal structures have been removed from the public header
3068 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3069 now redundant). Users should not attempt to access internal structures
3070 directly. Instead they should use the provided API functions.
3074 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3075 Access to deprecated functions can be re-enabled by running config with
3076 "enable-deprecated". In addition applications wishing to use deprecated
3077 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3078 will, by default, disable some transitive includes that previously existed
3079 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3083 * Added support for OCB mode. OpenSSL has been granted a patent license
3084 compatible with the OpenSSL license for use of OCB. Details are available
3085 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
3086 for OCB can be removed by calling config with no-ocb.
3090 * SSLv2 support has been removed. It still supports receiving a SSLv2
3091 compatible client hello.
3095 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3096 done while fixing the error code for the key-too-small case.
3098 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3100 * CA.sh has been removed; use CA.pl instead.
3104 * Removed old DES API.
3108 * Remove various unsupported platforms:
3114 Sinix/ReliantUNIX RM400
3119 16-bit platforms such as WIN16
3123 * Clean up OPENSSL_NO_xxx #define's
3124 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3125 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3126 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3127 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3128 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3129 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3130 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3131 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3132 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3133 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3137 * Cleaned up dead code
3138 Remove all but one '#ifdef undef' which is to be looked at.
3142 * Clean up calling of xxx_free routines.
3143 Just like free(), fix most of the xxx_free routines to accept
3144 NULL. Remove the non-null checks from callers. Save much code.
3148 * Add secure heap for storage of private keys (when possible).
3149 Add BIO_s_secmem(), CBIGNUM, etc.
3150 Contributed by Akamai Technologies under our Corporate CLA.
3154 * Experimental support for a new, fast, unbiased prime candidate generator,
3155 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3157 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3159 * New output format NSS in the sess_id command line tool. This allows
3160 exporting the session id and the master key in NSS keylog format.
3162 *Martin Kaiser <martin@kaiser.cx>*
3164 * Harmonize version and its documentation. -f flag is used to display
3167 *mancha <mancha1@zoho.com>*
3169 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3170 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3172 *mancha <mancha1@zoho.com>*
3174 * Fix some double frees. These are not thought to be exploitable.
3176 *mancha <mancha1@zoho.com>*
3178 * A missing bounds check in the handling of the TLS heartbeat extension
3179 can be used to reveal up to 64k of memory to a connected client or
3182 Thanks for Neel Mehta of Google Security for discovering this bug and to
3183 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3184 preparing the fix [CVE-2014-0160][]
3186 *Adam Langley, Bodo Moeller*
3188 * Fix for the attack described in the paper "Recovering OpenSSL
3189 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3190 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3191 http://eprint.iacr.org/2014/140
3193 Thanks to Yuval Yarom and Naomi Benger for discovering this
3194 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3196 *Yuval Yarom and Naomi Benger*
3198 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3199 this fixes a limitation in previous versions of OpenSSL.
3203 * Experimental encrypt-then-mac support.
3205 Experimental support for encrypt then mac from
3206 draft-gutmann-tls-encrypt-then-mac-02.txt
3208 To enable it set the appropriate extension number (0x42 for the test
3209 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3211 For non-compliant peers (i.e. just about everything) this should have no
3214 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3219 * Add EVP support for key wrapping algorithms, to avoid problems with
3220 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3221 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3222 algorithms and include tests cases.
3226 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3231 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3232 MGF1 digest and OAEP label.
3236 * Make openssl verify return errors.
3238 *Chris Palmer <palmer@google.com> and Ben Laurie*
3240 * New function ASN1_TIME_diff to calculate the difference between two
3241 ASN1_TIME structures or one structure and the current time.
3245 * Update fips_test_suite to support multiple command line options. New
3246 test to induce all self test errors in sequence and check expected
3251 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3252 sign or verify all in one operation.
3256 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3257 test programs and fips_test_suite. Includes functionality to parse
3258 the minimal script output of fipsalgest.pl directly.
3262 * Add authorisation parameter to FIPS_module_mode_set().
3266 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3270 * Use separate DRBG fields for internal and external flags. New function
3271 FIPS_drbg_health_check() to perform on demand health checking. Add
3272 generation tests to fips_test_suite with reduced health check interval to
3273 demonstrate periodic health checking. Add "nodh" option to
3274 fips_test_suite to skip very slow DH test.
3278 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3283 * More extensive health check for DRBG checking many more failure modes.
3284 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3285 combination: call this in fips_test_suite.
3289 * Add support for canonical generation of DSA parameter 'g'. See
3292 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3293 POST to handle HMAC cases.
3297 * Add functions FIPS_module_version() and FIPS_module_version_text()
3298 to return numerical and string versions of the FIPS module number.
3302 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3303 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3304 outside the validated module in the FIPS capable OpenSSL.
3308 * Minor change to DRBG entropy callback semantics. In some cases
3309 there is no multiple of the block length between min_len and
3310 max_len. Allow the callback to return more than max_len bytes
3311 of entropy but discard any extra: it is the callback's responsibility
3312 to ensure that the extra data discarded does not impact the
3313 requested amount of entropy.
3317 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3318 information in FIPS186-3, SP800-57 and SP800-131A.
3322 * CCM support via EVP. Interface is very similar to GCM case except we
3323 must supply all data in one chunk (i.e. no update, final) and the
3324 message length must be supplied if AAD is used. Add algorithm test
3329 * Initial version of POST overhaul. Add POST callback to allow the status
3330 of POST to be monitored and/or failures induced. Modify fips_test_suite
3331 to use callback. Always run all selftests even if one fails.
3335 * XTS support including algorithm test driver in the fips_gcmtest program.
3336 Note: this does increase the maximum key length from 32 to 64 bytes but
3337 there should be no binary compatibility issues as existing applications
3338 will never use XTS mode.
3342 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3343 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3344 performs algorithm blocking for unapproved PRNG types. Also do not
3345 set PRNG type in FIPS_mode_set(): leave this to the application.
3346 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3347 the standard OpenSSL PRNG: set additional data to a date time vector.
3351 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3352 This shouldn't present any incompatibility problems because applications
3353 shouldn't be using these directly and any that are will need to rethink
3354 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3358 * Extensive self tests and health checking required by SP800-90 DRBG.
3359 Remove strength parameter from FIPS_drbg_instantiate and always
3360 instantiate at maximum supported strength.
3364 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3368 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3372 * New function DH_compute_key_padded() to compute a DH key and pad with
3373 leading zeroes if needed: this complies with SP800-56A et al.
3377 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3378 anything, incomplete, subject to change and largely untested at present.
3382 * Modify fipscanisteronly build option to only build the necessary object
3383 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3387 * Add experimental option FIPSSYMS to give all symbols in
3388 fipscanister.o and FIPS or fips prefix. This will avoid
3389 conflicts with future versions of OpenSSL. Add perl script
3390 util/fipsas.pl to preprocess assembly language source files
3391 and rename any affected symbols.
3395 * Add selftest checks and algorithm block of non-fips algorithms in
3396 FIPS mode. Remove DES2 from selftests.
3400 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3401 return internal method without any ENGINE dependencies. Add new
3402 tiny fips sign and verify functions.
3406 * New build option no-ec2m to disable characteristic 2 code.
3410 * New build option "fipscanisteronly". This only builds fipscanister.o
3411 and (currently) associated fips utilities. Uses the file Makefile.fips
3412 instead of Makefile.org as the prototype.
3416 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3417 Update fips_gcmtest to use IV generator.
3421 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3422 setting output buffer to NULL. The *Final function must be
3423 called although it will not retrieve any additional data. The tag
3424 can be set or retrieved with a ctrl. The IV length is by default 12
3425 bytes (96 bits) but can be set to an alternative value. If the IV
3426 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3431 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3432 underlying do_cipher function handles all cipher semantics itself
3433 including padding and finalisation. This is useful if (for example)
3434 an ENGINE cipher handles block padding itself. The behaviour of
3435 do_cipher is subtly changed if this flag is set: the return value
3436 is the number of characters written to the output buffer (zero is
3437 no longer an error code) or a negative error code. Also if the
3438 input buffer is NULL and length 0 finalisation should be performed.
3442 * If a candidate issuer certificate is already part of the constructed
3443 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3447 * Improve forward-security support: add functions
3449 void SSL_CTX_set_not_resumable_session_callback(
3450 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3451 void SSL_set_not_resumable_session_callback(
3452 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3454 for use by SSL/TLS servers; the callback function will be called whenever a
3455 new session is created, and gets to decide whether the session may be
3456 cached to make it resumable (return 0) or not (return 1). (As by the
3457 SSL/TLS protocol specifications, the session_id sent by the server will be
3458 empty to indicate that the session is not resumable; also, the server will
3459 not generate RFC 4507 (RFC 5077) session tickets.)
3461 A simple reasonable callback implementation is to return is_forward_secure.
3462 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3463 by the SSL/TLS server library, indicating whether it can provide forward
3466 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3468 * New -verify_name option in command line utilities to set verification
3473 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3474 Add CMAC pkey methods.
3478 * Experimental renegotiation in s_server -www mode. If the client
3479 browses /reneg connection is renegotiated. If /renegcert it is
3480 renegotiated requesting a certificate.
3484 * Add an "external" session cache for debugging purposes to s_server. This
3485 should help trace issues which normally are only apparent in deployed
3486 multi-process servers.
3490 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3491 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3492 BIO_set_cipher() and some obscure PEM functions were changed so they
3493 can now return an error. The RAND changes required a change to the
3494 RAND_METHOD structure.
3498 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3499 a gcc attribute to warn if the result of a function is ignored. This
3500 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3501 whose return value is often ignored.
3505 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3506 These allow SCTs (signed certificate timestamps) to be requested and
3507 validated when establishing a connection.
3509 *Rob Percival <robpercival@google.com>*
3514 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019] ###
3516 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3517 used even when parsing explicit parameters, when loading a serialized key
3518 or calling `EC_GROUP_new_from_ecpkparameters()`/
3519 `EC_GROUP_new_from_ecparameters()`.
3520 This prevents bypass of security hardening and performance gains,
3521 especially for curves with specialized EC_METHODs.
3522 By default, if a key encoded with explicit parameters is loaded and later
3523 serialized, the output is still encoded with explicit parameters, even if
3524 internally a "named" EC_GROUP is used for computation.
3528 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3529 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3530 NULL. After this change, only the cofactor parameter can be NULL. It also
3531 does some minimal sanity checks on the passed order.
3536 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3537 An attack is simple, if the first CMS_recipientInfo is valid but the
3538 second CMS_recipientInfo is chosen ciphertext. If the second
3539 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3540 encryption key will be replaced by garbage, and the message cannot be
3541 decoded, but if the RSA decryption fails, the correct encryption key is
3542 used and the recipient will not notice the attack.
3543 As a work around for this potential attack the length of the decrypted
3544 key must be equal to the cipher default key length, in case the
3545 certifiate is not given and all recipientInfo are tried out.
3546 The old behaviour can be re-enabled in the CMS code by setting the
3547 CMS_DEBUG_DECRYPT flag.
3552 * Document issue with installation paths in diverse Windows builds
3554 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3555 binaries and run-time config file.
3560 ### Changes between 1.0.2r and 1.0.2s [28 May 2019] ###
3562 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3563 This changes the size when using the genpkey app when no size is given. It
3564 fixes an omission in earlier changes that changed all RSA, DSA and DH
3565 generation apps to use 2048 bits by default.
3569 * Add FIPS support for Android Arm 64-bit
3571 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3572 Module in Version 2.0.10. For some reason, the corresponding target
3573 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3574 built with FIPS support on Android Arm 64-bit. This omission has been
3577 *Matthias St. Pierre*
3579 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019] ###
3581 * 0-byte record padding oracle
3583 If an application encounters a fatal protocol error and then calls
3584 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3585 then OpenSSL can respond differently to the calling application if a 0 byte
3586 record is received with invalid padding compared to if a 0 byte record is
3587 received with an invalid MAC. If the application then behaves differently
3588 based on that in a way that is detectable to the remote peer, then this
3589 amounts to a padding oracle that could be used to decrypt data.
3591 In order for this to be exploitable "non-stitched" ciphersuites must be in
3592 use. Stitched ciphersuites are optimised implementations of certain
3593 commonly used ciphersuites. Also the application must call SSL_shutdown()
3594 twice even if a protocol error has occurred (applications should not do
3595 this but some do anyway).
3597 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3598 Aviram, with additional investigation by Steven Collison and Andrew
3599 Hourselt. It was reported to OpenSSL on 10th December 2018.
3604 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3608 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018] ###
3610 * Microarchitecture timing vulnerability in ECC scalar multiplication
3612 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3613 shown to be vulnerable to a microarchitecture timing side channel attack.
3614 An attacker with sufficient access to mount local timing attacks during
3615 ECDSA signature generation could recover the private key.
3617 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3618 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3624 * Timing vulnerability in DSA signature generation
3626 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3627 timing side channel attack. An attacker could use variations in the signing
3628 algorithm to recover the private key.
3630 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3635 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3636 Module, accidentally introduced while backporting security fixes from the
3637 development branch and hindering the use of ECC in FIPS mode.
3641 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018] ###
3643 * Client DoS due to large DH parameter
3645 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3646 malicious server can send a very large prime value to the client. This will
3647 cause the client to spend an unreasonably long period of time generating a
3648 key for this prime resulting in a hang until the client has finished. This
3649 could be exploited in a Denial Of Service attack.
3651 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3656 * Cache timing vulnerability in RSA Key Generation
3658 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3659 a cache timing side channel attack. An attacker with sufficient access to
3660 mount cache timing attacks during the RSA key generation process could
3661 recover the private key.
3663 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3664 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3669 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3670 parameter is no longer accepted, as it leads to a corrupt table. NULL
3671 pem_str is reserved for alias entries only.
3675 * Revert blinding in ECDSA sign and instead make problematic addition
3676 length-invariant. Switch even to fixed-length Montgomery multiplication.
3680 * Change generating and checking of primes so that the error rate of not
3681 being prime depends on the intended use based on the size of the input.
3682 For larger primes this will result in more rounds of Miller-Rabin.
3683 The maximal error rate for primes with more than 1080 bits is lowered
3686 *Kurt Roeckx, Annie Yousar*
3688 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3692 * Add blinding to ECDSA and DSA signatures to protect against side channel
3693 attacks discovered by Keegan Ryan (NCC Group).
3697 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3698 now allow empty (zero character) pass phrases.
3702 * Certificate time validation (X509_cmp_time) enforces stricter
3703 compliance with RFC 5280. Fractional seconds and timezone offsets
3704 are no longer allowed.
3708 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018] ###
3710 * Constructed ASN.1 types with a recursive definition could exceed the stack
3712 Constructed ASN.1 types with a recursive definition (such as can be found
3713 in PKCS7) could eventually exceed the stack given malicious input with
3714 excessive recursion. This could result in a Denial Of Service attack. There
3715 are no such structures used within SSL/TLS that come from untrusted sources
3716 so this is considered safe.
3718 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3724 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017] ###
3726 * Read/write after SSL object in error state
3728 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3729 mechanism. The intent was that if a fatal error occurred during a handshake
3730 then OpenSSL would move into the error state and would immediately fail if
3731 you attempted to continue the handshake. This works as designed for the
3732 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3733 SSL_connect()), however due to a bug it does not work correctly if
3734 SSL_read() or SSL_write() is called directly. In that scenario, if the
3735 handshake fails then a fatal error will be returned in the initial function
3736 call. If SSL_read()/SSL_write() is subsequently called by the application
3737 for the same SSL object then it will succeed and the data is passed without
3738 being decrypted/encrypted directly from the SSL/TLS record layer.
3740 In order to exploit this issue an application bug would have to be present
3741 that resulted in a call to SSL_read()/SSL_write() being issued after having
3742 already received a fatal error.
3744 This issue was reported to OpenSSL by David Benjamin (Google).
3749 * rsaz_1024_mul_avx2 overflow bug on x86_64
3751 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3752 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3753 Analysis suggests that attacks against RSA and DSA as a result of this
3754 defect would be very difficult to perform and are not believed likely.
3755 Attacks against DH1024 are considered just feasible, because most of the
3756 work necessary to deduce information about a private key may be performed
3757 offline. The amount of resources required for such an attack would be
3758 significant. However, for an attack on TLS to be meaningful, the server
3759 would have to share the DH1024 private key among multiple clients, which is
3760 no longer an option since CVE-2016-0701.
3762 This only affects processors that support the AVX2 but not ADX extensions
3763 like Intel Haswell (4th generation).
3765 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3766 was originally found via the OSS-Fuzz project.
3771 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017] ###
3773 * bn_sqrx8x_internal carry bug on x86_64
3775 There is a carry propagating bug in the x86_64 Montgomery squaring
3776 procedure. No EC algorithms are affected. Analysis suggests that attacks
3777 against RSA and DSA as a result of this defect would be very difficult to
3778 perform and are not believed likely. Attacks against DH are considered just
3779 feasible (although very difficult) because most of the work necessary to
3780 deduce information about a private key may be performed offline. The amount
3781 of resources required for such an attack would be very significant and
3782 likely only accessible to a limited number of attackers. An attacker would
3783 additionally need online access to an unpatched system using the target
3784 private key in a scenario with persistent DH parameters and a private
3785 key that is shared between multiple clients.
3787 This only affects processors that support the BMI1, BMI2 and ADX extensions
3788 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3790 This issue was reported to OpenSSL by the OSS-Fuzz project.
3795 * Malformed X.509 IPAddressFamily could cause OOB read
3797 If an X.509 certificate has a malformed IPAddressFamily extension,
3798 OpenSSL could do a one-byte buffer overread. The most likely result
3799 would be an erroneous display of the certificate in text format.
3801 This issue was reported to OpenSSL by the OSS-Fuzz project.
3806 ### Changes between 1.0.2k and 1.0.2l [25 May 2017] ###
3808 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3809 platform rather than 'mingw'.
3813 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017] ###
3815 * Truncated packet could crash via OOB read
3817 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3818 cipher is being used, then a truncated packet can cause that host to
3819 perform an out-of-bounds read, usually resulting in a crash.
3821 This issue was reported to OpenSSL by Robert Święcki of Google.
3826 * BN_mod_exp may produce incorrect results on x86_64
3828 There is a carry propagating bug in the x86_64 Montgomery squaring
3829 procedure. No EC algorithms are affected. Analysis suggests that attacks
3830 against RSA and DSA as a result of this defect would be very difficult to
3831 perform and are not believed likely. Attacks against DH are considered just
3832 feasible (although very difficult) because most of the work necessary to
3833 deduce information about a private key may be performed offline. The amount
3834 of resources required for such an attack would be very significant and
3835 likely only accessible to a limited number of attackers. An attacker would
3836 additionally need online access to an unpatched system using the target
3837 private key in a scenario with persistent DH parameters and a private
3838 key that is shared between multiple clients. For example this can occur by
3839 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3840 similar to CVE-2015-3193 but must be treated as a separate problem.
3842 This issue was reported to OpenSSL by the OSS-Fuzz project.
3847 * Montgomery multiplication may produce incorrect results
3849 There is a carry propagating bug in the Broadwell-specific Montgomery
3850 multiplication procedure that handles input lengths divisible by, but
3851 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3852 and DH private keys are impossible. This is because the subroutine in
3853 question is not used in operations with the private key itself and an input
3854 of the attacker's direct choice. Otherwise the bug can manifest itself as
3855 transient authentication and key negotiation failures or reproducible
3856 erroneous outcome of public-key operations with specially crafted input.
3857 Among EC algorithms only Brainpool P-512 curves are affected and one
3858 presumably can attack ECDH key negotiation. Impact was not analyzed in
3859 detail, because pre-requisites for attack are considered unlikely. Namely
3860 multiple clients have to choose the curve in question and the server has to
3861 share the private key among them, neither of which is default behaviour.
3862 Even then only clients that chose the curve will be affected.
3864 This issue was publicly reported as transient failures and was not
3865 initially recognized as a security issue. Thanks to Richard Morgan for
3866 providing reproducible case.
3871 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3872 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3873 prevent issues where no progress is being made and the peer continually
3874 sends unrecognised record types, using up resources processing them.
3878 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016] ###
3880 * Missing CRL sanity check
3882 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
3883 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
3884 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
3886 This issue only affects the OpenSSL 1.0.2i
3891 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016] ###
3893 * OCSP Status Request extension unbounded memory growth
3895 A malicious client can send an excessively large OCSP Status Request
3896 extension. If that client continually requests renegotiation, sending a
3897 large OCSP Status Request extension each time, then there will be unbounded
3898 memory growth on the server. This will eventually lead to a Denial Of
3899 Service attack through memory exhaustion. Servers with a default
3900 configuration are vulnerable even if they do not support OCSP. Builds using
3901 the "no-ocsp" build time option are not affected.
3903 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3908 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
3911 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
3917 * OOB write in MDC2_Update()
3919 An overflow can occur in MDC2_Update() either if called directly or
3920 through the EVP_DigestUpdate() function using MDC2. If an attacker
3921 is able to supply very large amounts of input data after a previous
3922 call to EVP_EncryptUpdate() with a partial block then a length check
3923 can overflow resulting in a heap corruption.
3925 The amount of data needed is comparable to SIZE_MAX which is impractical
3928 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3933 * Malformed SHA512 ticket DoS
3935 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
3936 DoS attack where a malformed ticket will result in an OOB read which will
3939 The use of SHA512 in TLS session tickets is comparatively rare as it requires
3940 a custom server callback and ticket lookup mechanism.
3942 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3947 * OOB write in BN_bn2dec()
3949 The function BN_bn2dec() does not check the return value of BN_div_word().
3950 This can cause an OOB write if an application uses this function with an
3951 overly large BIGNUM. This could be a problem if an overly large certificate
3952 or CRL is printed out from an untrusted source. TLS is not affected because
3953 record limits will reject an oversized certificate before it is parsed.
3955 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3960 * OOB read in TS_OBJ_print_bio()
3962 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
3963 the total length the OID text representation would use and not the amount
3964 of data written. This will result in OOB reads when large OIDs are
3967 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3972 * Pointer arithmetic undefined behaviour
3974 Avoid some undefined pointer arithmetic
3976 A common idiom in the codebase is to check limits in the following manner:
3979 Where "p" points to some malloc'd data of SIZE bytes and
3982 "len" here could be from some externally supplied data (e.g. from a TLS
3985 The rules of C pointer arithmetic are such that "p + len" is only well
3986 defined where len <= SIZE. Therefore the above idiom is actually
3987 undefined behaviour.
3989 For example this could cause problems if some malloc implementation
3990 provides an address for "p" such that "p + len" actually overflows for
3991 values of len that are too big and therefore p + len < limit.
3993 This issue was reported to OpenSSL by Guido Vranken
3998 * Constant time flag not preserved in DSA signing
4000 Operations in the DSA signing algorithm should run in constant time in
4001 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4002 implementation means that a non-constant time codepath is followed for
4003 certain operations. This has been demonstrated through a cache-timing
4004 attack to be sufficient for an attacker to recover the private DSA key.
4006 This issue was reported by César Pereida (Aalto University), Billy Brumley
4007 (Tampere University of Technology), and Yuval Yarom (The University of
4008 Adelaide and NICTA).
4013 * DTLS buffered message DoS
4015 In a DTLS connection where handshake messages are delivered out-of-order
4016 those messages that OpenSSL is not yet ready to process will be buffered
4017 for later use. Under certain circumstances, a flaw in the logic means that
4018 those messages do not get removed from the buffer even though the handshake
4019 has been completed. An attacker could force up to approx. 15 messages to
4020 remain in the buffer when they are no longer required. These messages will
4021 be cleared when the DTLS connection is closed. The default maximum size for
4022 a message is 100k. Therefore the attacker could force an additional 1500k
4023 to be consumed per connection. By opening many simulataneous connections an
4024 attacker could cause a DoS attack through memory exhaustion.
4026 This issue was reported to OpenSSL by Quan Luo.
4031 * DTLS replay protection DoS
4033 A flaw in the DTLS replay attack protection mechanism means that records
4034 that arrive for future epochs update the replay protection "window" before
4035 the MAC for the record has been validated. This could be exploited by an
4036 attacker by sending a record for the next epoch (which does not have to
4037 decrypt or have a valid MAC), with a very large sequence number. This means
4038 that all subsequent legitimate packets are dropped causing a denial of
4039 service for a specific DTLS connection.
4041 This issue was reported to OpenSSL by the OCAP audit team.
4046 * Certificate message OOB reads
4048 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4049 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4050 theoretical DoS risk but this has not been observed in practice on common
4053 The messages affected are client certificate, client certificate request
4054 and server certificate. As a result the attack can only be performed
4055 against a client or a server which enables client authentication.
4057 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4062 ### Changes between 1.0.2g and 1.0.2h [3 May 2016] ###
4064 * Prevent padding oracle in AES-NI CBC MAC check
4066 A MITM attacker can use a padding oracle attack to decrypt traffic
4067 when the connection uses an AES CBC cipher and the server support
4070 This issue was introduced as part of the fix for Lucky 13 padding
4071 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4072 constant time by making sure that always the same bytes are read and
4073 compared against either the MAC or padding bytes. But it no longer
4074 checked that there was enough data to have both the MAC and padding
4077 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4082 * Fix EVP_EncodeUpdate overflow
4084 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4085 Base64 encoding of binary data. If an attacker is able to supply very large
4086 amounts of input data then a length check can overflow resulting in a heap
4089 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4090 the PEM_write_bio* family of functions. These are mainly used within the
4091 OpenSSL command line applications, so any application which processes data
4092 from an untrusted source and outputs it as a PEM file should be considered
4093 vulnerable to this issue. User applications that call these APIs directly
4094 with large amounts of untrusted data may also be vulnerable.
4096 This issue was reported by Guido Vranken.
4101 * Fix EVP_EncryptUpdate overflow
4103 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4104 is able to supply very large amounts of input data after a previous call to
4105 EVP_EncryptUpdate() with a partial block then a length check can overflow
4106 resulting in a heap corruption. Following an analysis of all OpenSSL
4107 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4108 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4109 the first called function after an EVP_EncryptInit(), and therefore that
4110 specific call must be safe. The second form is where the length passed to
4111 EVP_EncryptUpdate() can be seen from the code to be some small value and
4112 therefore there is no possibility of an overflow. Since all instances are
4113 one of these two forms, it is believed that there can be no overflows in
4114 internal code due to this problem. It should be noted that
4115 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4116 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4117 of these calls have also been analysed too and it is believed there are no
4118 instances in internal usage where an overflow could occur.
4120 This issue was reported by Guido Vranken.
4125 * Prevent ASN.1 BIO excessive memory allocation
4127 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4128 a short invalid encoding can cause allocation of large amounts of memory
4129 potentially consuming excessive resources or exhausting memory.
4131 Any application parsing untrusted data through d2i BIO functions is
4132 affected. The memory based functions such as d2i_X509() are *not* affected.
4133 Since the memory based functions are used by the TLS library, TLS
4134 applications are not affected.
4136 This issue was reported by Brian Carpenter.
4143 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4144 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4145 in arbitrary stack data being returned in the buffer.
4147 This issue was reported by Guido Vranken.
4152 * Modify behavior of ALPN to invoke callback after SNI/servername
4153 callback, such that updates to the SSL_CTX affect ALPN.
4157 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4162 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4163 methods are enabled and ssl2 is disabled the methods return NULL.
4167 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016] ###
4169 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4170 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4171 provide any "EXPORT" or "LOW" strength ciphers.
4175 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4176 is by default disabled at build-time. Builds that are not configured with
4177 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4178 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4179 will need to explicitly call either of:
4181 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4183 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4185 as appropriate. Even if either of those is used, or the application
4186 explicitly uses the version-specific SSLv2_method() or its client and
4187 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4188 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4189 ciphers, and SSLv2 56-bit DES are no longer available.
4194 * Fix a double-free in DSA code
4196 A double free bug was discovered when OpenSSL parses malformed DSA private
4197 keys and could lead to a DoS attack or memory corruption for applications
4198 that receive DSA private keys from untrusted sources. This scenario is
4201 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4207 * Disable SRP fake user seed to address a server memory leak.
4209 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4211 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4212 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4213 was changed to ignore the "fake user" SRP seed, even if the seed
4216 Users should use SRP_VBASE_get1_by_user instead. Note that in
4217 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4218 also that even though configuring the SRP seed attempts to hide
4219 invalid usernames by continuing the handshake with fake
4220 credentials, this behaviour is not constant time and no strong
4221 guarantees are made that the handshake is indistinguishable from
4222 that of a valid user.
4227 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4229 In the BN_hex2bn function the number of hex digits is calculated using an
4230 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4231 large values of |i| this can result in |bn_expand| not allocating any
4232 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4233 field as NULL leading to a subsequent NULL ptr deref. For very large values
4234 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4235 In this case memory is allocated to the internal BIGNUM data field, but it
4236 is insufficiently sized leading to heap corruption. A similar issue exists
4237 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4238 is ever called by user applications with very large untrusted hex/dec data.
4239 This is anticipated to be a rare occurrence.
4241 All OpenSSL internal usage of these functions use data that is not expected
4242 to be untrusted, e.g. config file data or application command line
4243 arguments. If user developed applications generate config file data based
4244 on untrusted data then it is possible that this could also lead to security
4245 consequences. This is also anticipated to be rare.
4247 This issue was reported to OpenSSL by Guido Vranken.
4252 * Fix memory issues in BIO_*printf functions
4254 The internal |fmtstr| function used in processing a "%s" format string in
4255 the BIO_*printf functions could overflow while calculating the length of a
4256 string and cause an OOB read when printing very long strings.
4258 Additionally the internal |doapr_outch| function can attempt to write to an
4259 OOB memory location (at an offset from the NULL pointer) in the event of a
4260 memory allocation failure. In 1.0.2 and below this could be caused where
4261 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4262 could be in processing a very long "%s" format string. Memory leaks can
4265 The first issue may mask the second issue dependent on compiler behaviour.
4266 These problems could enable attacks where large amounts of untrusted data
4267 is passed to the BIO_*printf functions. If applications use these functions
4268 in this way then they could be vulnerable. OpenSSL itself uses these
4269 functions when printing out human-readable dumps of ASN.1 data. Therefore
4270 applications that print this data could be vulnerable if the data is from
4271 untrusted sources. OpenSSL command line applications could also be
4272 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4273 as command line arguments.
4275 Libssl is not considered directly vulnerable. Additionally certificates etc
4276 received via remote connections via libssl are also unlikely to be able to
4277 trigger these issues because of message size limits enforced within libssl.
4279 This issue was reported to OpenSSL Guido Vranken.
4284 * Side channel attack on modular exponentiation
4286 A side-channel attack was found which makes use of cache-bank conflicts on
4287 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4288 of RSA keys. The ability to exploit this issue is limited as it relies on
4289 an attacker who has control of code in a thread running on the same
4290 hyper-threaded core as the victim thread which is performing decryptions.
4292 This issue was reported to OpenSSL by Yuval Yarom, The University of
4293 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4294 Nadia Heninger, University of Pennsylvania with more information at
4295 http://cachebleed.info.
4300 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4301 if no keysize is specified with default_bits. This fixes an
4302 omission in an earlier change that changed all RSA/DSA key generation
4303 apps to use 2048 bits by default.
4307 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016] ###
4308 * DH small subgroups
4310 Historically OpenSSL only ever generated DH parameters based on "safe"
4311 primes. More recently (in version 1.0.2) support was provided for
4312 generating X9.42 style parameter files such as those required for RFC 5114
4313 support. The primes used in such files may not be "safe". Where an
4314 application is using DH configured with parameters based on primes that are
4315 not "safe" then an attacker could use this fact to find a peer's private
4316 DH exponent. This attack requires that the attacker complete multiple
4317 handshakes in which the peer uses the same private DH exponent. For example
4318 this could be used to discover a TLS server's private DH exponent if it's
4319 reusing the private DH exponent or it's using a static DH ciphersuite.
4321 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4322 TLS. It is not on by default. If the option is not set then the server
4323 reuses the same private DH exponent for the life of the server process and
4324 would be vulnerable to this attack. It is believed that many popular
4325 applications do set this option and would therefore not be at risk.
4327 The fix for this issue adds an additional check where a "q" parameter is
4328 available (as is the case in X9.42 based parameters). This detects the
4329 only known attack, and is the only possible defense for static DH
4330 ciphersuites. This could have some performance impact.
4332 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4333 default and cannot be disabled. This could have some performance impact.
4335 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4340 * SSLv2 doesn't block disabled ciphers
4342 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4343 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4344 been disabled, provided that the SSLv2 protocol was not also disabled via
4347 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4348 and Sebastian Schinzel.
4353 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015] ###
4355 * BN_mod_exp may produce incorrect results on x86_64
4357 There is a carry propagating bug in the x86_64 Montgomery squaring
4358 procedure. No EC algorithms are affected. Analysis suggests that attacks
4359 against RSA and DSA as a result of this defect would be very difficult to
4360 perform and are not believed likely. Attacks against DH are considered just
4361 feasible (although very difficult) because most of the work necessary to
4362 deduce information about a private key may be performed offline. The amount
4363 of resources required for such an attack would be very significant and
4364 likely only accessible to a limited number of attackers. An attacker would
4365 additionally need online access to an unpatched system using the target
4366 private key in a scenario with persistent DH parameters and a private
4367 key that is shared between multiple clients. For example this can occur by
4368 default in OpenSSL DHE based SSL/TLS ciphersuites.
4370 This issue was reported to OpenSSL by Hanno Böck.
4375 * Certificate verify crash with missing PSS parameter
4377 The signature verification routines will crash with a NULL pointer
4378 dereference if presented with an ASN.1 signature using the RSA PSS
4379 algorithm and absent mask generation function parameter. Since these
4380 routines are used to verify certificate signature algorithms this can be
4381 used to crash any certificate verification operation and exploited in a
4382 DoS attack. Any application which performs certificate verification is
4383 vulnerable including OpenSSL clients and servers which enable client
4386 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4391 * X509_ATTRIBUTE memory leak
4393 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4394 memory. This structure is used by the PKCS#7 and CMS routines so any
4395 application which reads PKCS#7 or CMS data from untrusted sources is
4396 affected. SSL/TLS is not affected.
4398 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4404 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4405 This changes the decoding behaviour for some invalid messages,
4406 though the change is mostly in the more lenient direction, and
4407 legacy behaviour is preserved as much as possible.
4411 * In DSA_generate_parameters_ex, if the provided seed is too short,
4414 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4416 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015] ###
4418 * Alternate chains certificate forgery
4420 During certificate verification, OpenSSL will attempt to find an
4421 alternative certificate chain if the first attempt to build such a chain
4422 fails. An error in the implementation of this logic can mean that an
4423 attacker could cause certain checks on untrusted certificates to be
4424 bypassed, such as the CA flag, enabling them to use a valid leaf
4425 certificate to act as a CA and "issue" an invalid certificate.
4427 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4432 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015] ###
4434 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4435 incompatibility in the handling of HMAC. The previous ABI has now been
4440 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015] ###
4442 * Malformed ECParameters causes infinite loop
4444 When processing an ECParameters structure OpenSSL enters an infinite loop
4445 if the curve specified is over a specially malformed binary polynomial
4448 This can be used to perform denial of service against any
4449 system which processes public keys, certificate requests or
4450 certificates. This includes TLS clients and TLS servers with
4451 client authentication enabled.
4453 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4458 * Exploitable out-of-bounds read in X509_cmp_time
4460 X509_cmp_time does not properly check the length of the ASN1_TIME
4461 string and can read a few bytes out of bounds. In addition,
4462 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4465 An attacker can use this to craft malformed certificates and CRLs of
4466 various sizes and potentially cause a segmentation fault, resulting in
4467 a DoS on applications that verify certificates or CRLs. TLS clients
4468 that verify CRLs are affected. TLS clients and servers with client
4469 authentication enabled may be affected if they use custom verification
4472 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4473 independently by Hanno Böck.
4478 * PKCS7 crash with missing EnvelopedContent
4480 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4481 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4482 with missing content and trigger a NULL pointer dereference on parsing.
4484 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4485 structures from untrusted sources are affected. OpenSSL clients and
4486 servers are not affected.
4488 This issue was reported to OpenSSL by Michal Zalewski (Google).
4493 * CMS verify infinite loop with unknown hash function
4495 When verifying a signedData message the CMS code can enter an infinite loop
4496 if presented with an unknown hash function OID. This can be used to perform
4497 denial of service against any system which verifies signedData messages using
4499 This issue was reported to OpenSSL by Johannes Bauer.
4504 * Race condition handling NewSessionTicket
4506 If a NewSessionTicket is received by a multi-threaded client when attempting to
4507 reuse a previous ticket then a race condition can occur potentially leading to
4508 a double free of the ticket data.
4513 * Only support 256-bit or stronger elliptic curves with the
4514 'ecdh_auto' setting (server) or by default (client). Of supported
4515 curves, prefer P-256 (both).
4519 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015] ###
4521 * ClientHello sigalgs DoS fix
4523 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4524 invalid signature algorithms extension a NULL pointer dereference will
4525 occur. This can be exploited in a DoS attack against the server.
4527 This issue was was reported to OpenSSL by David Ramos of Stanford
4531 *Stephen Henson and Matt Caswell*
4533 * Multiblock corrupted pointer fix
4535 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4536 feature only applies on 64 bit x86 architecture platforms that support AES
4537 NI instructions. A defect in the implementation of "multiblock" can cause
4538 OpenSSL's internal write buffer to become incorrectly set to NULL when
4539 using non-blocking IO. Typically, when the user application is using a
4540 socket BIO for writing, this will only result in a failed connection.
4541 However if some other BIO is used then it is likely that a segmentation
4542 fault will be triggered, thus enabling a potential DoS attack.
4544 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4549 * Segmentation fault in DTLSv1_listen fix
4551 The DTLSv1_listen function is intended to be stateless and processes the
4552 initial ClientHello from many peers. It is common for user code to loop
4553 over the call to DTLSv1_listen until a valid ClientHello is received with
4554 an associated cookie. A defect in the implementation of DTLSv1_listen means
4555 that state is preserved in the SSL object from one invocation to the next
4556 that can lead to a segmentation fault. Errors processing the initial
4557 ClientHello can trigger this scenario. An example of such an error could be
4558 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4561 This issue was reported to OpenSSL by Per Allansson.
4566 * Segmentation fault in ASN1_TYPE_cmp fix
4568 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4569 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4570 certificate signature algorithm consistency this can be used to crash any
4571 certificate verification operation and exploited in a DoS attack. Any
4572 application which performs certificate verification is vulnerable including
4573 OpenSSL clients and servers which enable client authentication.
4578 * Segmentation fault for invalid PSS parameters fix
4580 The signature verification routines will crash with a NULL pointer
4581 dereference if presented with an ASN.1 signature using the RSA PSS
4582 algorithm and invalid parameters. Since these routines are used to verify
4583 certificate signature algorithms this can be used to crash any
4584 certificate verification operation and exploited in a DoS attack. Any
4585 application which performs certificate verification is vulnerable including
4586 OpenSSL clients and servers which enable client authentication.
4588 This issue was was reported to OpenSSL by Brian Carpenter.
4593 * ASN.1 structure reuse memory corruption fix
4595 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4596 memory corruption via an invalid write. Such reuse is and has been
4597 strongly discouraged and is believed to be rare.
4599 Applications that parse structures containing CHOICE or ANY DEFINED BY
4600 components may be affected. Certificate parsing (d2i_X509 and related
4601 functions) are however not affected. OpenSSL clients and servers are
4607 * PKCS7 NULL pointer dereferences fix
4609 The PKCS#7 parsing code does not handle missing outer ContentInfo
4610 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4611 missing content and trigger a NULL pointer dereference on parsing.
4613 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4614 otherwise parse PKCS#7 structures from untrusted sources are
4615 affected. OpenSSL clients and servers are not affected.
4617 This issue was reported to OpenSSL by Michal Zalewski (Google).
4622 * DoS via reachable assert in SSLv2 servers fix
4624 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4625 servers that both support SSLv2 and enable export cipher suites by sending
4626 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4628 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4629 (OpenSSL development team).
4634 * Empty CKE with client auth and DHE fix
4636 If client auth is used then a server can seg fault in the event of a DHE
4637 ciphersuite being selected and a zero length ClientKeyExchange message
4638 being sent by the client. This could be exploited in a DoS attack.
4643 * Handshake with unseeded PRNG fix
4645 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4646 with an unseeded PRNG. The conditions are:
4647 - The client is on a platform where the PRNG has not been seeded
4648 automatically, and the user has not seeded manually
4649 - A protocol specific client method version has been used (i.e. not
4650 SSL_client_methodv23)
4651 - A ciphersuite is used that does not require additional random data from
4652 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4654 If the handshake succeeds then the client random that has been used will
4655 have been generated from a PRNG with insufficient entropy and therefore the
4656 output may be predictable.
4658 For example using the following command with an unseeded openssl will
4659 succeed on an unpatched platform:
4661 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4666 * Use After Free following d2i_ECPrivatekey error fix
4668 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4669 could cause a use after free condition. This, in turn, could cause a double
4670 free in several private key parsing functions (such as d2i_PrivateKey
4671 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4672 for applications that receive EC private keys from untrusted
4673 sources. This scenario is considered rare.
4675 This issue was discovered by the BoringSSL project and fixed in their
4681 * X509_to_X509_REQ NULL pointer deref fix
4683 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4684 the certificate key is invalid. This function is rarely used in practice.
4686 This issue was discovered by Brian Carpenter.
4691 * Removed the export ciphers from the DEFAULT ciphers
4695 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015] ###
4697 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4698 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4699 So far those who have to target multiple platforms would compromise
4700 and argue that binary targeting say ARMv5 would still execute on
4701 ARMv8. "Universal" build resolves this compromise by providing
4702 near-optimal performance even on newer platforms.
4706 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4707 (other platforms pending).
4709 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4711 * Add support for the SignedCertificateTimestampList certificate and
4712 OCSP response extensions from RFC6962.
4716 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4717 for corner cases. (Certain input points at infinity could lead to
4718 bogus results, with non-infinity inputs mapped to infinity too.)
4722 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4723 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4724 common cases are optimized and there still is room for further
4725 improvements. Vector Permutation AES for Altivec is also added.
4729 * Add support for little-endian ppc64 Linux target.
4731 *Marcelo Cerri (IBM)*
4733 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4734 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4735 are optimized and there still is room for further improvements.
4736 Both 32- and 64-bit modes are supported.
4738 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4740 * Improved ARMv7 NEON support.
4744 * Support for SPARC Architecture 2011 crypto extensions, first
4745 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4746 SHA256/512, MD5, GHASH and modular exponentiation.
4748 *Andy Polyakov, David Miller*
4750 * Accelerated modular exponentiation for Intel processors, a.k.a.
4753 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4755 * Support for new and upcoming Intel processors, including AVX2,
4756 BMI and SHA ISA extensions. This includes additional "stitched"
4757 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4760 This work was sponsored by Intel Corp.
4764 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4765 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4766 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4770 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4771 this fixes a limitation in previous versions of OpenSSL.
4775 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4776 MGF1 digest and OAEP label.
4780 * Add EVP support for key wrapping algorithms, to avoid problems with
4781 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4782 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4783 algorithms and include tests cases.
4787 * Add functions to allocate and set the fields of an ECDSA_METHOD
4790 *Douglas E. Engert, Steve Henson*
4792 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4793 difference in days and seconds between two tm or ASN1_TIME structures.
4797 * Add -rev test option to s_server to just reverse order of characters
4798 received by client and send back to server. Also prints an abbreviated
4799 summary of the connection parameters.
4803 * New option -brief for s_client and s_server to print out a brief summary
4804 of connection parameters.
4808 * Add callbacks for arbitrary TLS extensions.
4810 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4812 * New option -crl_download in several openssl utilities to download CRLs
4813 from CRLDP extension in certificates.
4817 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4821 * New function X509_CRL_diff to generate a delta CRL from the difference
4822 of two full CRLs. Add support to "crl" utility.
4826 * New functions to set lookup_crls function and to retrieve
4827 X509_STORE from X509_STORE_CTX.
4831 * Print out deprecated issuer and subject unique ID fields in
4836 * Extend OCSP I/O functions so they can be used for simple general purpose
4837 HTTP as well as OCSP. New wrapper function which can be used to download
4838 CRLs using the OCSP API.
4842 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
4846 * SSL_CONF* functions. These provide a common framework for application
4847 configuration using configuration files or command lines.
4851 * SSL/TLS tracing code. This parses out SSL/TLS records using the
4852 message callback and prints the results. Needs compile time option
4853 "enable-ssl-trace". New options to s_client and s_server to enable
4858 * New ctrl and macro to retrieve supported points extensions.
4859 Print out extension in s_server and s_client.
4863 * New functions to retrieve certificate signature and signature
4868 * Add functions to retrieve and manipulate the raw cipherlist sent by a
4873 * New Suite B modes for TLS code. These use and enforce the requirements
4874 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
4875 only use Suite B curves. The Suite B modes can be set by using the
4876 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
4880 * New chain verification flags for Suite B levels of security. Check
4881 algorithms are acceptable when flags are set in X509_verify_cert.
4885 * Make tls1_check_chain return a set of flags indicating checks passed
4886 by a certificate chain. Add additional tests to handle client
4887 certificates: checks for matching certificate type and issuer name
4892 * If an attempt is made to use a signature algorithm not in the peer
4893 preference list abort the handshake. If client has no suitable
4894 signature algorithms in response to a certificate request do not
4895 use the certificate.
4899 * If server EC tmp key is not in client preference list abort handshake.
4903 * Add support for certificate stores in CERT structure. This makes it
4904 possible to have different stores per SSL structure or one store in
4905 the parent SSL_CTX. Include distinct stores for certificate chain
4906 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
4907 to build and store a certificate chain in CERT structure: returning
4908 an error if the chain cannot be built: this will allow applications
4909 to test if a chain is correctly configured.
4911 Note: if the CERT based stores are not set then the parent SSL_CTX
4912 store is used to retain compatibility with existing behaviour.
4917 * New function ssl_set_client_disabled to set a ciphersuite disabled
4918 mask based on the current session, check mask when sending client
4919 hello and checking the requested ciphersuite.
4923 * New ctrls to retrieve and set certificate types in a certificate
4924 request message. Print out received values in s_client. If certificate
4925 types is not set with custom values set sensible values based on
4926 supported signature algorithms.
4930 * Support for distinct client and server supported signature algorithms.
4934 * Add certificate callback. If set this is called whenever a certificate
4935 is required by client or server. An application can decide which
4936 certificate chain to present based on arbitrary criteria: for example
4937 supported signature algorithms. Add very simple example to s_server.
4938 This fixes many of the problems and restrictions of the existing client
4939 certificate callback: for example you can now clear an existing
4940 certificate and specify the whole chain.
4944 * Add new "valid_flags" field to CERT_PKEY structure which determines what
4945 the certificate can be used for (if anything). Set valid_flags field
4946 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
4947 to have similar checks in it.
4949 Add new "cert_flags" field to CERT structure and include a "strict mode".
4950 This enforces some TLS certificate requirements (such as only permitting
4951 certificate signature algorithms contained in the supported algorithms
4952 extension) which some implementations ignore: this option should be used
4953 with caution as it could cause interoperability issues.
4957 * Update and tidy signature algorithm extension processing. Work out
4958 shared signature algorithms based on preferences and peer algorithms
4959 and print them out in s_client and s_server. Abort handshake if no
4960 shared signature algorithms.
4964 * Add new functions to allow customised supported signature algorithms
4965 for SSL and SSL_CTX structures. Add options to s_client and s_server
4970 * New function SSL_certs_clear() to delete all references to certificates
4971 from an SSL structure. Before this once a certificate had been added
4972 it couldn't be removed.
4976 * Integrate hostname, email address and IP address checking with certificate
4977 verification. New verify options supporting checking in openssl utility.
4981 * Fixes and wildcard matching support to hostname and email checking
4982 functions. Add manual page.
4984 *Florian Weimer (Red Hat Product Security Team)*
4986 * New functions to check a hostname email or IP address against a
4987 certificate. Add options x509 utility to print results of checks against
4992 * Fix OCSP checking.
4994 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
4996 * Initial experimental support for explicitly trusted non-root CAs.
4997 OpenSSL still tries to build a complete chain to a root but if an
4998 intermediate CA has a trust setting included that is used. The first
4999 setting is used: whether to trust (e.g., -addtrust option to the x509
5004 * Add -trusted_first option which attempts to find certificates in the
5005 trusted store even if an untrusted chain is also supplied.
5009 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5010 platform support for Linux and Android.
5014 * Support for linux-x32, ILP32 environment in x86_64 framework.
5018 * Experimental multi-implementation support for FIPS capable OpenSSL.
5019 When in FIPS mode the approved implementations are used as normal,
5020 when not in FIPS mode the internal unapproved versions are used instead.
5021 This means that the FIPS capable OpenSSL isn't forced to use the
5022 (often lower performance) FIPS implementations outside FIPS mode.
5026 * Transparently support X9.42 DH parameters when calling
5027 PEM_read_bio_DHparameters. This means existing applications can handle
5028 the new parameter format automatically.
5032 * Initial experimental support for X9.42 DH parameter format: mainly
5033 to support use of 'q' parameter for RFC5114 parameters.
5037 * Add DH parameters from RFC5114 including test data to dhtest.
5041 * Support for automatic EC temporary key parameter selection. If enabled
5042 the most preferred EC parameters are automatically used instead of
5043 hardcoded fixed parameters. Now a server just has to call:
5044 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5045 support ECDH and use the most appropriate parameters.
5049 * Enhance and tidy EC curve and point format TLS extension code. Use
5050 static structures instead of allocation if default values are used.
5051 New ctrls to set curves we wish to support and to retrieve shared curves.
5052 Print out shared curves in s_server. New options to s_server and s_client
5053 to set list of supported curves.
5057 * New ctrls to retrieve supported signature algorithms and
5058 supported curve values as an array of NIDs. Extend openssl utility
5059 to print out received values.
5063 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5064 between NIDs and the more common NIST names such as "P-256". Enhance
5065 ecparam utility and ECC method to recognise the NIST names for curves.
5069 * Enhance SSL/TLS certificate chain handling to support different
5070 chains for each certificate instead of one chain in the parent SSL_CTX.
5074 * Support for fixed DH ciphersuite client authentication: where both
5075 server and client use DH certificates with common parameters.
5079 * Support for fixed DH ciphersuites: those requiring DH server
5084 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5086 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5087 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5088 X509_CINF_get_signature were reverted post internal team review.
5094 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016] ###
5096 * OCSP Status Request extension unbounded memory growth
5098 A malicious client can send an excessively large OCSP Status Request
5099 extension. If that client continually requests renegotiation, sending a
5100 large OCSP Status Request extension each time, then there will be unbounded
5101 memory growth on the server. This will eventually lead to a Denial Of
5102 Service attack through memory exhaustion. Servers with a default
5103 configuration are vulnerable even if they do not support OCSP. Builds using
5104 the "no-ocsp" build time option are not affected.
5106 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5111 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5114 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5120 * OOB write in MDC2_Update()
5122 An overflow can occur in MDC2_Update() either if called directly or
5123 through the EVP_DigestUpdate() function using MDC2. If an attacker
5124 is able to supply very large amounts of input data after a previous
5125 call to EVP_EncryptUpdate() with a partial block then a length check
5126 can overflow resulting in a heap corruption.
5128 The amount of data needed is comparable to SIZE_MAX which is impractical
5131 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5136 * Malformed SHA512 ticket DoS
5138 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5139 DoS attack where a malformed ticket will result in an OOB read which will
5142 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5143 a custom server callback and ticket lookup mechanism.
5145 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5150 * OOB write in BN_bn2dec()
5152 The function BN_bn2dec() does not check the return value of BN_div_word().
5153 This can cause an OOB write if an application uses this function with an
5154 overly large BIGNUM. This could be a problem if an overly large certificate
5155 or CRL is printed out from an untrusted source. TLS is not affected because
5156 record limits will reject an oversized certificate before it is parsed.
5158 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5163 * OOB read in TS_OBJ_print_bio()
5165 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5166 the total length the OID text representation would use and not the amount
5167 of data written. This will result in OOB reads when large OIDs are
5170 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5175 * Pointer arithmetic undefined behaviour
5177 Avoid some undefined pointer arithmetic
5179 A common idiom in the codebase is to check limits in the following manner:
5182 Where "p" points to some malloc'd data of SIZE bytes and
5185 "len" here could be from some externally supplied data (e.g. from a TLS
5188 The rules of C pointer arithmetic are such that "p + len" is only well
5189 defined where len <= SIZE. Therefore the above idiom is actually
5190 undefined behaviour.
5192 For example this could cause problems if some malloc implementation
5193 provides an address for "p" such that "p + len" actually overflows for
5194 values of len that are too big and therefore p + len < limit.
5196 This issue was reported to OpenSSL by Guido Vranken
5201 * Constant time flag not preserved in DSA signing
5203 Operations in the DSA signing algorithm should run in constant time in
5204 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5205 implementation means that a non-constant time codepath is followed for
5206 certain operations. This has been demonstrated through a cache-timing
5207 attack to be sufficient for an attacker to recover the private DSA key.
5209 This issue was reported by César Pereida (Aalto University), Billy Brumley
5210 (Tampere University of Technology), and Yuval Yarom (The University of
5211 Adelaide and NICTA).
5216 * DTLS buffered message DoS
5218 In a DTLS connection where handshake messages are delivered out-of-order
5219 those messages that OpenSSL is not yet ready to process will be buffered
5220 for later use. Under certain circumstances, a flaw in the logic means that
5221 those messages do not get removed from the buffer even though the handshake
5222 has been completed. An attacker could force up to approx. 15 messages to
5223 remain in the buffer when they are no longer required. These messages will
5224 be cleared when the DTLS connection is closed. The default maximum size for
5225 a message is 100k. Therefore the attacker could force an additional 1500k
5226 to be consumed per connection. By opening many simulataneous connections an
5227 attacker could cause a DoS attack through memory exhaustion.
5229 This issue was reported to OpenSSL by Quan Luo.
5234 * DTLS replay protection DoS
5236 A flaw in the DTLS replay attack protection mechanism means that records
5237 that arrive for future epochs update the replay protection "window" before
5238 the MAC for the record has been validated. This could be exploited by an
5239 attacker by sending a record for the next epoch (which does not have to
5240 decrypt or have a valid MAC), with a very large sequence number. This means
5241 that all subsequent legitimate packets are dropped causing a denial of
5242 service for a specific DTLS connection.
5244 This issue was reported to OpenSSL by the OCAP audit team.
5249 * Certificate message OOB reads
5251 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5252 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5253 theoretical DoS risk but this has not been observed in practice on common
5256 The messages affected are client certificate, client certificate request
5257 and server certificate. As a result the attack can only be performed
5258 against a client or a server which enables client authentication.
5260 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5265 ### Changes between 1.0.1s and 1.0.1t [3 May 2016] ###
5267 * Prevent padding oracle in AES-NI CBC MAC check
5269 A MITM attacker can use a padding oracle attack to decrypt traffic
5270 when the connection uses an AES CBC cipher and the server support
5273 This issue was introduced as part of the fix for Lucky 13 padding
5274 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5275 constant time by making sure that always the same bytes are read and
5276 compared against either the MAC or padding bytes. But it no longer
5277 checked that there was enough data to have both the MAC and padding
5280 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5285 * Fix EVP_EncodeUpdate overflow
5287 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5288 Base64 encoding of binary data. If an attacker is able to supply very large
5289 amounts of input data then a length check can overflow resulting in a heap
5292 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5293 the PEM_write_bio* family of functions. These are mainly used within the
5294 OpenSSL command line applications, so any application which processes data
5295 from an untrusted source and outputs it as a PEM file should be considered
5296 vulnerable to this issue. User applications that call these APIs directly
5297 with large amounts of untrusted data may also be vulnerable.
5299 This issue was reported by Guido Vranken.
5304 * Fix EVP_EncryptUpdate overflow
5306 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5307 is able to supply very large amounts of input data after a previous call to
5308 EVP_EncryptUpdate() with a partial block then a length check can overflow
5309 resulting in a heap corruption. Following an analysis of all OpenSSL
5310 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5311 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5312 the first called function after an EVP_EncryptInit(), and therefore that
5313 specific call must be safe. The second form is where the length passed to
5314 EVP_EncryptUpdate() can be seen from the code to be some small value and
5315 therefore there is no possibility of an overflow. Since all instances are
5316 one of these two forms, it is believed that there can be no overflows in
5317 internal code due to this problem. It should be noted that
5318 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5319 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5320 of these calls have also been analysed too and it is believed there are no
5321 instances in internal usage where an overflow could occur.
5323 This issue was reported by Guido Vranken.
5328 * Prevent ASN.1 BIO excessive memory allocation
5330 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5331 a short invalid encoding can casuse allocation of large amounts of memory
5332 potentially consuming excessive resources or exhausting memory.
5334 Any application parsing untrusted data through d2i BIO functions is
5335 affected. The memory based functions such as d2i_X509() are *not* affected.
5336 Since the memory based functions are used by the TLS library, TLS
5337 applications are not affected.
5339 This issue was reported by Brian Carpenter.
5346 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5347 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5348 in arbitrary stack data being returned in the buffer.
5350 This issue was reported by Guido Vranken.
5355 * Modify behavior of ALPN to invoke callback after SNI/servername
5356 callback, such that updates to the SSL_CTX affect ALPN.
5360 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5365 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5366 methods are enabled and ssl2 is disabled the methods return NULL.
5370 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016] ###
5372 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5373 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5374 provide any "EXPORT" or "LOW" strength ciphers.
5378 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5379 is by default disabled at build-time. Builds that are not configured with
5380 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5381 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5382 will need to explicitly call either of:
5384 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5386 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5388 as appropriate. Even if either of those is used, or the application
5389 explicitly uses the version-specific SSLv2_method() or its client and
5390 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5391 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5392 ciphers, and SSLv2 56-bit DES are no longer available.
5397 * Fix a double-free in DSA code
5399 A double free bug was discovered when OpenSSL parses malformed DSA private
5400 keys and could lead to a DoS attack or memory corruption for applications
5401 that receive DSA private keys from untrusted sources. This scenario is
5404 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5410 * Disable SRP fake user seed to address a server memory leak.
5412 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5414 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5415 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5416 was changed to ignore the "fake user" SRP seed, even if the seed
5419 Users should use SRP_VBASE_get1_by_user instead. Note that in
5420 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5421 also that even though configuring the SRP seed attempts to hide
5422 invalid usernames by continuing the handshake with fake
5423 credentials, this behaviour is not constant time and no strong
5424 guarantees are made that the handshake is indistinguishable from
5425 that of a valid user.
5430 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5432 In the BN_hex2bn function the number of hex digits is calculated using an
5433 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5434 large values of |i| this can result in |bn_expand| not allocating any
5435 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5436 field as NULL leading to a subsequent NULL ptr deref. For very large values
5437 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5438 In this case memory is allocated to the internal BIGNUM data field, but it
5439 is insufficiently sized leading to heap corruption. A similar issue exists
5440 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5441 is ever called by user applications with very large untrusted hex/dec data.
5442 This is anticipated to be a rare occurrence.
5444 All OpenSSL internal usage of these functions use data that is not expected
5445 to be untrusted, e.g. config file data or application command line
5446 arguments. If user developed applications generate config file data based
5447 on untrusted data then it is possible that this could also lead to security
5448 consequences. This is also anticipated to be rare.
5450 This issue was reported to OpenSSL by Guido Vranken.
5455 * Fix memory issues in BIO_*printf functions
5457 The internal |fmtstr| function used in processing a "%s" format string in
5458 the BIO_*printf functions could overflow while calculating the length of a
5459 string and cause an OOB read when printing very long strings.
5461 Additionally the internal |doapr_outch| function can attempt to write to an
5462 OOB memory location (at an offset from the NULL pointer) in the event of a
5463 memory allocation failure. In 1.0.2 and below this could be caused where
5464 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5465 could be in processing a very long "%s" format string. Memory leaks can
5468 The first issue may mask the second issue dependent on compiler behaviour.
5469 These problems could enable attacks where large amounts of untrusted data
5470 is passed to the BIO_*printf functions. If applications use these functions
5471 in this way then they could be vulnerable. OpenSSL itself uses these
5472 functions when printing out human-readable dumps of ASN.1 data. Therefore
5473 applications that print this data could be vulnerable if the data is from
5474 untrusted sources. OpenSSL command line applications could also be
5475 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5476 as command line arguments.
5478 Libssl is not considered directly vulnerable. Additionally certificates etc
5479 received via remote connections via libssl are also unlikely to be able to
5480 trigger these issues because of message size limits enforced within libssl.
5482 This issue was reported to OpenSSL Guido Vranken.
5487 * Side channel attack on modular exponentiation
5489 A side-channel attack was found which makes use of cache-bank conflicts on
5490 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5491 of RSA keys. The ability to exploit this issue is limited as it relies on
5492 an attacker who has control of code in a thread running on the same
5493 hyper-threaded core as the victim thread which is performing decryptions.
5495 This issue was reported to OpenSSL by Yuval Yarom, The University of
5496 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5497 Nadia Heninger, University of Pennsylvania with more information at
5498 http://cachebleed.info.
5503 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5504 if no keysize is specified with default_bits. This fixes an
5505 omission in an earlier change that changed all RSA/DSA key generation
5506 apps to use 2048 bits by default.
5510 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016] ###
5512 * Protection for DH small subgroup attacks
5514 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5515 switched on by default and cannot be disabled. This could have some
5520 * SSLv2 doesn't block disabled ciphers
5522 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5523 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5524 been disabled, provided that the SSLv2 protocol was not also disabled via
5527 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5528 and Sebastian Schinzel.
5533 * Reject DH handshakes with parameters shorter than 1024 bits.
5537 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015] ###
5539 * Certificate verify crash with missing PSS parameter
5541 The signature verification routines will crash with a NULL pointer
5542 dereference if presented with an ASN.1 signature using the RSA PSS
5543 algorithm and absent mask generation function parameter. Since these
5544 routines are used to verify certificate signature algorithms this can be
5545 used to crash any certificate verification operation and exploited in a
5546 DoS attack. Any application which performs certificate verification is
5547 vulnerable including OpenSSL clients and servers which enable client
5550 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5555 * X509_ATTRIBUTE memory leak
5557 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5558 memory. This structure is used by the PKCS#7 and CMS routines so any
5559 application which reads PKCS#7 or CMS data from untrusted sources is
5560 affected. SSL/TLS is not affected.
5562 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5568 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5569 This changes the decoding behaviour for some invalid messages,
5570 though the change is mostly in the more lenient direction, and
5571 legacy behaviour is preserved as much as possible.
5575 * In DSA_generate_parameters_ex, if the provided seed is too short,
5576 use a random seed, as already documented.
5578 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5580 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015] ###
5582 * Alternate chains certificate forgery
5584 During certificate verfification, OpenSSL will attempt to find an
5585 alternative certificate chain if the first attempt to build such a chain
5586 fails. An error in the implementation of this logic can mean that an
5587 attacker could cause certain checks on untrusted certificates to be
5588 bypassed, such as the CA flag, enabling them to use a valid leaf
5589 certificate to act as a CA and "issue" an invalid certificate.
5591 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5597 * Race condition handling PSK identify hint
5599 If PSK identity hints are received by a multi-threaded client then
5600 the values are wrongly updated in the parent SSL_CTX structure. This can
5601 result in a race condition potentially leading to a double free of the
5607 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015] ###
5608 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5609 incompatibility in the handling of HMAC. The previous ABI has now been
5612 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015] ###
5614 * Malformed ECParameters causes infinite loop
5616 When processing an ECParameters structure OpenSSL enters an infinite loop
5617 if the curve specified is over a specially malformed binary polynomial
5620 This can be used to perform denial of service against any
5621 system which processes public keys, certificate requests or
5622 certificates. This includes TLS clients and TLS servers with
5623 client authentication enabled.
5625 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5630 * Exploitable out-of-bounds read in X509_cmp_time
5632 X509_cmp_time does not properly check the length of the ASN1_TIME
5633 string and can read a few bytes out of bounds. In addition,
5634 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5637 An attacker can use this to craft malformed certificates and CRLs of
5638 various sizes and potentially cause a segmentation fault, resulting in
5639 a DoS on applications that verify certificates or CRLs. TLS clients
5640 that verify CRLs are affected. TLS clients and servers with client
5641 authentication enabled may be affected if they use custom verification
5644 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5645 independently by Hanno Böck.
5650 * PKCS7 crash with missing EnvelopedContent
5652 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5653 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5654 with missing content and trigger a NULL pointer dereference on parsing.
5656 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5657 structures from untrusted sources are affected. OpenSSL clients and
5658 servers are not affected.
5660 This issue was reported to OpenSSL by Michal Zalewski (Google).
5665 * CMS verify infinite loop with unknown hash function
5667 When verifying a signedData message the CMS code can enter an infinite loop
5668 if presented with an unknown hash function OID. This can be used to perform
5669 denial of service against any system which verifies signedData messages using
5671 This issue was reported to OpenSSL by Johannes Bauer.
5676 * Race condition handling NewSessionTicket
5678 If a NewSessionTicket is received by a multi-threaded client when attempting to
5679 reuse a previous ticket then a race condition can occur potentially leading to
5680 a double free of the ticket data.
5685 * Reject DH handshakes with parameters shorter than 768 bits.
5687 *Kurt Roeckx and Emilia Kasper*
5689 * dhparam: generate 2048-bit parameters by default.
5691 *Kurt Roeckx and Emilia Kasper*
5693 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015] ###
5695 * Segmentation fault in ASN1_TYPE_cmp fix
5697 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5698 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5699 certificate signature algorithm consistency this can be used to crash any
5700 certificate verification operation and exploited in a DoS attack. Any
5701 application which performs certificate verification is vulnerable including
5702 OpenSSL clients and servers which enable client authentication.
5707 * ASN.1 structure reuse memory corruption fix
5709 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5710 memory corruption via an invalid write. Such reuse is and has been
5711 strongly discouraged and is believed to be rare.
5713 Applications that parse structures containing CHOICE or ANY DEFINED BY
5714 components may be affected. Certificate parsing (d2i_X509 and related
5715 functions) are however not affected. OpenSSL clients and servers are
5721 * PKCS7 NULL pointer dereferences fix
5723 The PKCS#7 parsing code does not handle missing outer ContentInfo
5724 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5725 missing content and trigger a NULL pointer dereference on parsing.
5727 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5728 otherwise parse PKCS#7 structures from untrusted sources are
5729 affected. OpenSSL clients and servers are not affected.
5731 This issue was reported to OpenSSL by Michal Zalewski (Google).
5736 * DoS via reachable assert in SSLv2 servers fix
5738 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5739 servers that both support SSLv2 and enable export cipher suites by sending
5740 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5742 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5743 (OpenSSL development team).
5748 * Use After Free following d2i_ECPrivatekey error fix
5750 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5751 could cause a use after free condition. This, in turn, could cause a double
5752 free in several private key parsing functions (such as d2i_PrivateKey
5753 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5754 for applications that receive EC private keys from untrusted
5755 sources. This scenario is considered rare.
5757 This issue was discovered by the BoringSSL project and fixed in their
5763 * X509_to_X509_REQ NULL pointer deref fix
5765 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5766 the certificate key is invalid. This function is rarely used in practice.
5768 This issue was discovered by Brian Carpenter.
5773 * Removed the export ciphers from the DEFAULT ciphers
5777 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015] ###
5779 * Build fixes for the Windows and OpenVMS platforms
5781 *Matt Caswell and Richard Levitte*
5783 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015] ###
5785 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5786 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5787 dereference. This could lead to a Denial Of Service attack. Thanks to
5788 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5793 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5794 dtls1_buffer_record function under certain conditions. In particular this
5795 could occur if an attacker sent repeated DTLS records with the same
5796 sequence number but for the next epoch. The memory leak could be exploited
5797 by an attacker in a Denial of Service attack through memory exhaustion.
5798 Thanks to Chris Mueller for reporting this issue.
5803 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5804 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5805 method would be set to NULL which could later result in a NULL pointer
5806 dereference. Thanks to Frank Schmirler for reporting this issue.
5811 * Abort handshake if server key exchange message is omitted for ephemeral
5814 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5815 reporting this issue.
5820 * Remove non-export ephemeral RSA code on client and server. This code
5821 violated the TLS standard by allowing the use of temporary RSA keys in
5822 non-export ciphersuites and could be used by a server to effectively
5823 downgrade the RSA key length used to a value smaller than the server
5824 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
5825 INRIA or reporting this issue.
5830 * Fixed issue where DH client certificates are accepted without verification.
5831 An OpenSSL server will accept a DH certificate for client authentication
5832 without the certificate verify message. This effectively allows a client to
5833 authenticate without the use of a private key. This only affects servers
5834 which trust a client certificate authority which issues certificates
5835 containing DH keys: these are extremely rare and hardly ever encountered.
5836 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
5842 * Ensure that the session ID context of an SSL is updated when its
5843 SSL_CTX is updated via SSL_set_SSL_CTX.
5845 The session ID context is typically set from the parent SSL_CTX,
5846 and can vary with the CTX.
5850 * Fix various certificate fingerprint issues.
5852 By using non-DER or invalid encodings outside the signed portion of a
5853 certificate the fingerprint can be changed without breaking the signature.
5854 Although no details of the signed portion of the certificate can be changed
5855 this can cause problems with some applications: e.g. those using the
5856 certificate fingerprint for blacklists.
5858 1. Reject signatures with non zero unused bits.
5860 If the BIT STRING containing the signature has non zero unused bits reject
5861 the signature. All current signature algorithms require zero unused bits.
5863 2. Check certificate algorithm consistency.
5865 Check the AlgorithmIdentifier inside TBS matches the one in the
5866 certificate signature. NB: this will result in signature failure
5867 errors for some broken certificates.
5869 Thanks to Konrad Kraszewski from Google for reporting this issue.
5871 3. Check DSA/ECDSA signatures use DER.
5873 Re-encode DSA/ECDSA signatures and compare with the original received
5874 signature. Return an error if there is a mismatch.
5876 This will reject various cases including garbage after signature
5877 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
5878 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
5879 (negative or with leading zeroes).
5881 Further analysis was conducted and fixes were developed by Stephen Henson
5882 of the OpenSSL core team.
5888 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
5889 results on some platforms, including x86_64. This bug occurs at random
5890 with a very low probability, and is not known to be exploitable in any
5891 way, though its exact impact is difficult to determine. Thanks to Pieter
5892 Wuille (Blockstream) who reported this issue and also suggested an initial
5893 fix. Further analysis was conducted by the OpenSSL development team and
5894 Adam Langley of Google. The final fix was developed by Andy Polyakov of
5895 the OpenSSL core team.
5900 *) Do not resume sessions on the server if the negotiated protocol
5901 version does not match the session's version. Resuming with a different
5902 version, while not strictly forbidden by the RFC, is of questionable
5903 sanity and breaks all known clients.
5905 *David Benjamin, Emilia Käsper*
5907 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
5908 early CCS messages during renegotiation. (Note that because
5909 renegotiation is encrypted, this early CCS was not exploitable.)
5913 *) Tighten client-side session ticket handling during renegotiation:
5914 ensure that the client only accepts a session ticket if the server sends
5915 the extension anew in the ServerHello. Previously, a TLS client would
5916 reuse the old extension state and thus accept a session ticket if one was
5917 announced in the initial ServerHello.
5919 Similarly, ensure that the client requires a session ticket if one
5920 was advertised in the ServerHello. Previously, a TLS client would
5921 ignore a missing NewSessionTicket message.
5925 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014] ###
5929 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
5930 sends a carefully crafted handshake message, to cause OpenSSL to fail
5931 to free up to 64k of memory causing a memory leak. This could be
5932 exploited in a Denial Of Service attack. This issue affects OpenSSL
5933 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
5934 whether SRTP is used or configured. Implementations of OpenSSL that
5935 have been compiled with OPENSSL_NO_SRTP defined are not affected.
5937 The fix was developed by the OpenSSL team.
5942 * Session Ticket Memory Leak.
5944 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
5945 integrity of that ticket is first verified. In the event of a session
5946 ticket integrity check failing, OpenSSL will fail to free memory
5947 causing a memory leak. By sending a large number of invalid session
5948 tickets an attacker could exploit this issue in a Denial Of Service
5954 * Build option no-ssl3 is incomplete.
5956 When OpenSSL is configured with "no-ssl3" as a build option, servers
5957 could accept and complete a SSL 3.0 handshake, and clients could be
5958 configured to send them.
5961 *Akamai and the OpenSSL team*
5963 * Add support for TLS_FALLBACK_SCSV.
5964 Client applications doing fallback retries should call
5965 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
5968 *Adam Langley, Bodo Moeller*
5970 * Add additional DigestInfo checks.
5972 Re-encode DigestInto in DER and check against the original when
5973 verifying RSA signature: this will reject any improperly encoded
5974 DigestInfo structures.
5976 Note: this is a precautionary measure and no attacks are currently known.
5981 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014] ###
5983 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
5984 SRP code can be overrun an internal buffer. Add sanity check that
5985 g, A, B < N to SRP code.
5987 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
5988 Group for discovering this issue.
5993 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
5994 TLS 1.0 instead of higher protocol versions when the ClientHello message
5995 is badly fragmented. This allows a man-in-the-middle attacker to force a
5996 downgrade to TLS 1.0 even if both the server and the client support a
5997 higher protocol version, by modifying the client's TLS records.
5999 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6000 researching this issue.
6005 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6006 to a denial of service attack. A malicious server can crash the client
6007 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6008 ciphersuite and sending carefully crafted handshake messages.
6010 Thanks to Felix Gröbert (Google) for discovering and researching this
6016 * By sending carefully crafted DTLS packets an attacker could cause openssl
6017 to leak memory. This can be exploited through a Denial of Service attack.
6018 Thanks to Adam Langley for discovering and researching this issue.
6023 * An attacker can force openssl to consume large amounts of memory whilst
6024 processing DTLS handshake messages. This can be exploited through a
6025 Denial of Service attack.
6026 Thanks to Adam Langley for discovering and researching this issue.
6031 * An attacker can force an error condition which causes openssl to crash
6032 whilst processing DTLS packets due to memory being freed twice. This
6033 can be exploited through a Denial of Service attack.
6034 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6040 * If a multithreaded client connects to a malicious server using a resumed
6041 session and the server sends an ec point format extension it could write
6042 up to 255 bytes to freed memory.
6044 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6050 * A malicious server can crash an OpenSSL client with a null pointer
6051 dereference (read) by specifying an SRP ciphersuite even though it was not
6052 properly negotiated with the client. This can be exploited through a
6053 Denial of Service attack.
6055 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6056 discovering and researching this issue.
6061 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6062 X509_name_oneline, X509_name_print_ex et al. to leak some information
6063 from the stack. Applications may be affected if they echo pretty printing
6064 output to the attacker.
6066 Thanks to Ivan Fratric (Google) for discovering this issue.
6069 *Emilia Käsper, and Steve Henson*
6071 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6072 for corner cases. (Certain input points at infinity could lead to
6073 bogus results, with non-infinity inputs mapped to infinity too.)
6077 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014] ###
6079 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6080 handshake can force the use of weak keying material in OpenSSL
6081 SSL/TLS clients and servers.
6083 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6084 researching this issue. [CVE-2014-0224][]
6086 *KIKUCHI Masashi, Steve Henson*
6088 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6089 OpenSSL DTLS client the code can be made to recurse eventually crashing
6092 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6095 *Imre Rad, Steve Henson*
6097 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6098 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6099 client or server. This is potentially exploitable to run arbitrary
6100 code on a vulnerable client or server.
6102 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6104 *Jüri Aedla, Steve Henson*
6106 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6107 are subject to a denial of service attack.
6109 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6110 this issue. [CVE-2014-3470][]
6112 *Felix Gröbert, Ivan Fratric, Steve Henson*
6114 * Harmonize version and its documentation. -f flag is used to display
6117 *mancha <mancha1@zoho.com>*
6119 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6120 in i2d_ECPrivateKey.
6122 *mancha <mancha1@zoho.com>*
6124 * Fix some double frees. These are not thought to be exploitable.
6126 *mancha <mancha1@zoho.com>*
6128 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014] ###
6130 * A missing bounds check in the handling of the TLS heartbeat extension
6131 can be used to reveal up to 64k of memory to a connected client or
6134 Thanks for Neel Mehta of Google Security for discovering this bug and to
6135 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6136 preparing the fix [CVE-2014-0160][]
6138 *Adam Langley, Bodo Moeller*
6140 * Fix for the attack described in the paper "Recovering OpenSSL
6141 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6142 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6143 http://eprint.iacr.org/2014/140
6145 Thanks to Yuval Yarom and Naomi Benger for discovering this
6146 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6148 *Yuval Yarom and Naomi Benger*
6150 * TLS pad extension: draft-agl-tls-padding-03
6152 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6153 TLS client Hello record length value would otherwise be > 255 and
6154 less that 512 pad with a dummy extension containing zeroes so it
6155 is at least 512 bytes long.
6158 *Adam Langley, Steve Henson*
6160 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014] ###
6162 * Fix for TLS record tampering bug. A carefully crafted invalid
6163 handshake could crash OpenSSL with a NULL pointer exception.
6164 Thanks to Anton Johansson for reporting this issues.
6167 * Keep original DTLS digest and encryption contexts in retransmission
6168 structures so we can use the previous session parameters if they need
6169 to be resent. [CVE-2013-6450][]
6173 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6174 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6175 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6176 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6177 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6178 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6180 *Rob Stradling, Adam Langley*
6182 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013] ###
6184 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6185 supporting platforms or when small records were transferred.
6187 *Andy Polyakov, Steve Henson*
6189 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013] ###
6191 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6193 This addresses the flaw in CBC record processing discovered by
6194 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6195 at: http://www.isg.rhul.ac.uk/tls/
6197 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6198 Security Group at Royal Holloway, University of London
6199 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6200 Emilia Käsper for the initial patch.
6203 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6205 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6206 ciphersuites which can be exploited in a denial of service attack.
6207 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6208 and detecting this bug and to Wolfgang Ettlinger
6209 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6214 * Return an error when checking OCSP signatures when key is NULL.
6215 This fixes a DoS attack. [CVE-2013-0166][]
6219 * Make openssl verify return errors.
6221 *Chris Palmer <palmer@google.com> and Ben Laurie*
6223 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6224 the right response is stapled. Also change SSL_get_certificate()
6225 so it returns the certificate actually sent.
6226 See http://rt.openssl.org/Ticket/Display.html?id=2836.
6228 *Rob Stradling <rob.stradling@comodo.com>*
6230 * Fix possible deadlock when decoding public keys.
6234 * Don't use TLS 1.0 record version number in initial client hello
6239 ### Changes between 1.0.1b and 1.0.1c [10 May 2012] ###
6241 * Sanity check record length before skipping explicit IV in TLS
6242 1.2, 1.1 and DTLS to fix DoS attack.
6244 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6245 fuzzing as a service testing platform.
6250 * Initialise tkeylen properly when encrypting CMS messages.
6251 Thanks to Solar Designer of Openwall for reporting this issue.
6255 * In FIPS mode don't try to use composite ciphers as they are not
6260 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012] ###
6262 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6263 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6264 mean any application compiled against OpenSSL 1.0.0 headers setting
6265 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6266 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6267 0x10000000L Any application which was previously compiled against
6268 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6269 will need to be recompiled as a result. Letting be results in
6270 inability to disable specifically TLS 1.1 and in client context,
6271 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6275 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6276 disable just protocol X, but all protocols above X *if* there are
6277 protocols *below* X still enabled. In more practical terms it means
6278 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6279 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6280 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6285 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012] ###
6287 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6288 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6289 in CRYPTO_realloc_clean.
6291 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6292 issue and to Adam Langley <agl@chromium.org> for fixing it.
6295 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6297 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6301 * Workarounds for some broken servers that "hang" if a client hello
6302 record length exceeds 255 bytes.
6304 1. Do not use record version number > TLS 1.0 in initial client
6305 hello: some (but not all) hanging servers will now work.
6306 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6307 the number of ciphers sent in the client hello. This should be
6308 set to an even number, such as 50, for example by passing:
6309 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6310 Most broken servers should now work.
6311 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6312 TLS 1.2 client support entirely.
6316 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6320 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012] ###
6322 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6323 STRING form instead of a DigestInfo.
6327 * The format used for MDC2 RSA signatures is inconsistent between EVP
6328 and the RSA_sign/RSA_verify functions. This was made more apparent when
6329 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6330 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6331 the correct format in RSA_verify so both forms transparently work.
6335 * Some servers which support TLS 1.0 can choke if we initially indicate
6336 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6337 encrypted premaster secret. As a workaround use the maximum permitted
6338 client version in client hello, this should keep such servers happy
6339 and still work with previous versions of OpenSSL.
6343 * Add support for TLS/DTLS heartbeats.
6345 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6347 * Add support for SCTP.
6349 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6351 * Improved PRNG seeding for VOS.
6353 *Paul Green <Paul.Green@stratus.com>*
6355 * Extensive assembler packs updates, most notably:
6357 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6358 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6359 - x86_64: bit-sliced AES implementation;
6360 - ARM: NEON support, contemporary platforms optimizations;
6361 - s390x: z196 support;
6362 - *: GHASH and GF(2^m) multiplication implementations;
6366 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6367 (removal of unnecessary code)
6369 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6371 * Add TLS key material exporter from RFC 5705.
6375 * Add DTLS-SRTP negotiation from RFC 5764.
6379 * Add Next Protocol Negotiation,
6380 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
6381 disabled with a no-npn flag to config or Configure. Code donated
6384 *Adam Langley <agl@google.com> and Ben Laurie*
6386 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6387 NIST-P256, NIST-P521, with constant-time single point multiplication on
6388 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6389 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6390 Code made available under Apache License version 2.0.
6392 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6393 line to include this in your build of OpenSSL, and run "make depend" (or
6394 "make update"). This enables the following EC_METHODs:
6396 EC_GFp_nistp224_method()
6397 EC_GFp_nistp256_method()
6398 EC_GFp_nistp521_method()
6400 EC_GROUP_new_by_curve_name() will automatically use these (while
6401 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6404 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6406 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6407 all platforms. Move ssize_t definition from e_os.h to the public
6408 header file e_os2.h as it now appears in public header file cms.h
6412 * New -sigopt option to the ca, req and x509 utilities. Additional
6413 signature parameters can be passed using this option and in
6418 * Add RSA PSS signing function. This will generate and set the
6419 appropriate AlgorithmIdentifiers for PSS based on those in the
6420 corresponding EVP_MD_CTX structure. No application support yet.
6424 * Support for companion algorithm specific ASN1 signing routines.
6425 New function ASN1_item_sign_ctx() signs a pre-initialised
6426 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6427 the appropriate parameters.
6431 * Add new algorithm specific ASN1 verification initialisation function
6432 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6433 handling will be the same no matter what EVP_PKEY_METHOD is used.
6434 Add a PSS handler to support verification of PSS signatures: checked
6435 against a number of sample certificates.
6439 * Add signature printing for PSS. Add PSS OIDs.
6441 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6443 * Add algorithm specific signature printing. An individual ASN1 method
6444 can now print out signatures instead of the standard hex dump.
6446 More complex signatures (e.g. PSS) can print out more meaningful
6447 information. Include DSA version that prints out the signature
6452 * Password based recipient info support for CMS library: implementing
6457 * Split password based encryption into PBES2 and PBKDF2 functions. This
6458 neatly separates the code into cipher and PBE sections and is required
6459 for some algorithms that split PBES2 into separate pieces (such as
6460 password based CMS).
6464 * Session-handling fixes:
6465 - Fix handling of connections that are resuming with a session ID,
6466 but also support Session Tickets.
6467 - Fix a bug that suppressed issuing of a new ticket if the client
6468 presented a ticket with an expired session.
6469 - Try to set the ticket lifetime hint to something reasonable.
6470 - Make tickets shorter by excluding irrelevant information.
6471 - On the client side, don't ignore renewed tickets.
6473 *Adam Langley, Bodo Moeller (Google)*
6475 * Fix PSK session representation.
6479 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6481 This work was sponsored by Intel.
6485 * Add GCM support to TLS library. Some custom code is needed to split
6486 the IV between the fixed (from PRF) and explicit (from TLS record)
6487 portions. This adds all GCM ciphersuites supported by RFC5288 and
6488 RFC5289. Generalise some AES* cipherstrings to include GCM and
6489 add a special AESGCM string for GCM only.
6493 * Expand range of ctrls for AES GCM. Permit setting invocation
6494 field on decrypt and retrieval of invocation field only on encrypt.
6498 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6499 As required by RFC5289 these ciphersuites cannot be used if for
6500 versions of TLS earlier than 1.2.
6504 * For FIPS capable OpenSSL interpret a NULL default public key method
6505 as unset and return the appropriate default but do *not* set the default.
6506 This means we can return the appropriate method in applications that
6507 switch between FIPS and non-FIPS modes.
6511 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6512 ENGINE is used then we cannot handle that in the FIPS module so we
6513 keep original code iff non-FIPS operations are allowed.
6517 * Add -attime option to openssl utilities.
6519 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6521 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6525 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6526 FIPS EC methods unconditionally for now.
6530 * New build option no-ec2m to disable characteristic 2 code.
6534 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6535 all cases can be covered as some introduce binary incompatibilities.
6539 * Redirect RSA operations to FIPS module including keygen,
6540 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6544 * Add similar low level API blocking to ciphers.
6548 * Low level digest APIs are not approved in FIPS mode: any attempt
6549 to use these will cause a fatal error. Applications that *really* want
6550 to use them can use the private_* version instead.
6554 * Redirect cipher operations to FIPS module for FIPS builds.
6558 * Redirect digest operations to FIPS module for FIPS builds.
6562 * Update build system to add "fips" flag which will link in fipscanister.o
6563 for static and shared library builds embedding a signature if needed.
6567 * Output TLS supported curves in preference order instead of numerical
6568 order. This is currently hardcoded for the highest order curves first.
6569 This should be configurable so applications can judge speed vs strength.
6573 * Add TLS v1.2 server support for client authentication.
6577 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6582 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6583 FIPS modules versions.
6587 * Add TLS v1.2 client side support for client authentication. Keep cache
6588 of handshake records longer as we don't know the hash algorithm to use
6589 until after the certificate request message is received.
6593 * Initial TLS v1.2 client support. Add a default signature algorithms
6594 extension including all the algorithms we support. Parse new signature
6595 format in client key exchange. Relax some ECC signing restrictions for
6596 TLS v1.2 as indicated in RFC5246.
6600 * Add server support for TLS v1.2 signature algorithms extension. Switch
6601 to new signature format when needed using client digest preference.
6602 All server ciphersuites should now work correctly in TLS v1.2. No client
6603 support yet and no support for client certificates.
6607 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6608 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6609 ciphersuites. At present only RSA key exchange ciphersuites work with
6610 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6611 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6612 and version checking.
6616 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6617 with this defined it will not be affected by any changes to ssl internal
6618 structures. Add several utility functions to allow openssl application
6619 to work with OPENSSL_NO_SSL_INTERN defined.
6623 * A long standing patch to add support for SRP from EdelWeb (Peter
6624 Sylvester and Christophe Renou) was integrated.
6625 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6626 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6629 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6633 * Permit abbreviated handshakes when renegotiating using the function
6634 SSL_renegotiate_abbreviated().
6636 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6638 * Add call to ENGINE_register_all_complete() to
6639 ENGINE_load_builtin_engines(), so some implementations get used
6640 automatically instead of needing explicit application support.
6644 * Add support for TLS key exporter as described in RFC5705.
6646 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6648 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6649 a few changes are required:
6651 Add SSL_OP_NO_TLSv1_1 flag.
6652 Add TLSv1_1 methods.
6653 Update version checking logic to handle version 1.1.
6654 Add explicit IV handling (ported from DTLS code).
6655 Add command line options to s_client/s_server.
6662 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015] ###
6664 * X509_ATTRIBUTE memory leak
6666 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6667 memory. This structure is used by the PKCS#7 and CMS routines so any
6668 application which reads PKCS#7 or CMS data from untrusted sources is
6669 affected. SSL/TLS is not affected.
6671 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6677 * Race condition handling PSK identify hint
6679 If PSK identity hints are received by a multi-threaded client then
6680 the values are wrongly updated in the parent SSL_CTX structure. This can
6681 result in a race condition potentially leading to a double free of the
6687 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015] ###
6689 * Malformed ECParameters causes infinite loop
6691 When processing an ECParameters structure OpenSSL enters an infinite loop
6692 if the curve specified is over a specially malformed binary polynomial
6695 This can be used to perform denial of service against any
6696 system which processes public keys, certificate requests or
6697 certificates. This includes TLS clients and TLS servers with
6698 client authentication enabled.
6700 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6705 * Exploitable out-of-bounds read in X509_cmp_time
6707 X509_cmp_time does not properly check the length of the ASN1_TIME
6708 string and can read a few bytes out of bounds. In addition,
6709 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6712 An attacker can use this to craft malformed certificates and CRLs of
6713 various sizes and potentially cause a segmentation fault, resulting in
6714 a DoS on applications that verify certificates or CRLs. TLS clients
6715 that verify CRLs are affected. TLS clients and servers with client
6716 authentication enabled may be affected if they use custom verification
6719 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6720 independently by Hanno Böck.
6725 * PKCS7 crash with missing EnvelopedContent
6727 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6728 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6729 with missing content and trigger a NULL pointer dereference on parsing.
6731 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6732 structures from untrusted sources are affected. OpenSSL clients and
6733 servers are not affected.
6735 This issue was reported to OpenSSL by Michal Zalewski (Google).
6740 * CMS verify infinite loop with unknown hash function
6742 When verifying a signedData message the CMS code can enter an infinite loop
6743 if presented with an unknown hash function OID. This can be used to perform
6744 denial of service against any system which verifies signedData messages using
6746 This issue was reported to OpenSSL by Johannes Bauer.
6751 * Race condition handling NewSessionTicket
6753 If a NewSessionTicket is received by a multi-threaded client when attempting to
6754 reuse a previous ticket then a race condition can occur potentially leading to
6755 a double free of the ticket data.
6760 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015] ###
6762 * Segmentation fault in ASN1_TYPE_cmp fix
6764 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6765 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6766 certificate signature algorithm consistency this can be used to crash any
6767 certificate verification operation and exploited in a DoS attack. Any
6768 application which performs certificate verification is vulnerable including
6769 OpenSSL clients and servers which enable client authentication.
6774 * ASN.1 structure reuse memory corruption fix
6776 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6777 memory corruption via an invalid write. Such reuse is and has been
6778 strongly discouraged and is believed to be rare.
6780 Applications that parse structures containing CHOICE or ANY DEFINED BY
6781 components may be affected. Certificate parsing (d2i_X509 and related
6782 functions) are however not affected. OpenSSL clients and servers are
6788 * PKCS7 NULL pointer dereferences fix
6790 The PKCS#7 parsing code does not handle missing outer ContentInfo
6791 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6792 missing content and trigger a NULL pointer dereference on parsing.
6794 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6795 otherwise parse PKCS#7 structures from untrusted sources are
6796 affected. OpenSSL clients and servers are not affected.
6798 This issue was reported to OpenSSL by Michal Zalewski (Google).
6803 * DoS via reachable assert in SSLv2 servers fix
6805 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6806 servers that both support SSLv2 and enable export cipher suites by sending
6807 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6809 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6810 (OpenSSL development team).
6815 * Use After Free following d2i_ECPrivatekey error fix
6817 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6818 could cause a use after free condition. This, in turn, could cause a double
6819 free in several private key parsing functions (such as d2i_PrivateKey
6820 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6821 for applications that receive EC private keys from untrusted
6822 sources. This scenario is considered rare.
6824 This issue was discovered by the BoringSSL project and fixed in their
6830 * X509_to_X509_REQ NULL pointer deref fix
6832 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6833 the certificate key is invalid. This function is rarely used in practice.
6835 This issue was discovered by Brian Carpenter.
6840 * Removed the export ciphers from the DEFAULT ciphers
6844 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015] ###
6846 * Build fixes for the Windows and OpenVMS platforms
6848 *Matt Caswell and Richard Levitte*
6850 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015] ###
6852 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6853 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6854 dereference. This could lead to a Denial Of Service attack. Thanks to
6855 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6860 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6861 dtls1_buffer_record function under certain conditions. In particular this
6862 could occur if an attacker sent repeated DTLS records with the same
6863 sequence number but for the next epoch. The memory leak could be exploited
6864 by an attacker in a Denial of Service attack through memory exhaustion.
6865 Thanks to Chris Mueller for reporting this issue.
6870 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6871 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6872 method would be set to NULL which could later result in a NULL pointer
6873 dereference. Thanks to Frank Schmirler for reporting this issue.
6878 * Abort handshake if server key exchange message is omitted for ephemeral
6881 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6882 reporting this issue.
6887 * Remove non-export ephemeral RSA code on client and server. This code
6888 violated the TLS standard by allowing the use of temporary RSA keys in
6889 non-export ciphersuites and could be used by a server to effectively
6890 downgrade the RSA key length used to a value smaller than the server
6891 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6892 INRIA or reporting this issue.
6897 * Fixed issue where DH client certificates are accepted without verification.
6898 An OpenSSL server will accept a DH certificate for client authentication
6899 without the certificate verify message. This effectively allows a client to
6900 authenticate without the use of a private key. This only affects servers
6901 which trust a client certificate authority which issues certificates
6902 containing DH keys: these are extremely rare and hardly ever encountered.
6903 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6909 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6910 results on some platforms, including x86_64. This bug occurs at random
6911 with a very low probability, and is not known to be exploitable in any
6912 way, though its exact impact is difficult to determine. Thanks to Pieter
6913 Wuille (Blockstream) who reported this issue and also suggested an initial
6914 fix. Further analysis was conducted by the OpenSSL development team and
6915 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6916 the OpenSSL core team.
6921 * Fix various certificate fingerprint issues.
6923 By using non-DER or invalid encodings outside the signed portion of a
6924 certificate the fingerprint can be changed without breaking the signature.
6925 Although no details of the signed portion of the certificate can be changed
6926 this can cause problems with some applications: e.g. those using the
6927 certificate fingerprint for blacklists.
6929 1. Reject signatures with non zero unused bits.
6931 If the BIT STRING containing the signature has non zero unused bits reject
6932 the signature. All current signature algorithms require zero unused bits.
6934 2. Check certificate algorithm consistency.
6936 Check the AlgorithmIdentifier inside TBS matches the one in the
6937 certificate signature. NB: this will result in signature failure
6938 errors for some broken certificates.
6940 Thanks to Konrad Kraszewski from Google for reporting this issue.
6942 3. Check DSA/ECDSA signatures use DER.
6944 Reencode DSA/ECDSA signatures and compare with the original received
6945 signature. Return an error if there is a mismatch.
6947 This will reject various cases including garbage after signature
6948 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6949 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6950 (negative or with leading zeroes).
6952 Further analysis was conducted and fixes were developed by Stephen Henson
6953 of the OpenSSL core team.
6959 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014] ###
6961 * Session Ticket Memory Leak.
6963 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6964 integrity of that ticket is first verified. In the event of a session
6965 ticket integrity check failing, OpenSSL will fail to free memory
6966 causing a memory leak. By sending a large number of invalid session
6967 tickets an attacker could exploit this issue in a Denial Of Service
6973 * Build option no-ssl3 is incomplete.
6975 When OpenSSL is configured with "no-ssl3" as a build option, servers
6976 could accept and complete a SSL 3.0 handshake, and clients could be
6977 configured to send them.
6980 *Akamai and the OpenSSL team*
6982 * Add support for TLS_FALLBACK_SCSV.
6983 Client applications doing fallback retries should call
6984 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6987 *Adam Langley, Bodo Moeller*
6989 * Add additional DigestInfo checks.
6991 Reencode DigestInto in DER and check against the original when
6992 verifying RSA signature: this will reject any improperly encoded
6993 DigestInfo structures.
6995 Note: this is a precautionary measure and no attacks are currently known.
7000 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014] ###
7002 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7003 to a denial of service attack. A malicious server can crash the client
7004 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7005 ciphersuite and sending carefully crafted handshake messages.
7007 Thanks to Felix Gröbert (Google) for discovering and researching this
7013 * By sending carefully crafted DTLS packets an attacker could cause openssl
7014 to leak memory. This can be exploited through a Denial of Service attack.
7015 Thanks to Adam Langley for discovering and researching this issue.
7020 * An attacker can force openssl to consume large amounts of memory whilst
7021 processing DTLS handshake messages. This can be exploited through a
7022 Denial of Service attack.
7023 Thanks to Adam Langley for discovering and researching this issue.
7028 * An attacker can force an error condition which causes openssl to crash
7029 whilst processing DTLS packets due to memory being freed twice. This
7030 can be exploited through a Denial of Service attack.
7031 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7037 * If a multithreaded client connects to a malicious server using a resumed
7038 session and the server sends an ec point format extension it could write
7039 up to 255 bytes to freed memory.
7041 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7047 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7048 X509_name_oneline, X509_name_print_ex et al. to leak some information
7049 from the stack. Applications may be affected if they echo pretty printing
7050 output to the attacker.
7052 Thanks to Ivan Fratric (Google) for discovering this issue.
7055 *Emilia Käsper, and Steve Henson*
7057 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7058 for corner cases. (Certain input points at infinity could lead to
7059 bogus results, with non-infinity inputs mapped to infinity too.)
7063 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014] ###
7065 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7066 handshake can force the use of weak keying material in OpenSSL
7067 SSL/TLS clients and servers.
7069 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7070 researching this issue. [CVE-2014-0224][]
7072 *KIKUCHI Masashi, Steve Henson*
7074 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7075 OpenSSL DTLS client the code can be made to recurse eventually crashing
7078 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7081 *Imre Rad, Steve Henson*
7083 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7084 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7085 client or server. This is potentially exploitable to run arbitrary
7086 code on a vulnerable client or server.
7088 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7090 *Jüri Aedla, Steve Henson*
7092 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7093 are subject to a denial of service attack.
7095 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7096 this issue. [CVE-2014-3470][]
7098 *Felix Gröbert, Ivan Fratric, Steve Henson*
7100 * Harmonize version and its documentation. -f flag is used to display
7103 *mancha <mancha1@zoho.com>*
7105 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7106 in i2d_ECPrivateKey.
7108 *mancha <mancha1@zoho.com>*
7110 * Fix some double frees. These are not thought to be exploitable.
7112 *mancha <mancha1@zoho.com>*
7114 * Fix for the attack described in the paper "Recovering OpenSSL
7115 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7116 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7117 http://eprint.iacr.org/2014/140
7119 Thanks to Yuval Yarom and Naomi Benger for discovering this
7120 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7122 *Yuval Yarom and Naomi Benger*
7124 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014] ###
7126 * Keep original DTLS digest and encryption contexts in retransmission
7127 structures so we can use the previous session parameters if they need
7128 to be resent. [CVE-2013-6450][]
7132 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7133 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7134 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7135 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7136 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7137 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7139 *Rob Stradling, Adam Langley*
7141 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013] ###
7143 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7145 This addresses the flaw in CBC record processing discovered by
7146 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7147 at: http://www.isg.rhul.ac.uk/tls/
7149 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7150 Security Group at Royal Holloway, University of London
7151 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7152 Emilia Käsper for the initial patch.
7155 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7157 * Return an error when checking OCSP signatures when key is NULL.
7158 This fixes a DoS attack. [CVE-2013-0166][]
7162 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7163 the right response is stapled. Also change SSL_get_certificate()
7164 so it returns the certificate actually sent.
7165 See http://rt.openssl.org/Ticket/Display.html?id=2836.
7166 (This is a backport)
7168 *Rob Stradling <rob.stradling@comodo.com>*
7170 * Fix possible deadlock when decoding public keys.
7174 ### Changes between 1.0.0i and 1.0.0j [10 May 2012] ###
7176 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7179 * Sanity check record length before skipping explicit IV in DTLS
7182 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7183 fuzzing as a service testing platform.
7188 * Initialise tkeylen properly when encrypting CMS messages.
7189 Thanks to Solar Designer of Openwall for reporting this issue.
7193 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012] ###
7195 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7196 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7197 in CRYPTO_realloc_clean.
7199 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7200 issue and to Adam Langley <agl@chromium.org> for fixing it.
7203 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7205 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012] ###
7207 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7208 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7209 content decryption and always return the same error. Note: this attack
7210 needs on average 2^20 messages so it only affects automated senders. The
7211 old behaviour can be re-enabled in the CMS code by setting the
7212 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7213 an MMA defence is not necessary.
7214 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7215 this issue. [CVE-2012-0884][]
7219 * Fix CVE-2011-4619: make sure we really are receiving a
7220 client hello before rejecting multiple SGC restarts. Thanks to
7221 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7225 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012] ###
7227 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7228 Thanks to Antonio Martin, Enterprise Secure Access Research and
7229 Development, Cisco Systems, Inc. for discovering this bug and
7230 preparing a fix. [CVE-2012-0050][]
7234 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012] ###
7236 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7237 of the Vaudenay padding oracle attack on CBC mode encryption
7238 which enables an efficient plaintext recovery attack against
7239 the OpenSSL implementation of DTLS. Their attack exploits timing
7240 differences arising during decryption processing. A research
7241 paper describing this attack can be found at:
7242 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
7243 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7244 Security Group at Royal Holloway, University of London
7245 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7246 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7247 for preparing the fix. [CVE-2011-4108][]
7249 *Robin Seggelmann, Michael Tuexen*
7251 * Clear bytes used for block padding of SSL 3.0 records.
7254 *Adam Langley (Google)*
7256 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7257 Kadianakis <desnacked@gmail.com> for discovering this issue and
7258 Adam Langley for preparing the fix. [CVE-2011-4619][]
7260 *Adam Langley (Google)*
7262 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7264 *Andrey Kulikov <amdeich@gmail.com>*
7266 * Prevent malformed RFC3779 data triggering an assertion failure.
7267 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7268 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7270 *Rob Austein <sra@hactrn.net>*
7272 * Improved PRNG seeding for VOS.
7274 *Paul Green <Paul.Green@stratus.com>*
7276 * Fix ssl_ciph.c set-up race.
7278 *Adam Langley (Google)*
7280 * Fix spurious failures in ecdsatest.c.
7282 *Emilia Käsper (Google)*
7284 * Fix the BIO_f_buffer() implementation (which was mixing different
7285 interpretations of the '..._len' fields).
7287 *Adam Langley (Google)*
7289 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7290 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7291 threads won't reuse the same blinding coefficients.
7293 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7294 lock to call BN_BLINDING_invert_ex, and avoids one use of
7295 BN_BLINDING_update for each BN_BLINDING structure (previously,
7296 the last update always remained unused).
7298 *Emilia Käsper (Google)*
7300 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7302 *Bob Buckholz (Google)*
7304 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011] ###
7306 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7307 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7309 *Kaspar Brand <ossl@velox.ch>*
7311 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7312 for multi-threaded use of ECDH. [CVE-2011-3210][]
7314 *Adam Langley (Google)*
7316 * Fix x509_name_ex_d2i memory leak on bad inputs.
7320 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7321 signature public key algorithm by using OID xref utilities instead.
7322 Before this you could only use some ECC ciphersuites with SHA1 only.
7326 * Add protection against ECDSA timing attacks as mentioned in the paper
7327 by Billy Bob Brumley and Nicola Tuveri, see:
7329 http://eprint.iacr.org/2011/232.pdf
7332 *Billy Bob Brumley and Nicola Tuveri*
7334 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011] ###
7336 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7338 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7340 * Fix bug in string printing code: if *any* escaping is enabled we must
7341 escape the escape character (backslash) or the resulting string is
7346 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010] ###
7348 * Disable code workaround for ancient and obsolete Netscape browsers
7349 and servers: an attacker can use it in a ciphersuite downgrade attack.
7350 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7354 * Fixed J-PAKE implementation error, originally discovered by
7355 Sebastien Martini, further info and confirmation from Stefan
7356 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7360 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010] ###
7362 * Fix extension code to avoid race conditions which can result in a buffer
7363 overrun vulnerability: resumed sessions must not be modified as they can
7364 be shared by multiple threads. CVE-2010-3864
7368 * Fix WIN32 build system to correctly link an ENGINE directory into
7373 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010] ###
7375 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7378 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7380 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010] ###
7382 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7383 context. The operation can be customised via the ctrl mechanism in
7384 case ENGINEs want to include additional functionality.
7388 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7392 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7393 output hashes compatible with older versions of OpenSSL.
7395 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7397 * Fix compression algorithm handling: if resuming a session use the
7398 compression algorithm of the resumed session instead of determining
7399 it from client hello again. Don't allow server to change algorithm.
7403 * Add load_crls() function to apps tidying load_certs() too. Add option
7404 to verify utility to allow additional CRLs to be included.
7408 * Update OCSP request code to permit adding custom headers to the request:
7409 some responders need this.
7413 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7416 *Julia Lawall <julia@diku.dk>*
7418 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7419 needlessly dereferenced structures, used obsolete functions and
7420 didn't handle all updated verify codes correctly.
7424 * Disable MD2 in the default configuration.
7428 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7429 indicate the initial BIO being pushed or popped. This makes it possible
7430 to determine whether the BIO is the one explicitly called or as a result
7431 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7432 it handles reference counts correctly and doesn't zero out the I/O bio
7433 when it is not being explicitly popped. WARNING: applications which
7434 included workarounds for the old buggy behaviour will need to be modified
7435 or they could free up already freed BIOs.
7439 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7440 renaming to all platforms (within the 0.9.8 branch, this was
7441 done conditionally on Netware platforms to avoid a name clash).
7443 *Guenter <lists@gknw.net>*
7445 * Add ECDHE and PSK support to DTLS.
7447 *Michael Tuexen <tuexen@fh-muenster.de>*
7449 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7454 * Add "missing" function EVP_MD_flags() (without this the only way to
7455 retrieve a digest flags is by accessing the structure directly. Update
7456 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
7457 or cipher is registered as in the "from" argument. Print out all
7458 registered digests in the dgst usage message instead of manually
7459 attempting to work them out.
7463 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7464 this allows the use of compression and extensions. Change default cipher
7465 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7466 by default unless an application cipher string requests it.
7470 * Alter match criteria in PKCS12_parse(). It used to try to use local
7471 key ids to find matching certificates and keys but some PKCS#12 files
7472 don't follow the (somewhat unwritten) rules and this strategy fails.
7473 Now just gather all certificates together and the first private key
7474 then look for the first certificate that matches the key.
7478 * Support use of registered digest and cipher names for dgst and cipher
7479 commands instead of having to add each one as a special case. So now
7486 openssl dgst -sha256 foo
7488 and this works for ENGINE based algorithms too.
7493 * Update Gost ENGINE to support parameter files.
7495 *Victor B. Wagner <vitus@cryptocom.ru>*
7497 * Support GeneralizedTime in ca utility.
7499 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7501 * Enhance the hash format used for certificate directory links. The new
7502 form uses the canonical encoding (meaning equivalent names will work
7503 even if they aren't identical) and uses SHA1 instead of MD5. This form
7504 is incompatible with the older format and as a result c_rehash should
7505 be used to rebuild symbolic links.
7509 * Make PKCS#8 the default write format for private keys, replacing the
7510 traditional format. This form is standardised, more secure and doesn't
7511 include an implicit MD5 dependency.
7515 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7516 committed to OpenSSL should pass this lot as a minimum.
7520 * Add session ticket override functionality for use by EAP-FAST.
7522 *Jouni Malinen <j@w1.fi>*
7524 * Modify HMAC functions to return a value. Since these can be implemented
7525 in an ENGINE errors can occur.
7529 * Type-checked OBJ_bsearch_ex.
7533 * Type-checked OBJ_bsearch. Also some constification necessitated
7534 by type-checking. Still to come: TXT_DB, bsearch(?),
7535 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7540 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7541 seconds to a tm structure directly, instead of going through OS
7542 specific date routines. This avoids any issues with OS routines such
7543 as the year 2038 bug. New *_adj() functions for ASN1 time structures
7544 and X509_time_adj_ex() to cover the extended range. The existing
7545 X509_time_adj() is still usable and will no longer have any date issues.
7549 * Delta CRL support. New use deltas option which will attempt to locate
7550 and search any appropriate delta CRLs available.
7552 This work was sponsored by Google.
7556 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7557 code and add additional score elements. Validate alternate CRL paths
7558 as part of the CRL checking and indicate a new error "CRL path validation
7559 error" in this case. Applications wanting additional details can use
7560 the verify callback and check the new "parent" field. If this is not
7561 NULL CRL path validation is taking place. Existing applications won't
7562 see this because it requires extended CRL support which is off by
7565 This work was sponsored by Google.
7569 * Support for freshest CRL extension.
7571 This work was sponsored by Google.
7575 * Initial indirect CRL support. Currently only supported in the CRLs
7576 passed directly and not via lookup. Process certificate issuer
7577 CRL entry extension and lookup CRL entries by bother issuer name
7578 and serial number. Check and process CRL issuer entry in IDP extension.
7580 This work was sponsored by Google.
7584 * Add support for distinct certificate and CRL paths. The CRL issuer
7585 certificate is validated separately in this case. Only enabled if
7586 an extended CRL support flag is set: this flag will enable additional
7587 CRL functionality in future.
7589 This work was sponsored by Google.
7593 * Add support for policy mappings extension.
7595 This work was sponsored by Google.
7599 * Fixes to pathlength constraint, self issued certificate handling,
7600 policy processing to align with RFC3280 and PKITS tests.
7602 This work was sponsored by Google.
7606 * Support for name constraints certificate extension. DN, email, DNS
7607 and URI types are currently supported.
7609 This work was sponsored by Google.
7613 * To cater for systems that provide a pointer-based thread ID rather
7614 than numeric, deprecate the current numeric thread ID mechanism and
7615 replace it with a structure and associated callback type. This
7616 mechanism allows a numeric "hash" to be extracted from a thread ID in
7617 either case, and on platforms where pointers are larger than 'long',
7618 mixing is done to help ensure the numeric 'hash' is usable even if it
7619 can't be guaranteed unique. The default mechanism is to use "&errno"
7620 as a pointer-based thread ID to distinguish between threads.
7622 Applications that want to provide their own thread IDs should now use
7623 CRYPTO_THREADID_set_callback() to register a callback that will call
7624 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7626 Note that ERR_remove_state() is now deprecated, because it is tied
7627 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7628 to free the current thread's error state should be replaced by
7629 ERR_remove_thread_state(NULL).
7631 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7632 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7633 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7634 application was previously providing a numeric thread callback that
7635 was inappropriate for distinguishing threads, then uniqueness might
7636 have been obtained with &errno that happened immediately in the
7637 intermediate development versions of OpenSSL; this is no longer the
7638 case, the numeric thread callback will now override the automatic use
7641 *Geoff Thorpe, with help from Bodo Moeller*
7643 * Initial support for different CRL issuing certificates. This covers a
7644 simple case where the self issued certificates in the chain exist and
7645 the real CRL issuer is higher in the existing chain.
7647 This work was sponsored by Google.
7651 * Removed effectively defunct crypto/store from the build.
7655 * Revamp of STACK to provide stronger type-checking. Still to come:
7656 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7657 ASN1_STRING, CONF_VALUE.
7661 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7662 RAM on SSL connections. This option can save about 34k per idle SSL.
7666 * Revamp of LHASH to provide stronger type-checking. Still to come:
7667 STACK, TXT_DB, bsearch, qsort.
7671 * Initial support for Cryptographic Message Syntax (aka CMS) based
7672 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7673 support for data, signedData, compressedData, digestedData and
7674 encryptedData, envelopedData types included. Scripts to check against
7675 RFC4134 examples draft and interop and consistency checks of many
7676 content types and variants.
7680 * Add options to enc utility to support use of zlib compression BIO.
7684 * Extend mk1mf to support importing of options and assembly language
7685 files from Configure script, currently only included in VC-WIN32.
7686 The assembly language rules can now optionally generate the source
7687 files from the associated perl scripts.
7691 * Implement remaining functionality needed to support GOST ciphersuites.
7692 Interop testing has been performed using CryptoPro implementations.
7694 *Victor B. Wagner <vitus@cryptocom.ru>*
7696 * s390x assembler pack.
7700 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7705 * Implement Opaque PRF Input TLS extension as specified in
7706 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7707 official specification yet and no extension type assignment by
7708 IANA exists, this extension (for now) will have to be explicitly
7709 enabled when building OpenSSL by providing the extension number
7710 to use. For example, specify an option
7712 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7714 to the "config" or "Configure" script to enable the extension,
7715 assuming extension number 0x9527 (which is a completely arbitrary
7716 and unofficial assignment based on the MD5 hash of the Internet
7717 Draft). Note that by doing so, you potentially lose
7718 interoperability with other TLS implementations since these might
7719 be using the same extension number for other purposes.
7721 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7722 opaque PRF input value to use in the handshake. This will create
7723 an internal copy of the length-'len' string at 'src', and will
7724 return non-zero for success.
7726 To get more control and flexibility, provide a callback function
7729 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7730 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7734 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7737 Callback function 'cb' will be called in handshakes, and is
7738 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7739 Argument 'arg' is for application purposes (the value as given to
7740 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7741 be provided to the callback function). The callback function
7742 has to return non-zero to report success: usually 1 to use opaque
7743 PRF input just if possible, or 2 to enforce use of the opaque PRF
7744 input. In the latter case, the library will abort the handshake
7745 if opaque PRF input is not successfully negotiated.
7747 Arguments 'peerinput' and 'len' given to the callback function
7748 will always be NULL and 0 in the case of a client. A server will
7749 see the client's opaque PRF input through these variables if
7750 available (NULL and 0 otherwise). Note that if the server
7751 provides an opaque PRF input, the length must be the same as the
7752 length of the client's opaque PRF input.
7754 Note that the callback function will only be called when creating
7755 a new session (session resumption can resume whatever was
7756 previously negotiated), and will not be called in SSL 2.0
7757 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7758 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7759 for applications that need to enforce opaque PRF input.
7764 * Update ssl code to support digests other than SHA1+MD5 for handshake
7768 *Victor B. Wagner <vitus@cryptocom.ru>*
7770 * Add RFC4507 support to OpenSSL. This includes the corrections in
7771 RFC4507bis. The encrypted ticket format is an encrypted encoded
7772 SSL_SESSION structure, that way new session features are automatically
7775 If a client application caches session in an SSL_SESSION structure
7776 support is transparent because tickets are now stored in the encoded
7779 The SSL_CTX structure automatically generates keys for ticket
7780 protection in servers so again support should be possible
7781 with no application modification.
7783 If a client or server wishes to disable RFC4507 support then the option
7784 SSL_OP_NO_TICKET can be set.
7786 Add a TLS extension debugging callback to allow the contents of any client
7787 or server extensions to be examined.
7789 This work was sponsored by Google.
7793 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7794 OpenSSL should now compile cleanly on gcc 4.2
7796 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7798 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7799 support including streaming MAC support: this is required for GOST
7800 ciphersuite support.
7802 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7804 * Add option -stream to use PKCS#7 streaming in smime utility. New
7805 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7806 to output in BER and PEM format.
7810 * Experimental support for use of HMAC via EVP_PKEY interface. This
7811 allows HMAC to be handled via the EVP_DigestSign*() interface. The
7812 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7813 ENGINE support for HMAC keys which are unextractable. New -mac and
7814 -macopt options to dgst utility.
7818 * New option -sigopt to dgst utility. Update dgst to use
7819 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
7820 alternative signing parameters such as X9.31 or PSS in the dgst
7825 * Change ssl_cipher_apply_rule(), the internal function that does
7826 the work each time a ciphersuite string requests enabling
7827 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7828 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7829 the order of disabled ciphersuites such that those ciphersuites
7830 that most recently went from enabled to disabled not only stay
7831 in order with respect to each other, but also have higher priority
7832 than other disabled ciphersuites the next time ciphersuites are
7835 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
7836 the same ciphersuites as with "HIGH" alone, but in a specific
7837 order where the PSK ciphersuites come first (since they are the
7838 most recently disabled ciphersuites when "HIGH" is parsed).
7840 Also, change ssl_create_cipher_list() (using this new
7841 functionality) such that between otherwise identical
7842 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
7847 * Change ssl_create_cipher_list() so that it automatically
7848 arranges the ciphersuites in reasonable order before starting
7849 to process the rule string. Thus, the definition for "DEFAULT"
7850 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
7851 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
7852 This makes it much easier to arrive at a reasonable default order
7853 in applications for which anonymous ciphers are OK (meaning
7854 that you can't actually use DEFAULT).
7856 *Bodo Moeller; suggested by Victor Duchovni*
7858 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
7859 processing) into multiple integers instead of setting
7860 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
7861 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
7862 (These masks as well as the individual bit definitions are hidden
7863 away into the non-exported interface ssl/ssl_locl.h, so this
7864 change to the definition of the SSL_CIPHER structure shouldn't
7865 affect applications.) This give us more bits for each of these
7866 categories, so there is no longer a need to coagulate AES128 and
7867 AES256 into a single algorithm bit, and to coagulate Camellia128
7868 and Camellia256 into a single algorithm bit, which has led to all
7871 Thus, among other things, the kludge introduced in 0.9.7m and
7872 0.9.8e for masking out AES256 independently of AES128 or masking
7873 out Camellia256 independently of AES256 is not needed here in 0.9.9.
7875 With the change, we also introduce new ciphersuite aliases that
7876 so far were missing: "AES128", "AES256", "CAMELLIA128", and
7881 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
7882 Use the leftmost N bytes of the signature input if the input is
7883 larger than the prime q (with N being the size in bytes of q).
7887 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
7888 it yet and it is largely untested.
7892 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
7896 * Initial incomplete changes to avoid need for function casts in OpenSSL
7897 some compilers (gcc 4.2 and later) reject their use. Safestack is
7898 reimplemented. Update ASN1 to avoid use of legacy functions.
7902 * Win32/64 targets are linked with Winsock2.
7906 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
7907 to external functions. This can be used to increase CRL handling
7908 efficiency especially when CRLs are very large by (for example) storing
7909 the CRL revoked certificates in a database.
7913 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
7914 new CRLs added to a directory can be used. New command line option
7915 -verify_return_error to s_client and s_server. This causes real errors
7916 to be returned by the verify callback instead of carrying on no matter
7917 what. This reflects the way a "real world" verify callback would behave.
7921 * GOST engine, supporting several GOST algorithms and public key formats.
7922 Kindly donated by Cryptocom.
7926 * Partial support for Issuing Distribution Point CRL extension. CRLs
7927 partitioned by DP are handled but no indirect CRL or reason partitioning
7928 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
7929 selected via a scoring technique which handles IDP and AKID in CRLs.
7933 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
7934 will ultimately be used for all verify operations: this will remove the
7935 X509_STORE dependency on certificate verification and allow alternative
7936 lookup methods. X509_STORE based implementations of these two callbacks.
7940 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
7941 Modify get_crl() to find a valid (unexpired) CRL if possible.
7945 * New function X509_CRL_match() to check if two CRLs are identical. Normally
7946 this would be called X509_CRL_cmp() but that name is already used by
7947 a function that just compares CRL issuer names. Cache several CRL
7948 extensions in X509_CRL structure and cache CRLDP in X509.
7952 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
7953 this maps equivalent X509_NAME structures into a consistent structure.
7954 Name comparison can then be performed rapidly using memcmp().
7958 * Non-blocking OCSP request processing. Add -timeout option to ocsp
7963 * Allow digests to supply their own micalg string for S/MIME type using
7964 the ctrl EVP_MD_CTRL_MICALG.
7968 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
7969 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
7970 ctrl. It can then customise the structure before and/or after signing
7975 * New function OBJ_add_sigid() to allow application defined signature OIDs
7976 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
7977 to free up any added signature OIDs.
7981 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
7982 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
7983 digest and cipher tables. New options added to openssl utility:
7984 list-message-digest-algorithms and list-cipher-algorithms.
7988 * Change the array representation of binary polynomials: the list
7989 of degrees of non-zero coefficients is now terminated with -1.
7990 Previously it was terminated with 0, which was also part of the
7991 value; thus, the array representation was not applicable to
7992 polynomials where t^0 has coefficient zero. This change makes
7993 the array representation useful in a more general context.
7997 * Various modifications and fixes to SSL/TLS cipher string
7998 handling. For ECC, the code now distinguishes between fixed ECDH
7999 with RSA certificates on the one hand and with ECDSA certificates
8000 on the other hand, since these are separate ciphersuites. The
8001 unused code for Fortezza ciphersuites has been removed.
8003 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8004 (not "ECDHE"). For consistency with the code for DH
8005 certificates, use of ECDH certificates is now considered ECDH
8006 authentication, not RSA or ECDSA authentication (the latter is
8007 merely the CA's signing algorithm and not actively used in the
8010 The temporary ciphersuite alias "ECCdraft" is no longer
8011 available, and ECC ciphersuites are no longer excluded from "ALL"
8012 and "DEFAULT". The following aliases now exist for RFC 4492
8013 ciphersuites, most of these by analogy with the DH case:
8015 kECDHr - ECDH cert, signed with RSA
8016 kECDHe - ECDH cert, signed with ECDSA
8017 kECDH - ECDH cert (signed with either RSA or ECDSA)
8018 kEECDH - ephemeral ECDH
8019 ECDH - ECDH cert or ephemeral ECDH
8025 AECDH - anonymous ECDH
8026 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8031 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8032 Use correct micalg parameters depending on digest(s) in signed message.
8036 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8037 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8041 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8042 an engine to register a method. Add ENGINE lookups for methods and
8043 functional reference processing.
8047 * New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
8048 EVP_{Sign,Verify}* which allow an application to customise the signature
8053 * New -resign option to smime utility. This adds one or more signers
8054 to an existing PKCS#7 signedData structure. Also -md option to use an
8055 alternative message digest algorithm for signing.
8059 * Tidy up PKCS#7 routines and add new functions to make it easier to
8060 create PKCS7 structures containing multiple signers. Update smime
8061 application to support multiple signers.
8065 * New -macalg option to pkcs12 utility to allow setting of an alternative
8070 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8071 Reorganize PBE internals to lookup from a static table using NIDs,
8072 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8073 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8074 PRF which will be automatically used with PBES2.
8078 * Replace the algorithm specific calls to generate keys in "req" with the
8083 * Update PKCS#7 enveloped data routines to use new API. This is now
8084 supported by any public key method supporting the encrypt operation. A
8085 ctrl is added to allow the public key algorithm to examine or modify
8086 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8091 * Add a ctrl to asn1 method to allow a public key algorithm to express
8092 a default digest type to use. In most cases this will be SHA1 but some
8093 algorithms (such as GOST) need to specify an alternative digest. The
8094 return value indicates how strong the preference is 1 means optional and
8095 2 is mandatory (that is it is the only supported type). Modify
8096 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8097 use the default md. Update openssl utilities to use the default digest
8098 type for signing if it is not explicitly indicated.
8102 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8103 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8104 signing method from the key type. This effectively removes the link
8105 between digests and public key types.
8109 * Add an OID cross reference table and utility functions. Its purpose is to
8110 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8111 rsaEncryption. This will allow some of the algorithm specific hackery
8112 needed to use the correct OID to be removed.
8116 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8117 structures for PKCS7_sign(). They are now set up by the relevant public
8122 * Add provisional EC pkey method with support for ECDSA and ECDH.
8126 * Add support for key derivation (agreement) in the API, DH method and
8131 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8132 public and private key formats. As a side effect these add additional
8133 command line functionality not previously available: DSA signatures can be
8134 generated and verified using pkeyutl and DH key support and generation in
8141 *Oliver Tappe <zooey@hirschkaefer.de>*
8143 * New make target "install_html_docs" installs HTML renditions of the
8146 *Oliver Tappe <zooey@hirschkaefer.de>*
8148 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8149 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8150 support key and parameter generation and add initial key generation
8151 functionality for RSA.
8155 * Add functions for main EVP_PKEY_method operations. The undocumented
8156 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
8157 EVP_PKEY_{encrypt,decrypt}_old.
8161 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8162 key API, doesn't do much yet.
8166 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8167 public key algorithms. New option to openssl utility:
8168 "list-public-key-algorithms" to print out info.
8172 * Implement the Supported Elliptic Curves Extension for
8173 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8177 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8178 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8182 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8183 utilities such as rsa, dsa, dsaparam etc except they process any key
8188 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8189 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8190 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8195 * Initial support for pluggable public key ASN1.
8196 De-spaghettify the public key ASN1 handling. Move public and private
8197 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8198 algorithm specific handling to a single module within the relevant
8199 algorithm directory. Add functions to allow (near) opaque processing
8200 of public and private key structures.
8204 * Implement the Supported Point Formats Extension for
8205 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8209 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8210 for the psk identity [hint] and the psk callback functions to the
8211 SSL_SESSION, SSL and SSL_CTX structure.
8214 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8218 SSL_CTX_use_psk_identity_hint
8219 SSL_get_psk_identity_hint
8220 SSL_get_psk_identity
8221 SSL_use_psk_identity_hint
8224 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8226 * Add RFC 3161 compliant time stamp request creation, response generation
8227 and response verification functionality.
8229 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8231 * Add initial support for TLS extensions, specifically for the server_name
8232 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8233 have new members for a host name. The SSL data structure has an
8234 additional member SSL_CTX *initial_ctx so that new sessions can be
8235 stored in that context to allow for session resumption, even after the
8236 SSL has been switched to a new SSL_CTX in reaction to a client's
8237 server_name extension.
8239 New functions (subject to change):
8241 SSL_get_servername()
8242 SSL_get_servername_type()
8245 New CTRL codes and macros (subject to change):
8247 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8248 - SSL_CTX_set_tlsext_servername_callback()
8249 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8250 - SSL_CTX_set_tlsext_servername_arg()
8251 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8253 openssl s_client has a new '-servername ...' option.
8255 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8256 '-key2 ...', '-servername_fatal' (subject to change). This allows
8257 testing the HostName extension for a specific single host name ('-cert'
8258 and '-key' remain fallbacks for handshakes without HostName
8259 negotiation). If the unrecognized_name alert has to be sent, this by
8260 default is a warning; it becomes fatal with the '-servername_fatal'
8264 *Peter Sylvester, Remy Allais, Christophe Renou*
8266 * Whirlpool hash implementation is added.
8270 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8271 bn(64,32). Because of instruction set limitations it doesn't have
8272 any negative impact on performance. This was done mostly in order
8273 to make it possible to share assembler modules, such as bn_mul_mont
8274 implementations, between 32- and 64-bit builds without hassle.
8278 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8279 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8284 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8285 dedicated Montgomery multiplication procedure, is introduced.
8286 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8287 "64-bit" performance on certain 32-bit targets.
8291 * New option SSL_OP_NO_COMP to disable use of compression selectively
8292 in SSL structures. New SSL ctrl to set maximum send fragment size.
8293 Save memory by setting the I/O buffer sizes dynamically instead of
8294 using the maximum available value.
8298 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8299 in addition to the text details.
8303 * Very, very preliminary EXPERIMENTAL support for printing of general
8304 ASN1 structures. This currently produces rather ugly output and doesn't
8305 handle several customised structures at all.
8309 * Integrated support for PVK file format and some related formats such
8310 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8311 these in the 'rsa' and 'dsa' utilities.
8315 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8319 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8320 place for the (very old) "NETSCAPE" format certificates which are now
8321 handled using new ASN1 code equivalents.
8325 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8326 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8327 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8331 * Modify CRL distribution points extension code to print out previously
8332 unsupported fields. Enhance extension setting code to allow setting of
8337 * Add print and set support for Issuing Distribution Point CRL extension.
8341 * Change 'Configure' script to enable Camellia by default.
8348 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010] ###
8350 * When rejecting SSL/TLS records due to an incorrect version number, never
8351 update s->server with a new major version number. As of
8352 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8353 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8354 the previous behavior could result in a read attempt at NULL when
8355 receiving specific incorrect SSL/TLS records once record payload
8356 protection is active. [CVE-2010-0740][]
8358 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8360 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8361 could be crashed if the relevant tables were not present (e.g. chrooted).
8363 *Tomas Hoger <thoger@redhat.com>*
8365 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010] ###
8367 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8369 *Martin Olsson, Neel Mehta*
8371 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8372 accommodate for stack sorting, always a write lock!).
8376 * On some versions of WIN32 Heap32Next is very slow. This can cause
8377 excessive delays in the RAND_poll(): over a minute. As a workaround
8378 include a time check in the inner Heap32Next loop too.
8382 * The code that handled flushing of data in SSL/TLS originally used the
8383 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8384 the problem outlined in PR#1949. The fix suggested there however can
8385 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8386 of Apache). So instead simplify the code to flush unconditionally.
8387 This should be fine since flushing with no data to flush is a no op.
8391 * Handle TLS versions 2.0 and later properly and correctly use the
8392 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8393 off ancient servers have a habit of sticking around for a while...
8397 * Modify compression code so it frees up structures without using the
8398 ex_data callbacks. This works around a problem where some applications
8399 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8400 restarting) then use compression (e.g. SSL with compression) later.
8401 This results in significant per-connection memory leaks and
8402 has caused some security issues including CVE-2008-1678 and
8407 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8408 change when encrypting or decrypting.
8412 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8413 connect and renegotiate with servers which do not support RI.
8414 Until RI is more widely deployed this option is enabled by default.
8418 * Add "missing" ssl ctrls to clear options and mode.
8422 * If client attempts to renegotiate and doesn't support RI respond with
8423 a no_renegotiation alert as required by RFC5746. Some renegotiating
8424 TLS clients will continue a connection gracefully when they receive
8425 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8426 waiting for a server hello which it will never receive. Now we treat a
8427 received no_renegotiation alert as a fatal error. This is because
8428 applications requesting a renegotiation might well expect it to succeed
8429 and would have no code in place to handle the server denying it so the
8430 only safe thing to do is to terminate the connection.
8434 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8435 peer supports secure renegotiation and 0 otherwise. Print out peer
8436 renegotiation support in s_client/s_server.
8440 * Replace the highly broken and deprecated SPKAC certification method with
8441 the updated NID creation version. This should correctly handle UTF8.
8445 * Implement RFC5746. Re-enable renegotiation but require the extension
8446 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8447 turns out to be a bad idea. It has been replaced by
8448 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8449 SSL_CTX_set_options(). This is really not recommended unless you
8450 know what you are doing.
8452 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8454 * Fixes to stateless session resumption handling. Use initial_ctx when
8455 issuing and attempting to decrypt tickets in case it has changed during
8456 servername handling. Use a non-zero length session ID when attempting
8457 stateless session resumption: this makes it possible to determine if
8458 a resumption has occurred immediately after receiving server hello
8459 (several places in OpenSSL subtly assume this) instead of later in
8464 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8465 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8466 fixes for a few places where the return code is not checked
8469 *Julia Lawall <julia@diku.dk>*
8471 * Add --strict-warnings option to Configure script to include devteam
8472 warnings in other configurations.
8476 * Add support for --libdir option and LIBDIR variable in makefiles. This
8477 makes it possible to install openssl libraries in locations which
8478 have names other than "lib", for example "/usr/lib64" which some
8481 *Steve Henson, based on patch from Jeremy Utley*
8483 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8484 X690 8.9.12 and can produce some misleading textual output of OIDs.
8486 *Steve Henson, reported by Dan Kaminsky*
8488 * Delete MD2 from algorithm tables. This follows the recommendation in
8489 several standards that it is not used in new applications due to
8490 several cryptographic weaknesses. For binary compatibility reasons
8491 the MD2 API is still compiled in by default.
8495 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8500 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8501 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8504 *Guenter <lists@gknw.net>*
8506 * Fix the server certificate chain building code to use X509_verify_cert(),
8507 it used to have an ad-hoc builder which was unable to cope with anything
8508 other than a simple chain.
8510 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8512 * Don't check self signed certificate signatures in X509_verify_cert()
8513 by default (a flag can override this): it just wastes time without
8514 adding any security. As a useful side effect self signed root CAs
8515 with non-FIPS digests are now usable in FIPS mode.
8519 * In dtls1_process_out_of_seq_message() the check if the current message
8520 is already buffered was missing. For every new message was memory
8521 allocated, allowing an attacker to perform an denial of service attack
8522 with sending out of seq handshake messages until there is no memory
8523 left. Additionally every future message was buffered, even if the
8524 sequence number made no sense and would be part of another handshake.
8525 So only messages with sequence numbers less than 10 in advance will be
8526 buffered. [CVE-2009-1378][]
8528 *Robin Seggelmann, discovered by Daniel Mentz*
8530 * Records are buffered if they arrive with a future epoch to be
8531 processed after finishing the corresponding handshake. There is
8532 currently no limitation to this buffer allowing an attacker to perform
8533 a DOS attack with sending records with future epochs until there is no
8534 memory left. This patch adds the pqueue_size() function to determine
8535 the size of a buffer and limits the record buffer to 100 entries.
8538 *Robin Seggelmann, discovered by Daniel Mentz*
8540 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8541 parent structure is freed. [CVE-2009-1379][]
8545 * Handle non-blocking I/O properly in SSL_shutdown() call.
8547 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8551 *Ilya O. <vrghost@gmail.com>*
8553 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009] ###
8555 * Disable renegotiation completely - this fixes a severe security
8556 problem [CVE-2009-3555][] at the cost of breaking all
8557 renegotiation. Renegotiation can be re-enabled by setting
8558 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8559 run-time. This is really not recommended unless you know what
8564 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009] ###
8566 * Don't set val to NULL when freeing up structures, it is freed up by
8567 underlying code. If sizeof(void *) > sizeof(long) this can result in
8568 zeroing past the valid field. [CVE-2009-0789][]
8570 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8572 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8573 checked correctly. This would allow some invalid signed attributes to
8574 appear to verify correctly. [CVE-2009-0591][]
8576 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8578 * Reject UniversalString and BMPString types with invalid lengths. This
8579 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8580 a legal length. [CVE-2009-0590][]
8584 * Set S/MIME signing as the default purpose rather than setting it
8585 unconditionally. This allows applications to override it at the store
8590 * Permit restricted recursion of ASN1 strings. This is needed in practice
8591 to handle some structures.
8595 * Improve efficiency of mem_gets: don't search whole buffer each time
8598 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8600 * New -hex option for openssl rand.
8604 * Print out UTF8String and NumericString when parsing ASN1.
8608 * Support NumericString type for name components.
8612 * Allow CC in the environment to override the automatically chosen
8613 compiler. Note that nothing is done to ensure flags work with the
8618 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009] ###
8620 * Properly check EVP_VerifyFinal() and similar return values
8623 *Ben Laurie, Bodo Moeller, Google Security Team*
8625 * Enable TLS extensions by default.
8629 * Allow the CHIL engine to be loaded, whether the application is
8630 multithreaded or not. (This does not release the developer from the
8631 obligation to set up the dynamic locking callbacks.)
8633 *Sander Temme <sander@temme.net>*
8635 * Use correct exit code if there is an error in dgst command.
8637 *Steve Henson; problem pointed out by Roland Dirlewanger*
8639 * Tweak Configure so that you need to say "experimental-jpake" to enable
8640 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8644 * Add experimental JPAKE support, including demo authentication in
8645 s_client and s_server.
8649 * Set the comparison function in v3_addr_canonize().
8651 *Rob Austein <sra@hactrn.net>*
8653 * Add support for XMPP STARTTLS in s_client.
8655 *Philip Paeps <philip@freebsd.org>*
8657 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8658 to ensure that even with this option, only ciphersuites in the
8659 server's preference list will be accepted. (Note that the option
8660 applies only when resuming a session, so the earlier behavior was
8661 just about the algorithm choice for symmetric cryptography.)
8665 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008] ###
8667 * Fix NULL pointer dereference if a DTLS server received
8668 ChangeCipherSpec as first record [CVE-2009-1386][].
8672 * Fix a state transition in s3_srvr.c and d1_srvr.c
8673 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
8677 * The fix in 0.9.8c that supposedly got rid of unsafe
8678 double-checked locking was incomplete for RSA blinding,
8679 addressing just one layer of what turns out to have been
8680 doubly unsafe triple-checked locking.
8682 So now fix this for real by retiring the MONT_HELPER macro
8683 in crypto/rsa/rsa_eay.c.
8686 *Bodo Moeller; problem pointed out by Marius Schilder*
8688 * Various precautionary measures:
8690 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8692 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8693 (NB: This would require knowledge of the secret session ticket key
8694 to exploit, in which case you'd be SOL either way.)
8696 - Change bn_nist.c so that it will properly handle input BIGNUMs
8697 outside the expected range.
8699 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8703 *Neel Mehta, Bodo Moeller*
8705 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8706 the load fails. Useful for distros.
8708 *Ben Laurie and the FreeBSD team*
8710 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8714 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8718 * Expand ENGINE to support engine supplied SSL client certificate functions.
8720 This work was sponsored by Logica.
8724 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8725 keystores. Support for SSL/TLS client authentication too.
8726 Not compiled unless enable-capieng specified to Configure.
8728 This work was sponsored by Logica.
8732 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8733 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8734 attribute creation routines such as certificate requests and PKCS#12
8739 ### Changes between 0.9.8g and 0.9.8h [28 May 2008] ###
8741 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8742 handshake which could lead to a client crash as found using the
8743 Codenomicon TLS test suite [CVE-2008-1672][]
8745 *Steve Henson, Mark Cox*
8747 * Fix double free in TLS server name extensions which could lead to
8748 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8752 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8754 Clear the error queue to ensure that error entries left from
8755 older function calls do not interfere with the correct operation.
8757 *Lutz Jaenicke, Erik de Castro Lopo*
8759 * Remove root CA certificates of commercial CAs:
8761 The OpenSSL project does not recommend any specific CA and does not
8762 have any policy with respect to including or excluding any CA.
8763 Therefore it does not make any sense to ship an arbitrary selection
8764 of root CA certificates with the OpenSSL software.
8768 * RSA OAEP patches to fix two separate invalid memory reads.
8769 The first one involves inputs when 'lzero' is greater than
8770 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8771 before the beginning of from). The second one involves inputs where
8772 the 'db' section contains nothing but zeroes (there is a one-byte
8773 invalid read after the end of 'db').
8775 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8777 * Partial backport from 0.9.9-dev:
8779 Introduce bn_mul_mont (dedicated Montgomery multiplication
8780 procedure) as a candidate for BIGNUM assembler implementation.
8781 While 0.9.9-dev uses assembler for various architectures, only
8782 x86_64 is available by default here in the 0.9.8 branch, and
8783 32-bit x86 is available through a compile-time setting.
8785 To try the 32-bit x86 assembler implementation, use Configure
8786 option "enable-montasm" (which exists only for this backport).
8788 As "enable-montasm" for 32-bit x86 disclaims code stability
8789 anyway, in this constellation we activate additional code
8790 backported from 0.9.9-dev for further performance improvements,
8791 namely BN_from_montgomery_word. (To enable this otherwise,
8792 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8795 *Andy Polyakov (backport partially by Bodo Moeller)*
8797 * Add TLS session ticket callback. This allows an application to set
8798 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8799 values. This is useful for key rollover for example where several key
8800 sets may exist with different names.
8804 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8805 This was broken until now in 0.9.8 releases, such that the only way
8806 a registered ENGINE could be used (assuming it initialises
8807 successfully on the host) was to explicitly set it as the default
8808 for the relevant algorithms. This is in contradiction with 0.9.7
8809 behaviour and the documentation. With this fix, when an ENGINE is
8810 registered into a given algorithm's table of implementations, the
8811 'uptodate' flag is reset so that auto-discovery will be used next
8812 time a new context for that algorithm attempts to select an
8815 *Ian Lister (tweaked by Geoff Thorpe)*
8817 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8818 implementation in the following ways:
8820 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8823 Lack of BER streaming support means one pass streaming processing is
8824 only supported if data is detached: setting the streaming flag is
8825 ignored for embedded content.
8827 CMS support is disabled by default and must be explicitly enabled
8828 with the enable-cms configuration option.
8832 * Update the GMP engine glue to do direct copies between BIGNUM and
8833 mpz_t when openssl and GMP use the same limb size. Otherwise the
8834 existing "conversion via a text string export" trick is still used.
8836 *Paul Sheer <paulsheer@gmail.com>*
8838 * Zlib compression BIO. This is a filter BIO which compressed and
8839 uncompresses any data passed through it.
8843 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
8844 RFC3394 compatible AES key wrapping.
8848 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
8849 sets string data without copying. X509_ALGOR_set0() and
8850 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
8851 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
8852 from an X509_ATTRIBUTE structure optionally checking it occurs only
8853 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
8858 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
8859 to get the expected BN_FLG_CONSTTIME behavior.
8861 *Bodo Moeller (Google)*
8865 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
8866 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
8867 - added some more tests to do_tests.pl
8868 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
8869 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
8870 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
8871 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
8872 - various changes to netware.pl to enable gcc-cross builds on Win32
8874 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
8875 - various changes to fix missing prototype warnings
8876 - fixed x86nasm.pl to create correct asm files for NASM COFF output
8877 - added AES, WHIRLPOOL and CPUID assembler code to build files
8878 - added missing AES assembler make rules to mk1mf.pl
8879 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
8881 *Guenter Knauf <eflash@gmx.net>*
8883 * Implement certificate status request TLS extension defined in RFC3546.
8884 A client can set the appropriate parameters and receive the encoded
8885 OCSP response via a callback. A server can query the supplied parameters
8886 and set the encoded OCSP response in the callback. Add simplified examples
8887 to s_client and s_server.
8891 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007] ###
8894 + Binary incompatibility of ssl_ctx_st structure
8895 + DTLS interoperation with non-compliant servers
8896 + Don't call get_session_cb() without proposed session
8897 + Fix ia64 assembler code
8899 *Andy Polyakov, Steve Henson*
8901 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007] ###
8903 * DTLS Handshake overhaul. There were longstanding issues with
8904 OpenSSL DTLS implementation, which were making it impossible for
8905 RFC 4347 compliant client to communicate with OpenSSL server.
8906 Unfortunately just fixing these incompatibilities would "cut off"
8907 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
8908 server keeps tolerating non RFC compliant syntax. The opposite is
8909 not true, 0.9.8f client can not communicate with earlier server.
8910 This update even addresses CVE-2007-4995.
8914 * Changes to avoid need for function casts in OpenSSL: some compilers
8915 (gcc 4.2 and later) reject their use.
8916 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
8919 * Add RFC4507 support to OpenSSL. This includes the corrections in
8920 RFC4507bis. The encrypted ticket format is an encrypted encoded
8921 SSL_SESSION structure, that way new session features are automatically
8924 If a client application caches session in an SSL_SESSION structure
8925 support is transparent because tickets are now stored in the encoded
8928 The SSL_CTX structure automatically generates keys for ticket
8929 protection in servers so again support should be possible
8930 with no application modification.
8932 If a client or server wishes to disable RFC4507 support then the option
8933 SSL_OP_NO_TICKET can be set.
8935 Add a TLS extension debugging callback to allow the contents of any client
8936 or server extensions to be examined.
8938 This work was sponsored by Google.
8942 * Add initial support for TLS extensions, specifically for the server_name
8943 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8944 have new members for a host name. The SSL data structure has an
8945 additional member SSL_CTX *initial_ctx so that new sessions can be
8946 stored in that context to allow for session resumption, even after the
8947 SSL has been switched to a new SSL_CTX in reaction to a client's
8948 server_name extension.
8950 New functions (subject to change):
8952 SSL_get_servername()
8953 SSL_get_servername_type()
8956 New CTRL codes and macros (subject to change):
8958 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8959 - SSL_CTX_set_tlsext_servername_callback()
8960 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8961 - SSL_CTX_set_tlsext_servername_arg()
8962 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8964 openssl s_client has a new '-servername ...' option.
8966 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8967 '-key2 ...', '-servername_fatal' (subject to change). This allows
8968 testing the HostName extension for a specific single host name ('-cert'
8969 and '-key' remain fallbacks for handshakes without HostName
8970 negotiation). If the unrecognized_name alert has to be sent, this by
8971 default is a warning; it becomes fatal with the '-servername_fatal'
8975 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
8977 * Add AES and SSE2 assembly language support to VC++ build.
8981 * Mitigate attack on final subtraction in Montgomery reduction.
8985 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
8986 (which previously caused an internal error).
8990 * Squeeze another 10% out of IGE mode when in != out.
8994 * AES IGE mode speedup.
8996 *Dean Gaudet (Google)*
8998 * Add the Korean symmetric 128-bit cipher SEED (see
8999 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
9000 add SEED ciphersuites from RFC 4162:
9002 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9003 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9004 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9005 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9007 To minimize changes between patchlevels in the OpenSSL 0.9.8
9008 series, SEED remains excluded from compilation unless OpenSSL
9009 is configured with 'enable-seed'.
9011 *KISA, Bodo Moeller*
9013 * Mitigate branch prediction attacks, which can be practical if a
9014 single processor is shared, allowing a spy process to extract
9015 information. For detailed background information, see
9016 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
9017 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9018 and Necessary Software Countermeasures"). The core of the change
9019 are new versions BN_div_no_branch() and
9020 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9021 respectively, which are slower, but avoid the security-relevant
9022 conditional branches. These are automatically called by BN_div()
9023 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9024 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9025 remove a conditional branch.
9027 BN_FLG_CONSTTIME is the new name for the previous
9028 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9029 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9030 in the exponent causes BN_mod_exp_mont() to use the alternative
9031 implementation in BN_mod_exp_mont_consttime().) The old name
9032 remains as a deprecated alias.
9034 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9035 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9036 constant-time implementations for more than just exponentiation.
9037 Here too the old name is kept as a deprecated alias.
9039 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9040 the BN_BLINDING structure gets an independent copy of the
9041 modulus. This means that the previous "BIGNUM *m" argument to
9042 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9043 essentially becomes "const BIGNUM *m", although we can't actually
9044 change this in the header file before 0.9.9. It allows
9045 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9046 enable BN_FLG_CONSTTIME.
9049 *Matthew D Wood (Intel Corp)*
9051 * In the SSL/TLS server implementation, be strict about session ID
9052 context matching (which matters if an application uses a single
9053 external cache for different purposes). Previously,
9054 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9055 set. This did ensure strict client verification, but meant that,
9056 with applications using a single external cache for quite
9057 different requirements, clients could circumvent ciphersuite
9058 restrictions for a given session ID context by starting a session
9059 in a different context.
9063 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9064 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9065 authentication-only ciphersuites.
9069 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9070 not complete and could lead to a possible single byte overflow
9071 [CVE-2007-5135][] [Ben Laurie]
9073 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007] ###
9075 * Since AES128 and AES256 (and similarly Camellia128 and
9076 Camellia256) share a single mask bit in the logic of
9077 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9078 kludge to work properly if AES128 is available and AES256 isn't
9079 (or if Camellia128 is available and Camellia256 isn't).
9083 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9084 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9085 When a point or a seed is encoded in a BIT STRING, we need to
9086 prevent the removal of trailing zero bits to get the proper DER
9087 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9088 of a NamedBitList, for which trailing 0 bits need to be removed.)
9092 * Have SSL/TLS server implementation tolerate "mismatched" record
9093 protocol version while receiving ClientHello even if the
9094 ClientHello is fragmented. (The server can't insist on the
9095 particular protocol version it has chosen before the ServerHello
9096 message has informed the client about his choice.)
9100 * Add RFC 3779 support.
9102 *Rob Austein for ARIN, Ben Laurie*
9104 * Load error codes if they are not already present instead of using a
9105 static variable. This allows them to be cleanly unloaded and reloaded.
9106 Improve header file function name parsing.
9110 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9111 or CAPABILITY handshake as required by RFCs.
9115 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006] ###
9117 * Introduce limits to prevent malicious keys being able to
9118 cause a denial of service. [CVE-2006-2940][]
9120 *Steve Henson, Bodo Moeller*
9122 * Fix ASN.1 parsing of certain invalid structures that can result
9123 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9125 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9126 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9128 * Fix SSL client code which could crash if connecting to a
9129 malicious SSLv2 server. [CVE-2006-4343][]
9131 *Tavis Ormandy and Will Drewry, Google Security Team*
9133 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9134 match only those. Before that, "AES256-SHA" would be interpreted
9135 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9136 the same strength classification in 0.9.7h) as we currently only
9137 have a single AES bit in the ciphersuite description bitmap.
9138 That change, however, also applied to ciphersuite strings such as
9139 "RC4-MD5" that intentionally matched multiple ciphersuites --
9140 namely, SSL 2.0 ciphersuites in addition to the more common ones
9141 from SSL 3.0/TLS 1.0.
9143 So we change the selection algorithm again: Naming an explicit
9144 ciphersuite selects this one ciphersuite, and any other similar
9145 ciphersuite (same bitmap) from *other* protocol versions.
9146 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9147 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9149 Since SSL 2.0 does not have any ciphersuites for which the
9150 128/256 bit distinction would be relevant, this works for now.
9151 The proper fix will be to use different bits for AES128 and
9152 AES256, which would have avoided the problems from the beginning;
9153 however, bits are scarce, so we can only do this in a new release
9154 (not just a patchlevel) when we can change the SSL_CIPHER
9155 definition to split the single 'unsigned long mask' bitmap into
9156 multiple values to extend the available space.
9161 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006] ###
9163 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9164 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9166 * Add AES IGE and biIGE modes.
9170 * Change the Unix randomness entropy gathering to use poll() when
9171 possible instead of select(), since the latter has some
9172 undesirable limitations.
9174 *Darryl Miles via Richard Levitte and Bodo Moeller*
9176 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9177 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9178 cannot be implicitly activated as part of, e.g., the "AES" alias.
9179 However, please upgrade to OpenSSL 0.9.9[-dev] for
9180 non-experimental use of the ECC ciphersuites to get TLS extension
9181 support, which is required for curve and point format negotiation
9182 to avoid potential handshake problems.
9186 * Disable rogue ciphersuites:
9188 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9189 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9190 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9192 The latter two were purportedly from
9193 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9196 Also deactivate the remaining ciphersuites from
9197 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9198 unofficial, and the ID has long expired.
9202 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9203 dual-core machines) and other potential thread-safety issues.
9207 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9208 versions), which is now available for royalty-free use
9209 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
9210 Also, add Camellia TLS ciphersuites from RFC 4132.
9212 To minimize changes between patchlevels in the OpenSSL 0.9.8
9213 series, Camellia remains excluded from compilation unless OpenSSL
9214 is configured with 'enable-camellia'.
9218 * Disable the padding bug check when compression is in use. The padding
9219 bug check assumes the first packet is of even length, this is not
9220 necessarily true if compression is enabled and can result in false
9221 positives causing handshake failure. The actual bug test is ancient
9222 code so it is hoped that implementations will either have fixed it by
9223 now or any which still have the bug do not support compression.
9227 ### Changes between 0.9.8a and 0.9.8b [04 May 2006] ###
9229 * When applying a cipher rule check to see if string match is an explicit
9230 cipher suite and only match that one cipher suite if it is.
9234 * Link in manifests for VC++ if needed.
9236 *Austin Ziegler <halostatue@gmail.com>*
9238 * Update support for ECC-based TLS ciphersuites according to
9239 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9240 TLS extensions, which are supported starting with the 0.9.9
9241 branch, not in the OpenSSL 0.9.8 branch).
9245 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9246 opaque EVP_CIPHER_CTX handling.
9250 * Fixes and enhancements to zlib compression code. We now only use
9251 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9252 to conform with the standards mentioned here:
9253 http://www.zlib.net/DLL_FAQ.txt
9254 Static zlib linking now works on Windows and the new --with-zlib-include
9255 --with-zlib-lib options to Configure can be used to supply the location
9256 of the headers and library. Gracefully handle case where zlib library
9261 * Several fixes and enhancements to the OID generation code. The old code
9262 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9263 handle numbers larger than ULONG_MAX, truncated printing and had a
9264 non standard OBJ_obj2txt() behaviour.
9268 * Add support for building of engines under engine/ as shared libraries
9269 under VC++ build system.
9273 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9274 Hopefully, we will not see any false combination of paths any more.
9278 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005] ###
9280 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9281 (part of SSL_OP_ALL). This option used to disable the
9282 countermeasure against man-in-the-middle protocol-version
9283 rollback in the SSL 2.0 server implementation, which is a bad
9284 idea. [CVE-2005-2969][]
9286 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9287 for Information Security, National Institute of Advanced Industrial
9288 Science and Technology [AIST*, Japan)]
9290 * Add two function to clear and return the verify parameter flags.
9294 * Keep cipherlists sorted in the source instead of sorting them at
9295 runtime, thus removing the need for a lock.
9299 * Avoid some small subgroup attacks in Diffie-Hellman.
9301 *Nick Mathewson and Ben Laurie*
9303 * Add functions for well-known primes.
9307 * Extended Windows CE support.
9309 *Satoshi Nakamura and Andy Polyakov*
9311 * Initialize SSL_METHOD structures at compile time instead of during
9312 runtime, thus removing the need for a lock.
9316 * Make PKCS7_decrypt() work even if no certificate is supplied by
9317 attempting to decrypt each encrypted key in turn. Add support to
9322 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005] ###
9324 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9327 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9331 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9332 key into the same file any more.
9336 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9340 * Add -utf8 command line and config file option to 'ca'.
9342 *Stefan <stf@udoma.org*
9344 * Removed the macro des_crypt(), as it seems to conflict with some
9345 libraries. Use DES_crypt().
9349 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9350 involves renaming the source and generated shared-libs for
9351 both. The engines will accept the corrected or legacy ids
9352 ('ncipher' and '4758_cca' respectively) when binding. NB,
9353 this only applies when building 'shared'.
9355 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9357 * Add attribute functions to EVP_PKEY structure. Modify
9358 PKCS12_create() to recognize a CSP name attribute and
9359 use it. Make -CSP option work again in pkcs12 utility.
9363 * Add new functionality to the bn blinding code:
9364 - automatic re-creation of the BN_BLINDING parameters after
9365 a fixed number of uses (currently 32)
9366 - add new function for parameter creation
9367 - introduce flags to control the update behaviour of the
9368 BN_BLINDING parameters
9369 - hide BN_BLINDING structure
9370 Add a second BN_BLINDING slot to the RSA structure to improve
9371 performance when a single RSA object is shared among several
9376 * Add support for DTLS.
9378 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9380 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9381 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9385 * Remove buggy and incomplete DH cert support from
9386 ssl/ssl_rsa.c and ssl/s3_both.c
9390 * Use SHA-1 instead of MD5 as the default digest algorithm for
9391 the apps/openssl applications.
9395 * Compile clean with "-Wall -Wmissing-prototypes
9396 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9397 DEBUG_SAFESTACK must also be set.
9401 * Change ./Configure so that certain algorithms can be disabled by default.
9402 The new counterpiece to "no-xxx" is "enable-xxx".
9404 The patented RC5 and MDC2 algorithms will now be disabled unless
9405 "enable-rc5" and "enable-mdc2", respectively, are specified.
9407 (IDEA remains enabled despite being patented. This is because IDEA
9408 is frequently required for interoperability, and there is no license
9409 fee for non-commercial use. As before, "no-idea" can be used to
9410 avoid this algorithm.)
9415 * Add processing of proxy certificates (see RFC 3820). This work was
9416 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9417 EGEE (Enabling Grids for E-science in Europe).
9421 * RC4 performance overhaul on modern architectures/implementations, such
9422 as Intel P4, IA-64 and AMD64.
9426 * New utility extract-section.pl. This can be used specify an alternative
9427 section number in a pod file instead of having to treat each file as
9428 a separate case in Makefile. This can be done by adding two lines to the
9431 =for comment openssl_section:XXX
9433 The blank line is mandatory.
9438 * New arguments -certform, -keyform and -pass for s_client and s_server
9439 to allow alternative format key and certificate files and passphrase
9444 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9445 update associated structures and add various utility functions.
9447 Add new policy related verify parameters, include policy checking in
9448 standard verify code. Enhance 'smime' application with extra parameters
9449 to support policy checking and print out.
9453 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9454 Nehemiah processors. These extensions support AES encryption in hardware
9455 as well as RNG (though RNG support is currently disabled).
9457 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9459 * Deprecate BN_[get|set]_params() functions (they were ignored internally).
9463 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9465 *Andy Polyakov and a number of other people*
9467 * Improved PowerPC platform support. Most notably BIGNUM assembler
9468 implementation contributed by IBM.
9470 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9472 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9473 exponent rather than 'unsigned long'. There is a corresponding change to
9474 the new 'rsa_keygen' element of the RSA_METHOD structure.
9476 *Jelte Jansen, Geoff Thorpe*
9478 * Functionality for creating the initial serial number file is now
9479 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9481 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9482 number file to 1, which is bound to cause problems. To avoid
9483 the problems while respecting compatibility between different 0.9.7
9484 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9485 CA.pl for serial number initialization. With the new release 0.9.8,
9486 we can fix the problem directly in the 'ca' utility.)
9490 * Reduced header interdependencies by declaring more opaque objects in
9491 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9492 give fewer recursive includes, which could break lazy source code - so
9493 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9494 developers should define this symbol when building and using openssl to
9495 ensure they track the recommended behaviour, interfaces, [etc], but
9496 backwards-compatible behaviour prevails when this isn't defined.
9500 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9504 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9505 This will generate a random key of the appropriate length based on the
9506 cipher context. The EVP_CIPHER can provide its own random key generation
9507 routine to support keys of a specific form. This is used in the des and
9508 3des routines to generate a key of the correct parity. Update S/MIME
9509 code to use new functions and hence generate correct parity DES keys.
9510 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9511 valid (weak or incorrect parity).
9515 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9516 as looking them up. This is useful when the verified structure may contain
9517 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9518 present unless the new PKCS7_NO_CRL flag is asserted.
9522 * Extend ASN1 oid configuration module. It now additionally accepts the
9525 shortName = some long name, 1.2.3.4
9529 * Reimplemented the BN_CTX implementation. There is now no more static
9530 limitation on the number of variables it can handle nor the depth of the
9531 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9532 information can now expand as required, and rather than having a single
9533 static array of bignums, BN_CTX now uses a linked-list of such arrays
9534 allowing it to expand on demand whilst maintaining the usefulness of
9535 BN_CTX's "bundling".
9539 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9540 to allow all RSA operations to function using a single BN_CTX.
9544 * Preliminary support for certificate policy evaluation and checking. This
9545 is initially intended to pass the tests outlined in "Conformance Testing
9546 of Relying Party Client Certificate Path Processing Logic" v1.07.
9550 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9551 remained unused and not that useful. A variety of other little bignum
9552 tweaks and fixes have also been made continuing on from the audit (see
9557 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9558 associated ASN1, EVP and SSL functions and old ASN1 macros.
9562 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9563 and this should never fail. So the return value from the use of
9564 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9565 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9569 * BN_CTX_get() should return zero-valued bignums, providing the same
9570 initialised value as BN_new().
9572 *Geoff Thorpe, suggested by Ulf Möller*
9574 * Support for inhibitAnyPolicy certificate extension.
9578 * An audit of the BIGNUM code is underway, for which debugging code is
9579 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9580 is considered valid when processing BIGNUMs, and causes execution to
9581 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9582 further steps are taken to deliberately pollute unused data in BIGNUM
9583 structures to try and expose faulty code further on. For now, openssl will
9584 (in its default mode of operation) continue to tolerate the inconsistent
9585 forms that it has tolerated in the past, but authors and packagers should
9586 consider trying openssl and their own applications when compiled with
9587 these debugging symbols defined. It will help highlight potential bugs in
9588 their own code, and will improve the test coverage for OpenSSL itself. At
9589 some point, these tighter rules will become openssl's default to improve
9590 maintainability, though the assert()s and other overheads will remain only
9591 in debugging configurations. See bn.h for more details.
9593 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9595 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9596 that can only be obtained through BN_CTX_new() (which implicitly
9597 initialises it). The presence of this function only made it possible
9598 to overwrite an existing structure (and cause memory leaks).
9602 * Because of the callback-based approach for implementing LHASH as a
9603 template type, lh_insert() adds opaque objects to hash-tables and
9604 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9605 to clean up those corresponding objects before destroying the hash table
9606 (and losing the object pointers). So some over-zealous constifications in
9607 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9608 objects as "const" and the lh_doall[_arg] callback wrappers are not
9609 prototyped to have "const" restrictions on the object pointers they are
9610 given (and so aren't required to cast them away any more).
9614 * The tmdiff.h API was so ugly and minimal that our own timing utility
9615 (speed) prefers to use its own implementation. The two implementations
9616 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9617 its object type properly exposed (MS_TM) instead of casting to/from "char
9618 *". This may still change yet if someone realises MS_TM and "ms_time_***"
9619 aren't necessarily the greatest nomenclatures - but this is what was used
9620 internally to the implementation so I've used that for now.
9624 * Ensure that deprecated functions do not get compiled when
9625 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9626 the self-tests were still using deprecated key-generation functions so
9627 these have been updated also.
9631 * Reorganise PKCS#7 code to separate the digest location functionality
9632 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9633 New function PKCS7_set_digest() to set the digest type for PKCS#7
9634 digestedData type. Add additional code to correctly generate the
9635 digestedData type and add support for this type in PKCS7 initialization
9640 * New function PKCS7_set0_type_other() this initializes a PKCS7
9641 structure of type "other".
9645 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9646 sure the loop does correctly stop and breaking ("division by zero")
9647 modulus operations are not performed. The (pre-generated) prime
9648 table crypto/bn/bn_prime.h was already correct, but it could not be
9649 re-generated on some platforms because of the "division by zero"
9650 situation in the script.
9652 *Ralf S. Engelschall*
9654 * Update support for ECC-based TLS ciphersuites according to
9655 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9656 SHA-1 now is only used for "small" curves (where the
9657 representation of a field element takes up to 24 bytes); for
9658 larger curves, the field element resulting from ECDH is directly
9659 used as premaster secret.
9661 *Douglas Stebila (Sun Microsystems Laboratories)*
9663 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9664 curve secp160r1 to the tests.
9666 *Douglas Stebila (Sun Microsystems Laboratories)*
9668 * Add the possibility to load symbols globally with DSO.
9670 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9672 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9673 control of the error stack.
9677 * Add support for STORE in ENGINE.
9681 * Add the STORE type. The intention is to provide a common interface
9682 to certificate and key stores, be they simple file-based stores, or
9683 HSM-type store, or LDAP stores, or...
9684 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9688 * Add a generic structure called OPENSSL_ITEM. This can be used to
9689 pass a list of arguments to any function as well as provide a way
9690 for a function to pass data back to the caller.
9694 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9695 works like BUF_strdup() but can be used to duplicate a portion of
9696 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9701 * Add the function sk_find_ex() which works like sk_find(), but will
9702 return an index to an element even if an exact match couldn't be
9703 found. The index is guaranteed to point at the element where the
9704 searched-for key would be inserted to preserve sorting order.
9708 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9709 takes an extra flags argument for optional functionality. Currently,
9710 the following flags are defined:
9712 OBJ_BSEARCH_VALUE_ON_NOMATCH
9713 This one gets OBJ_bsearch_ex() to return a pointer to the first
9714 element where the comparing function returns a negative or zero
9717 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9718 This one gets OBJ_bsearch_ex() to return a pointer to the first
9719 element where the comparing function returns zero. This is useful
9720 if there are more than one element where the comparing function
9725 * Make it possible to create self-signed certificates with 'openssl ca'
9726 in such a way that the self-signed certificate becomes part of the
9727 CA database and uses the same mechanisms for serial number generation
9728 as all other certificate signing. The new flag '-selfsign' enables
9729 this functionality. Adapt CA.sh and CA.pl.in.
9733 * Add functionality to check the public key of a certificate request
9734 against a given private. This is useful to check that a certificate
9735 request can be signed by that key (self-signing).
9739 * Make it possible to have multiple active certificates with the same
9740 subject in the CA index file. This is done only if the keyword
9741 'unique_subject' is set to 'no' in the main CA section (default
9742 if 'CA_default') of the configuration file. The value is saved
9743 with the database itself in a separate index attribute file,
9744 named like the index file with '.attr' appended to the name.
9748 * Generate multi-valued AVAs using '+' notation in config files for
9753 * Support for nameConstraints certificate extension.
9757 * Support for policyConstraints certificate extension.
9761 * Support for policyMappings certificate extension.
9765 * Make sure the default DSA_METHOD implementation only uses its
9766 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9767 and change its own handlers to be NULL so as to remove unnecessary
9768 indirection. This lets alternative implementations fallback to the
9769 default implementation more easily.
9773 * Support for directoryName in GeneralName related extensions
9778 * Make it possible to link applications using Makefile.shared.
9779 Make that possible even when linking against static libraries!
9783 * Support for single pass processing for S/MIME signing. This now
9784 means that S/MIME signing can be done from a pipe, in addition
9785 cleartext signing (multipart/signed type) is effectively streaming
9786 and the signed data does not need to be all held in memory.
9788 This is done with a new flag PKCS7_STREAM. When this flag is set
9789 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9790 is done after the data is output (and digests calculated) in
9791 SMIME_write_PKCS7().
9795 * Add full support for -rpath/-R, both in shared libraries and
9796 applications, at least on the platforms where it's known how
9801 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9802 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9803 will now compute a table of multiples of the generator that
9804 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9805 faster (notably in the case of a single point multiplication,
9806 scalar * generator).
9808 *Nils Larsch, Bodo Moeller*
9810 * IPv6 support for certificate extensions. The various extensions
9811 which use the IP:a.b.c.d can now take IPv6 addresses using the
9812 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9817 * Added an ENGINE that implements RSA by performing private key
9818 exponentiations with the GMP library. The conversions to and from
9819 GMP's mpz_t format aren't optimised nor are any montgomery forms
9820 cached, and on x86 it appears OpenSSL's own performance has caught up.
9821 However there are likely to be other architectures where GMP could
9822 provide a boost. This ENGINE is not built in by default, but it can be
9823 specified at Configure time and should be accompanied by the necessary
9824 linker additions, eg;
9825 ./config -DOPENSSL_USE_GMP -lgmp
9829 * "openssl engine" will not display ENGINE/DSO load failure errors when
9830 testing availability of engines with "-t" - the old behaviour is
9831 produced by increasing the feature's verbosity with "-tt".
9835 * ECDSA routines: under certain error conditions uninitialized BN objects
9836 could be freed. Solution: make sure initialization is performed early
9837 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
9842 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
9843 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9844 software implementations. For DSA and DH, parameter generation can
9845 also be overridden by providing the appropriate method callbacks.
9849 * Change the "progress" mechanism used in key-generation and
9850 primality testing to functions that take a new BN_GENCB pointer in
9851 place of callback/argument pairs. The new API functions have "_ex"
9852 postfixes and the older functions are reimplemented as wrappers for
9853 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
9854 declarations of the old functions to help (graceful) attempts to
9855 migrate to the new functions. Also, the new key-generation API
9856 functions operate on a caller-supplied key-structure and return
9857 success/failure rather than returning a key or NULL - this is to
9858 help make "keygen" another member function of RSA_METHOD etc.
9860 Example for using the new callback interface:
9862 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
9866 BN_GENCB_set(&my_cb, my_callback, my_arg);
9868 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
9869 /* For the meaning of a, b in calls to my_callback(), see the
9870 * documentation of the function that calls the callback.
9871 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
9872 * my_callback should return 1 if it wants BN_is_prime_ex()
9873 * to continue, or 0 to stop.
9878 * Change the ZLIB compression method to be stateful, and make it
9879 available to TLS with the number defined in
9880 draft-ietf-tls-compression-04.txt.
9884 * Add the ASN.1 structures and functions for CertificatePair, which
9885 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
9887 CertificatePair ::= SEQUENCE {
9888 forward [0] Certificate OPTIONAL,
9889 reverse [1] Certificate OPTIONAL,
9890 -- at least one of the pair shall be present -- }
9892 Also implement the PEM functions to read and write certificate
9893 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
9895 This needed to be defined, mostly for the sake of the LDAP
9896 attribute crossCertificatePair, but may prove useful elsewhere as
9901 * Make it possible to inhibit symlinking of shared libraries in
9902 Makefile.shared, for Cygwin's sake.
9906 * Extend the BIGNUM API by creating a function
9907 void BN_set_negative(BIGNUM *a, int neg);
9908 and a macro that behave like
9909 int BN_is_negative(const BIGNUM *a);
9911 to avoid the need to access 'a->neg' directly in applications.
9915 * Implement fast modular reduction for pseudo-Mersenne primes
9916 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
9917 EC_GROUP_new_curve_GFp() will now automatically use this
9920 *Nils Larsch <nla@trustcenter.de>*
9922 * Add new lock type (CRYPTO_LOCK_BN).
9926 * Change the ENGINE framework to automatically load engines
9927 dynamically from specific directories unless they could be
9928 found to already be built in or loaded. Move all the
9929 current engines except for the cryptodev one to a new
9931 The engines in engines/ are built as shared libraries if
9932 the "shared" options was given to ./Configure or ./config.
9933 Otherwise, they are inserted in libcrypto.a.
9934 /usr/local/ssl/engines is the default directory for dynamic
9935 engines, but that can be overridden at configure time through
9936 the usual use of --prefix and/or --openssldir, and at run
9937 time with the environment variable OPENSSL_ENGINES.
9939 *Geoff Thorpe and Richard Levitte*
9941 * Add Makefile.shared, a helper makefile to build shared
9942 libraries. Adapt Makefile.org.
9946 * Add version info to Win32 DLLs.
9948 *Peter 'Luna' Runestig" <peter@runestig.com>*
9950 * Add new 'medium level' PKCS#12 API. Certificates and keys
9951 can be added using this API to created arbitrary PKCS#12
9952 files while avoiding the low level API.
9954 New options to PKCS12_create(), key or cert can be NULL and
9955 will then be omitted from the output file. The encryption
9956 algorithm NIDs can be set to -1 for no encryption, the mac
9957 iteration count can be set to 0 to omit the mac.
9959 Enhance pkcs12 utility by making the -nokeys and -nocerts
9960 options work when creating a PKCS#12 file. New option -nomac
9961 to omit the mac, NONE can be set for an encryption algorithm.
9962 New code is modified to use the enhanced PKCS12_create()
9963 instead of the low level API.
9967 * Extend ASN1 encoder to support indefinite length constructed
9968 encoding. This can output sequences tags and octet strings in
9969 this form. Modify pk7_asn1.c to support indefinite length
9970 encoding. This is experimental and needs additional code to
9971 be useful, such as an ASN1 bio and some enhanced streaming
9974 Extend template encode functionality so that tagging is passed
9975 down to the template encoder.
9979 * Let 'openssl req' fail if an argument to '-newkey' is not
9980 recognized instead of using RSA as a default.
9984 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
9985 As these are not official, they are not included in "ALL";
9986 the "ECCdraft" ciphersuite group alias can be used to select them.
9988 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
9990 * Add ECDH engine support.
9992 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
9994 * Add ECDH in new directory crypto/ecdh/.
9996 *Douglas Stebila (Sun Microsystems Laboratories)*
9998 * Let BN_rand_range() abort with an error after 100 iterations
9999 without success (which indicates a broken PRNG).
10003 * Change BN_mod_sqrt() so that it verifies that the input value
10004 is really the square of the return value. (Previously,
10005 BN_mod_sqrt would show GIGO behaviour.)
10009 * Add named elliptic curves over binary fields from X9.62, SECG,
10010 and WAP/WTLS; add OIDs that were still missing.
10012 *Sheueling Chang Shantz and Douglas Stebila
10013 (Sun Microsystems Laboratories)*
10015 * Extend the EC library for elliptic curves over binary fields
10016 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10019 EC_GF2m_simple_method
10023 EC_GROUP_new_curve_GF2m
10024 EC_GROUP_set_curve_GF2m
10025 EC_GROUP_get_curve_GF2m
10026 EC_POINT_set_affine_coordinates_GF2m
10027 EC_POINT_get_affine_coordinates_GF2m
10028 EC_POINT_set_compressed_coordinates_GF2m
10030 Point compression for binary fields is disabled by default for
10031 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10034 As binary polynomials are represented as BIGNUMs, various members
10035 of the EC_GROUP and EC_POINT data structures can be shared
10036 between the implementations for prime fields and binary fields;
10037 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
10038 are essentially identical to their ..._GFp counterparts.
10039 (For simplicity, the '..._GFp' prefix has been dropped from
10040 various internal method names.)
10042 An internal 'field_div' method (similar to 'field_mul' and
10043 'field_sqr') has been added; this is used only for binary fields.
10045 *Sheueling Chang Shantz and Douglas Stebila
10046 (Sun Microsystems Laboratories)*
10048 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10049 through methods ('mul', 'precompute_mult').
10051 The generic implementations (now internally called 'ec_wNAF_mul'
10052 and 'ec_wNAF_precomputed_mult') remain the default if these
10053 methods are undefined.
10055 *Sheueling Chang Shantz and Douglas Stebila
10056 (Sun Microsystems Laboratories)*
10058 * New function EC_GROUP_get_degree, which is defined through
10059 EC_METHOD. For curves over prime fields, this returns the bit
10060 length of the modulus.
10062 *Sheueling Chang Shantz and Douglas Stebila
10063 (Sun Microsystems Laboratories)*
10065 * New functions EC_GROUP_dup, EC_POINT_dup.
10066 (These simply call ..._new and ..._copy).
10068 *Sheueling Chang Shantz and Douglas Stebila
10069 (Sun Microsystems Laboratories)*
10071 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10072 Polynomials are represented as BIGNUMs (where the sign bit is not
10073 used) in the following functions [macros]:
10076 BN_GF2m_sub [= BN_GF2m_add]
10077 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10078 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10079 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10081 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10082 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10083 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10084 BN_GF2m_cmp [= BN_ucmp]
10086 (Note that only the 'mod' functions are actually for fields GF(2^m).
10087 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10089 For some functions, an the irreducible polynomial defining a
10090 field can be given as an 'unsigned int[]' with strictly
10091 decreasing elements giving the indices of those bits that are set;
10092 i.e., p[] represents the polynomial
10093 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10095 p[0] > p[1] > ... > p[k] = 0.
10096 This applies to the following functions:
10099 BN_GF2m_mod_mul_arr
10100 BN_GF2m_mod_sqr_arr
10101 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10102 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10103 BN_GF2m_mod_exp_arr
10104 BN_GF2m_mod_sqrt_arr
10105 BN_GF2m_mod_solve_quad_arr
10109 Conversion can be performed by the following functions:
10114 bntest.c has additional tests for binary polynomial arithmetic.
10116 Two implementations for BN_GF2m_mod_div() are available.
10117 The default algorithm simply uses BN_GF2m_mod_inv() and
10118 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10119 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10120 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10122 *Sheueling Chang Shantz and Douglas Stebila
10123 (Sun Microsystems Laboratories)*
10125 * Add new error code 'ERR_R_DISABLED' that can be used when some
10126 functionality is disabled at compile-time.
10128 *Douglas Stebila <douglas.stebila@sun.com>*
10130 * Change default behaviour of 'openssl asn1parse' so that more
10131 information is visible when viewing, e.g., a certificate:
10133 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10134 mode the content of non-printable OCTET STRINGs is output in a
10135 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10136 avoid the appearance of a printable string.
10138 *Nils Larsch <nla@trustcenter.de>*
10140 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10142 EC_GROUP_set_asn1_flag()
10143 EC_GROUP_get_asn1_flag()
10144 EC_GROUP_set_point_conversion_form()
10145 EC_GROUP_get_point_conversion_form()
10146 These control ASN1 encoding details:
10147 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10148 has been set to OPENSSL_EC_NAMED_CURVE.
10149 - Points are encoded in uncompressed form by default; options for
10150 asn1_for are as for point2oct, namely
10151 POINT_CONVERSION_COMPRESSED
10152 POINT_CONVERSION_UNCOMPRESSED
10153 POINT_CONVERSION_HYBRID
10155 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10157 EC_GROUP_set_seed()
10158 EC_GROUP_get0_seed()
10159 EC_GROUP_get_seed_len()
10160 This is used only for ASN1 purposes (so far).
10162 *Nils Larsch <nla@trustcenter.de>*
10164 * Add 'field_type' member to EC_METHOD, which holds the NID
10165 of the appropriate field type OID. The new function
10166 EC_METHOD_get_field_type() returns this value.
10168 *Nils Larsch <nla@trustcenter.de>*
10171 EC_POINT_point2bn()
10172 EC_POINT_bn2point()
10173 EC_POINT_point2hex()
10174 EC_POINT_hex2point()
10175 providing useful interfaces to EC_POINT_point2oct() and
10176 EC_POINT_oct2point().
10178 *Nils Larsch <nla@trustcenter.de>*
10180 * Change internals of the EC library so that the functions
10181 EC_GROUP_set_generator()
10182 EC_GROUP_get_generator()
10183 EC_GROUP_get_order()
10184 EC_GROUP_get_cofactor()
10185 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10186 to methods, which would lead to unnecessary code duplication when
10187 adding different types of curves.
10189 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10191 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10192 arithmetic, and such that modified wNAFs are generated
10193 (which avoid length expansion in many cases).
10197 * Add a function EC_GROUP_check_discriminant() (defined via
10198 EC_METHOD) that verifies that the curve discriminant is non-zero.
10200 Add a function EC_GROUP_check() that makes some sanity tests
10201 on a EC_GROUP, its generator and order. This includes
10202 EC_GROUP_check_discriminant().
10204 *Nils Larsch <nla@trustcenter.de>*
10206 * Add ECDSA in new directory crypto/ecdsa/.
10208 Add applications 'openssl ecparam' and 'openssl ecdsa'
10209 (these are based on 'openssl dsaparam' and 'openssl dsa').
10211 ECDSA support is also included in various other files across the
10212 library. Most notably,
10213 - 'openssl req' now has a '-newkey ecdsa:file' option;
10214 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10215 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10216 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10217 them suitable for ECDSA where domain parameters must be
10218 extracted before the specific public key;
10219 - ECDSA engine support has been added.
10221 *Nils Larsch <nla@trustcenter.de>*
10223 * Include some named elliptic curves, and add OIDs from X9.62,
10224 SECG, and WAP/WTLS. Each curve can be obtained from the new
10226 EC_GROUP_new_by_curve_name(),
10227 and the list of available named curves can be obtained with
10228 EC_get_builtin_curves().
10229 Also add a 'curve_name' member to EC_GROUP objects, which can be
10231 EC_GROUP_set_curve_name()
10232 EC_GROUP_get_curve_name()
10234 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10236 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10237 was actually never needed) and in BN_mul(). The removal in BN_mul()
10238 required a small change in bn_mul_part_recursive() and the addition
10239 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10240 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10241 bn_sub_words() and bn_add_words() except they take arrays with
10246 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007] ###
10248 * Cleanse PEM buffers before freeing them since they may contain
10251 *Benjamin Bennett <ben@psc.edu>*
10253 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10254 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10255 authentication-only ciphersuites.
10259 * Since AES128 and AES256 share a single mask bit in the logic of
10260 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10261 kludge to work properly if AES128 is available and AES256 isn't.
10265 * Expand security boundary to match 1.1.1 module.
10269 * Remove redundant features: hash file source, editing of test vectors
10270 modify fipsld to use external fips_premain.c signature.
10274 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10275 run algorithm test programs.
10279 * Make algorithm test programs more tolerant of whitespace.
10283 * Have SSL/TLS server implementation tolerate "mismatched" record
10284 protocol version while receiving ClientHello even if the
10285 ClientHello is fragmented. (The server can't insist on the
10286 particular protocol version it has chosen before the ServerHello
10287 message has informed the client about his choice.)
10291 * Load error codes if they are not already present instead of using a
10292 static variable. This allows them to be cleanly unloaded and reloaded.
10296 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006] ###
10298 * Introduce limits to prevent malicious keys being able to
10299 cause a denial of service. [CVE-2006-2940][]
10301 *Steve Henson, Bodo Moeller*
10303 * Fix ASN.1 parsing of certain invalid structures that can result
10304 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10306 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10307 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10309 * Fix SSL client code which could crash if connecting to a
10310 malicious SSLv2 server. [CVE-2006-4343][]
10312 *Tavis Ormandy and Will Drewry, Google Security Team*
10314 * Change ciphersuite string processing so that an explicit
10315 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10316 will no longer include "AES128-SHA"), and any other similar
10317 ciphersuite (same bitmap) from *other* protocol versions (so that
10318 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10319 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10320 changes from 0.9.8b and 0.9.8d.
10324 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006] ###
10326 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10327 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10329 * Change the Unix randomness entropy gathering to use poll() when
10330 possible instead of select(), since the latter has some
10331 undesirable limitations.
10333 *Darryl Miles via Richard Levitte and Bodo Moeller*
10335 * Disable rogue ciphersuites:
10337 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10338 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10339 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10341 The latter two were purportedly from
10342 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10345 Also deactivate the remaining ciphersuites from
10346 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10347 unofficial, and the ID has long expired.
10351 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10352 dual-core machines) and other potential thread-safety issues.
10356 ### Changes between 0.9.7i and 0.9.7j [04 May 2006] ###
10358 * Adapt fipsld and the build system to link against the validated FIPS
10359 module in FIPS mode.
10363 * Fixes for VC++ 2005 build under Windows.
10367 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10368 from a Windows bash shell such as MSYS. It is autodetected from the
10369 "config" script when run from a VC++ environment. Modify standard VC++
10370 build to use fipscanister.o from the GNU make build.
10374 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005] ###
10376 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10377 The value now differs depending on if you build for FIPS or not.
10378 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10379 safely run with a non-FIPSed libcrypto, as it may crash because of
10380 the difference induced by this change.
10384 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005] ###
10386 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10387 (part of SSL_OP_ALL). This option used to disable the
10388 countermeasure against man-in-the-middle protocol-version
10389 rollback in the SSL 2.0 server implementation, which is a bad
10390 idea. [CVE-2005-2969][]
10392 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10393 for Information Security, National Institute of Advanced Industrial
10394 Science and Technology [AIST*, Japan)]
10396 * Minimal support for X9.31 signatures and PSS padding modes. This is
10397 mainly for FIPS compliance and not fully integrated at this stage.
10401 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10402 the exponentiation using a fixed-length exponent. (Otherwise,
10403 the information leaked through timing could expose the secret key
10404 after many signatures; cf. Bleichenbacher's attack on DSA with
10409 * Make a new fixed-window mod_exp implementation the default for
10410 RSA, DSA, and DH private-key operations so that the sequence of
10411 squares and multiplies and the memory access pattern are
10412 independent of the particular secret key. This will mitigate
10413 cache-timing and potential related attacks.
10415 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10416 and this is automatically used by BN_mod_exp_mont() if the new flag
10417 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10418 will use this BN flag for private exponents unless the flag
10419 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10420 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10423 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10425 * Change the client implementation for SSLv23_method() and
10426 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10427 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10428 (Previously, the SSL 2.0 backwards compatible Client Hello
10429 message format would be used even with SSL_OP_NO_SSLv2.)
10433 * Add support for smime-type MIME parameter in S/MIME messages which some
10438 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10439 a threadsafe manner. Modify rsa code to use new function and add calls
10440 to dsa and dh code (which had race conditions before).
10444 * Include the fixed error library code in the C error file definitions
10445 instead of fixing them up at runtime. This keeps the error code
10446 structures constant.
10450 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005] ###
10452 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10455 * Fixes for newer kerberos headers. NB: the casts are needed because
10456 the 'length' field is signed on one version and unsigned on another
10457 with no (?) obvious way to tell the difference, without these VC++
10458 complains. Also the "definition" of FAR (blank) is no longer included
10459 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10460 some needed definitions.
10464 * Undo Cygwin change.
10468 * Added support for proxy certificates according to RFC 3820.
10469 Because they may be a security thread to unaware applications,
10470 they must be explicitly allowed in run-time. See
10471 docs/HOWTO/proxy_certificates.txt for further information.
10475 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005] ###
10477 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10478 server and client random values. Previously
10479 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10480 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10482 This change has negligible security impact because:
10484 1. Server and client random values still have 24 bytes of pseudo random
10487 2. Server and client random values are sent in the clear in the initial
10490 3. The master secret is derived using the premaster secret (48 bytes in
10491 size for static RSA ciphersuites) as well as client server and random
10494 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10497 *Stephen Henson, reported by UK NISCC*
10499 * Use Windows randomness collection on Cygwin.
10503 * Fix hang in EGD/PRNGD query when communication socket is closed
10504 prematurely by EGD/PRNGD.
10506 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10508 * Prompt for pass phrases when appropriate for PKCS12 input format.
10512 * Back-port of selected performance improvements from development
10513 branch, as well as improved support for PowerPC platforms.
10517 * Add lots of checks for memory allocation failure, error codes to indicate
10518 failure and freeing up memory if a failure occurs.
10520 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10522 * Add new -passin argument to dgst.
10526 * Perform some character comparisons of different types in X509_NAME_cmp:
10527 this is needed for some certificates that re-encode DNs into UTF8Strings
10528 (in violation of RFC3280) and can't or won't issue name rollover
10533 * Make an explicit check during certificate validation to see that
10534 the CA setting in each certificate on the chain is correct. As a
10535 side effect always do the following basic checks on extensions,
10536 not just when there's an associated purpose to the check:
10538 - if there is an unhandled critical extension (unless the user
10539 has chosen to ignore this fault)
10540 - if the path length has been exceeded (if one is set at all)
10541 - that certain extensions fit the associated purpose (if one has
10546 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004] ###
10548 * Avoid a race condition when CRLs are checked in a multi threaded
10549 environment. This would happen due to the reordering of the revoked
10550 entries during signature checking and serial number lookup. Now the
10551 encoding is cached and the serial number sort performed under a lock.
10552 Add new STACK function sk_is_sorted().
10556 * Add Delta CRL to the extension code.
10560 * Various fixes to s3_pkt.c so alerts are sent properly.
10562 *David Holmes <d.holmes@f5.com>*
10564 * Reduce the chances of duplicate issuer name and serial numbers (in
10565 violation of RFC3280) using the OpenSSL certificate creation utilities.
10566 This is done by creating a random 64 bit value for the initial serial
10567 number when a serial number file is created or when a self signed
10568 certificate is created using 'openssl req -x509'. The initial serial
10569 number file is created using 'openssl x509 -next_serial' in CA.pl
10570 rather than being initialized to 1.
10574 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004] ###
10576 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10577 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10579 *Joe Orton, Steve Henson*
10581 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10584 *Joe Orton, Steve Henson*
10586 * Make it possible to have multiple active certificates with the same
10587 subject in the CA index file. This is done only if the keyword
10588 'unique_subject' is set to 'no' in the main CA section (default
10589 if 'CA_default') of the configuration file. The value is saved
10590 with the database itself in a separate index attribute file,
10591 named like the index file with '.attr' appended to the name.
10595 * X509 verify fixes. Disable broken certificate workarounds when
10596 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10597 keyUsage extension present. Don't accept CRLs with unhandled critical
10598 extensions: since verify currently doesn't process CRL extensions this
10599 rejects a CRL with *any* critical extensions. Add new verify error codes
10604 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10605 A clarification of RFC2560 will require the use of OCTET STRINGs and
10606 some implementations cannot handle the current raw format. Since OpenSSL
10607 copies and compares OCSP nonces as opaque blobs without any attempt at
10608 parsing them this should not create any compatibility issues.
10612 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10613 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10614 this HMAC (and other) operations are several times slower than OpenSSL
10619 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10621 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10623 * Use the correct content when signing type "other".
10627 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003] ###
10629 * Fix various bugs revealed by running the NISCC test suite:
10631 Stop out of bounds reads in the ASN1 code when presented with
10632 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10634 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10636 If verify callback ignores invalid public key errors don't try to check
10637 certificate signature with the NULL public key.
10642 * New -ignore_err option in ocsp application to stop the server
10643 exiting on the first error in a request.
10647 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10648 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10653 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10654 extra data after the compression methods not only for TLS 1.0
10655 but also for SSL 3.0 (as required by the specification).
10657 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10659 * Change X509_certificate_type() to mark the key as exported/exportable
10660 when it's 512 *bits* long, not 512 bytes.
10664 * Change AES_cbc_encrypt() so it outputs exact multiple of
10665 blocks during encryption.
10669 * Various fixes to base64 BIO and non blocking I/O. On write
10670 flushes were not handled properly if the BIO retried. On read
10671 data was not being buffered properly and had various logic bugs.
10672 This also affects blocking I/O when the data being decoded is a
10677 * Various S/MIME bugfixes and compatibility changes:
10678 output correct application/pkcs7 MIME type if
10679 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10680 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10681 of files as .eml work). Correctly handle very long lines in MIME
10686 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003] ###
10688 * Countermeasure against the Klima-Pokorny-Rosa extension of
10689 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10690 a protocol version number mismatch like a decryption error
10691 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10695 * Turn on RSA blinding by default in the default implementation
10696 to avoid a timing attack. Applications that don't want it can call
10697 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10698 They would be ill-advised to do so in most cases.
10700 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10702 * Change RSA blinding code so that it works when the PRNG is not
10703 seeded (in this case, the secret RSA exponent is abused as
10704 an unpredictable seed -- if it is not unpredictable, there
10705 is no point in blinding anyway). Make RSA blinding thread-safe
10706 by remembering the creator's thread ID in rsa->blinding and
10707 having all other threads use local one-time blinding factors
10708 (this requires more computation than sharing rsa->blinding, but
10709 avoids excessive locking; and if an RSA object is not shared
10710 between threads, blinding will still be very fast).
10714 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10715 ENGINE as defaults for all supported algorithms irrespective of
10716 the 'flags' parameter. 'flags' is now honoured, so applications
10717 should make sure they are passing it correctly.
10721 * Target "mingw" now allows native Windows code to be generated in
10722 the Cygwin environment as well as with the MinGW compiler.
10726 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003] ###
10728 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10729 via timing by performing a MAC computation even if incorrect
10730 block cipher padding has been found. This is a countermeasure
10731 against active attacks where the attacker has to distinguish
10732 between bad padding and a MAC verification error. [CVE-2003-0078][]
10734 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10735 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10736 Martin Vuagnoux (EPFL, Ilion)*
10738 * Make the no-err option work as intended. The intention with no-err
10739 is not to have the whole error stack handling routines removed from
10740 libcrypto, it's only intended to remove all the function name and
10741 reason texts, thereby removing some of the footprint that may not
10742 be interesting if those errors aren't displayed anyway.
10744 NOTE: it's still possible for any application or module to have its
10745 own set of error texts inserted. The routines are there, just not
10746 used by default when no-err is given.
10750 * Add support for FreeBSD on IA64.
10752 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10754 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10755 Kerberos function mit_des_cbc_cksum(). Before this change,
10756 the value returned by DES_cbc_cksum() was like the one from
10757 mit_des_cbc_cksum(), except the bytes were swapped.
10759 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10761 * Allow an application to disable the automatic SSL chain building.
10762 Before this a rather primitive chain build was always performed in
10763 ssl3_output_cert_chain(): an application had no way to send the
10764 correct chain if the automatic operation produced an incorrect result.
10766 Now the chain builder is disabled if either:
10768 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10770 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10772 The reasoning behind this is that an application would not want the
10773 auto chain building to take place if extra chain certificates are
10774 present and it might also want a means of sending no additional
10775 certificates (for example the chain has two certificates and the
10780 * Add the possibility to build without the ENGINE framework.
10782 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10784 * Under Win32 gmtime() can return NULL: check return value in
10785 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10789 * DSA routines: under certain error conditions uninitialized BN objects
10790 could be freed. Solution: make sure initialization is performed early
10791 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10792 Nils Larsch <nla@trustcenter.de> via PR#459)
10796 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10797 checked on reconnect on the client side, therefore session resumption
10798 could still fail with a "ssl session id is different" error. This
10799 behaviour is masked when SSL_OP_ALL is used due to
10800 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10801 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10802 followup to PR #377.
10806 * IA-32 assembler support enhancements: unified ELF targets, support
10807 for SCO/Caldera platforms, fix for Cygwin shared build.
10811 * Add support for FreeBSD on sparc64. As a consequence, support for
10812 FreeBSD on non-x86 processors is separate from x86 processors on
10813 the config script, much like the NetBSD support.
10815 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10817 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002] ###
10819 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10822 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10823 code (06) was taken as the first octet of the session ID and the last
10824 octet was ignored consequently. As a result SSLv2 client side session
10825 caching could not have worked due to the session ID mismatch between
10827 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10832 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10833 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10838 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10839 seems that in spite of existing for more than a year, many application
10840 author have done nothing to provide the necessary callbacks, which
10841 means that this particular engine will not work properly anywhere.
10842 This is a very unfortunate situation which forces us, in the name
10843 of usability, to give the hw_ncipher.c a static lock, which is part
10845 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10846 appear in 0.9.8 or later. We EXPECT application authors to have
10847 dealt properly with this when 0.9.8 is released (unless we actually
10848 make such changes in the libcrypto locking code that changes will
10849 have to be made anyway).
10853 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
10854 octets have been read, EOF or an error occurs. Without this change
10855 some truncated ASN1 structures will not produce an error.
10859 * Disable Heimdal support, since it hasn't been fully implemented.
10860 Still give the possibility to force the use of Heimdal, but with
10861 warnings and a request that patches get sent to openssl-dev.
10865 * Add the VC-CE target, introduce the WINCE sysname, and add
10866 INSTALL.WCE and appropriate conditionals to make it build.
10868 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10870 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
10871 cygssl-x.y.z.dll, where x, y and z are the major, minor and
10872 edit numbers of the version.
10874 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10876 * Introduce safe string copy and catenation functions
10877 (BUF_strlcpy() and BUF_strlcat()).
10879 *Ben Laurie (CHATS) and Richard Levitte*
10881 * Avoid using fixed-size buffers for one-line DNs.
10883 *Ben Laurie (CHATS)*
10885 * Add BUF_MEM_grow_clean() to avoid information leakage when
10886 resizing buffers containing secrets, and use where appropriate.
10888 *Ben Laurie (CHATS)*
10890 * Avoid using fixed size buffers for configuration file location.
10892 *Ben Laurie (CHATS)*
10894 * Avoid filename truncation for various CA files.
10896 *Ben Laurie (CHATS)*
10898 * Use sizeof in preference to magic numbers.
10900 *Ben Laurie (CHATS)*
10902 * Avoid filename truncation in cert requests.
10904 *Ben Laurie (CHATS)*
10906 * Add assertions to check for (supposedly impossible) buffer
10909 *Ben Laurie (CHATS)*
10911 * Don't cache truncated DNS entries in the local cache (this could
10912 potentially lead to a spoofing attack).
10914 *Ben Laurie (CHATS)*
10916 * Fix various buffers to be large enough for hex/decimal
10917 representations in a platform independent manner.
10919 *Ben Laurie (CHATS)*
10921 * Add CRYPTO_realloc_clean() to avoid information leakage when
10922 resizing buffers containing secrets, and use where appropriate.
10924 *Ben Laurie (CHATS)*
10926 * Add BIO_indent() to avoid much slightly worrying code to do
10929 *Ben Laurie (CHATS)*
10931 * Convert sprintf()/BIO_puts() to BIO_printf().
10933 *Ben Laurie (CHATS)*
10935 * buffer_gets() could terminate with the buffer only half
10938 *Ben Laurie (CHATS)*
10940 * Add assertions to prevent user-supplied crypto functions from
10941 overflowing internal buffers by having large block sizes, etc.
10943 *Ben Laurie (CHATS)*
10945 * New OPENSSL_assert() macro (similar to assert(), but enabled
10948 *Ben Laurie (CHATS)*
10950 * Eliminate unused copy of key in RC4.
10952 *Ben Laurie (CHATS)*
10954 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
10956 *Ben Laurie (CHATS)*
10958 * Fix off-by-one error in EGD path.
10960 *Ben Laurie (CHATS)*
10962 * If RANDFILE path is too long, ignore instead of truncating.
10964 *Ben Laurie (CHATS)*
10966 * Eliminate unused and incorrectly sized X.509 structure
10969 *Ben Laurie (CHATS)*
10971 * Eliminate unused and dangerous function knumber().
10973 *Ben Laurie (CHATS)*
10975 * Eliminate unused and dangerous structure, KSSL_ERR.
10977 *Ben Laurie (CHATS)*
10979 * Protect against overlong session ID context length in an encoded
10980 session object. Since these are local, this does not appear to be
10983 *Ben Laurie (CHATS)*
10985 * Change from security patch (see 0.9.6e below) that did not affect
10986 the 0.9.6 release series:
10988 Remote buffer overflow in SSL3 protocol - an attacker could
10989 supply an oversized master key in Kerberos-enabled versions.
10992 *Ben Laurie (CHATS)*
10994 * Change the SSL kerb5 codes to match RFC 2712.
10998 * Make -nameopt work fully for req and add -reqopt switch.
11000 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11002 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11004 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11006 * Make sure tests can be performed even if the corresponding algorithms
11007 have been removed entirely. This was also the last step to make
11008 OpenSSL compilable with DJGPP under all reasonable conditions.
11010 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11012 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11013 to allow version independent disabling of normally unselected ciphers,
11014 which may be activated as a side-effect of selecting a single cipher.
11016 (E.g., cipher list string "RSA" enables ciphersuites that are left
11017 out of "ALL" because they do not provide symmetric encryption.
11018 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11020 *Lutz Jaenicke, Bodo Moeller*
11022 * Add appropriate support for separate platform-dependent build
11023 directories. The recommended way to make a platform-dependent
11024 build directory is the following (tested on Linux), maybe with
11027 # Place yourself outside of the OpenSSL source tree. In
11028 # this example, the environment variable OPENSSL_SOURCE
11029 # is assumed to contain the absolute OpenSSL source directory.
11030 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11031 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11032 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11033 mkdir -p `dirname $F`
11034 ln -s $OPENSSL_SOURCE/$F $F
11037 To be absolutely sure not to disturb the source tree, a "make clean"
11038 is a good thing. If it isn't successful, don't worry about it,
11039 it probably means the source directory is very clean.
11043 * Make sure any ENGINE control commands make local copies of string
11044 pointers passed to them whenever necessary. Otherwise it is possible
11045 the caller may have overwritten (or deallocated) the original string
11046 data when a later ENGINE operation tries to use the stored values.
11048 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11050 * Improve diagnostics in file reading and command-line digests.
11052 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11054 * Add AES modes CFB and OFB to the object database. Correct an
11055 error in AES-CFB decryption.
11059 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11060 allows existing EVP_CIPHER_CTX structures to be reused after
11061 calling EVP_*Final(). This behaviour is used by encryption
11062 BIOs and some applications. This has the side effect that
11063 applications must explicitly clean up cipher contexts with
11064 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11068 * Check the values of dna and dnb in bn_mul_recursive before calling
11069 bn_mul_comba (a non zero value means the a or b arrays do not contain
11070 n2 elements) and fallback to bn_mul_normal if either is not zero.
11074 * Fix escaping of non-ASCII characters when using the -subj option
11075 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11079 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11080 form for "surname", serialNumber has no short form.
11081 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11082 therefore remove "mail" short name for "internet 7".
11083 The OID for unique identifiers in X509 certificates is
11084 x500UniqueIdentifier, not uniqueIdentifier.
11085 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11089 * Add an "init" command to the ENGINE config module and auto initialize
11090 ENGINEs. Without any "init" command the ENGINE will be initialized
11091 after all ctrl commands have been executed on it. If init=1 the
11092 ENGINE is initialized at that point (ctrls before that point are run
11093 on the uninitialized ENGINE and after on the initialized one). If
11094 init=0 then the ENGINE will not be initialized at all.
11098 * Fix the 'app_verify_callback' interface so that the user-defined
11099 argument is actually passed to the callback: In the
11100 SSL_CTX_set_cert_verify_callback() prototype, the callback
11101 declaration has been changed from
11104 int (*cb)(X509_STORE_CTX *,void *);
11105 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11106 i=s->ctx->app_verify_callback(&ctx)
11107 has been changed into
11108 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11110 To update applications using SSL_CTX_set_cert_verify_callback(),
11111 a dummy argument can be added to their callback functions.
11113 *D. K. Smetters <smetters@parc.xerox.com>*
11115 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11117 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11119 * Add and OPENSSL_LOAD_CONF define which will cause
11120 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11121 This allows older applications to transparently support certain
11122 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11123 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11124 load the config file and OPENSSL_add_all_algorithms_conf() which will
11125 always load it have also been added.
11129 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11130 Adjust NIDs and EVP layer.
11132 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11134 * Config modules support in openssl utility.
11136 Most commands now load modules from the config file,
11137 though in a few (such as version) this isn't done
11138 because it couldn't be used for anything.
11140 In the case of ca and req the config file used is
11141 the same as the utility itself: that is the -config
11142 command line option can be used to specify an
11147 * Move default behaviour from OPENSSL_config(). If appname is NULL
11148 use "openssl_conf" if filename is NULL use default openssl config file.
11152 * Add an argument to OPENSSL_config() to allow the use of an alternative
11153 config section name. Add a new flag to tolerate a missing config file
11154 and move code to CONF_modules_load_file().
11158 * Support for crypto accelerator cards from Accelerated Encryption
11159 Processing, www.aep.ie. (Use engine 'aep')
11160 The support was copied from 0.9.6c [engine] and adapted/corrected
11161 to work with the new engine framework.
11163 *AEP Inc. and Richard Levitte*
11165 * Support for SureWare crypto accelerator cards from Baltimore
11166 Technologies. (Use engine 'sureware')
11167 The support was copied from 0.9.6c [engine] and adapted
11168 to work with the new engine framework.
11172 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11173 make the newer ENGINE framework commands for the CHIL engine work.
11175 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11177 * Make it possible to produce shared libraries on ReliantUNIX.
11179 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11181 * Add the configuration target debug-linux-ppro.
11182 Make 'openssl rsa' use the general key loading routines
11183 implemented in apps.c, and make those routines able to
11184 handle the key format FORMAT_NETSCAPE and the variant
11187 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11189 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11191 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11193 * Add -keyform to rsautl, and document -engine.
11195 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11197 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11198 BIO_R_NO_SUCH_FILE error code rather than the generic
11199 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11203 * Add new functions
11204 ERR_peek_last_error
11205 ERR_peek_last_error_line
11206 ERR_peek_last_error_line_data.
11207 These are similar to
11209 ERR_peek_error_line
11210 ERR_peek_error_line_data,
11211 but report on the latest error recorded rather than the first one
11212 still in the error queue.
11214 *Ben Laurie, Bodo Moeller*
11216 * default_algorithms option in ENGINE config module. This allows things
11218 default_algorithms = ALL
11219 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11223 * Preliminary ENGINE config module.
11227 * New experimental application configuration code.
11231 * Change the AES code to follow the same name structure as all other
11232 symmetric ciphers, and behave the same way. Move everything to
11233 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11235 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11237 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11239 *Ben Laurie and Theo de Raadt*
11241 * Add option to output public keys in req command.
11243 *Massimiliano Pala madwolf@openca.org*
11245 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11246 (up to about 10% better than before for P-192 and P-224).
11250 * New functions/macros
11252 SSL_CTX_set_msg_callback(ctx, cb)
11253 SSL_CTX_set_msg_callback_arg(ctx, arg)
11254 SSL_set_msg_callback(ssl, cb)
11255 SSL_set_msg_callback_arg(ssl, arg)
11257 to request calling a callback function
11259 void cb(int write_p, int version, int content_type,
11260 const void *buf, size_t len, SSL *ssl, void *arg)
11262 whenever a protocol message has been completely received
11263 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11264 protocol version according to which the SSL library interprets
11265 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11266 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11267 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11268 specification (change_cipher_spec(20), alert(21), handshake(22)).
11269 'buf' and 'len' point to the actual message, 'ssl' to the
11270 SSL object, and 'arg' is the application-defined value set by
11271 SSL[_CTX]_set_msg_callback_arg().
11273 'openssl s_client' and 'openssl s_server' have new '-msg' options
11274 to enable a callback that displays all protocol messages.
11278 * Change the shared library support so shared libraries are built as
11279 soon as the corresponding static library is finished, and thereby get
11280 openssl and the test programs linked against the shared library.
11281 This still only happens when the keyword "shard" has been given to
11282 the configuration scripts.
11284 NOTE: shared library support is still an experimental thing, and
11285 backward binary compatibility is still not guaranteed.
11287 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11289 * Add support for Subject Information Access extension.
11291 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11293 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11294 additional bytes when new memory had to be allocated, not just
11295 when reusing an existing buffer.
11299 * New command line and configuration option 'utf8' for the req command.
11300 This allows field values to be specified as UTF8 strings.
11304 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11305 runs for the former and machine-readable output for the latter.
11309 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11310 of the e-mail address in the DN (i.e., it will go into a certificate
11311 extension only). The new configuration file option 'email_in_dn = no'
11312 has the same effect.
11314 *Massimiliano Pala madwolf@openca.org*
11316 * Change all functions with names starting with des_ to be starting
11317 with DES_ instead. Add wrappers that are compatible with libdes,
11318 but are named _ossl_old_des_*. Finally, add macros that map the
11319 des_* symbols to the corresponding _ossl_old_des_* if libdes
11320 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11321 desired, the des_* symbols will be mapped to DES_*, with one
11324 Since we provide two compatibility mappings, the user needs to
11325 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11326 compatibility is desired. The default (i.e., when that macro
11327 isn't defined) is OpenSSL 0.9.6c compatibility.
11329 There are also macros that enable and disable the support of old
11330 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11331 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11332 are defined, the default will apply: to support the old des routines.
11334 In either case, one must include openssl/des.h to get the correct
11335 definitions. Do not try to just include openssl/des_old.h, that
11338 NOTE: This is a major break of an old API into a new one. Software
11339 authors are encouraged to switch to the DES_ style functions. Some
11340 time in the future, des_old.h and the libdes compatibility functions
11341 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11342 default), and then completely removed.
11346 * Test for certificates which contain unsupported critical extensions.
11347 If such a certificate is found during a verify operation it is
11348 rejected by default: this behaviour can be overridden by either
11349 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11350 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11351 X509_supported_extension() has also been added which returns 1 if a
11352 particular extension is supported.
11356 * Modify the behaviour of EVP cipher functions in similar way to digests
11357 to retain compatibility with existing code.
11361 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11362 compatibility with existing code. In particular the 'ctx' parameter does
11363 not have to be to be initialized before the call to EVP_DigestInit() and
11364 it is tidied up after a call to EVP_DigestFinal(). New function
11365 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11366 EVP_MD_CTX_copy() changed to not require the destination to be
11367 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11368 requires the destination to be valid.
11370 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11371 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11375 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11376 so that complete 'Handshake' protocol structures are kept in memory
11377 instead of overwriting 'msg_type' and 'length' with 'body' data.
11381 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11383 *Massimo Santin via Richard Levitte*
11385 * Major restructuring to the underlying ENGINE code. This includes
11386 reduction of linker bloat, separation of pure "ENGINE" manipulation
11387 (initialisation, etc) from functionality dealing with implementations
11388 of specific crypto interfaces. This change also introduces integrated
11389 support for symmetric ciphers and digest implementations - so ENGINEs
11390 can now accelerate these by providing EVP_CIPHER and EVP_MD
11391 implementations of their own. This is detailed in crypto/engine/README
11392 as it couldn't be adequately described here. However, there are a few
11393 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11394 were changed in the original introduction of ENGINE code have now
11395 reverted back - the hooking from this code to ENGINE is now a good
11396 deal more passive and at run-time, operations deal directly with
11397 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11398 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11399 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
11400 they were not being used by the framework as there is no concept of a
11401 BIGNUM_METHOD and they could not be generalised to the new
11402 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11403 ENGINE_cpy() has been removed as it cannot be consistently defined in
11408 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11412 * Change mkdef.pl to sort symbols that get the same entry number,
11413 and make sure the automatically generated functions ERR_load_*
11414 become part of libeay.num as well.
11418 * New function SSL_renegotiate_pending(). This returns true once
11419 renegotiation has been requested (either SSL_renegotiate() call
11420 or HelloRequest/ClientHello received from the peer) and becomes
11421 false once a handshake has been completed.
11422 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11423 sends a HelloRequest, but does not ensure that a handshake takes
11424 place. SSL_renegotiate_pending() is useful for checking if the
11425 client has followed the request.)
11429 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11430 By default, clients may request session resumption even during
11431 renegotiation (if session ID contexts permit); with this option,
11432 session resumption is possible only in the first handshake.
11434 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11435 more bits available for options that should not be part of
11436 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11440 * Add some demos for certificate and certificate request creation.
11444 * Make maximum certificate chain size accepted from the peer application
11445 settable (SSL*_get/set_max_cert_list()), as proposed by
11446 "Douglas E. Engert" <deengert@anl.gov>.
11450 * Add support for shared libraries for Unixware-7
11451 (Boyd Lynn Gerber <gerberb@zenez.com>).
11455 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11456 be done prior to destruction. Use this to unload error strings from
11457 ENGINEs that load their own error strings. NB: This adds two new API
11458 functions to "get" and "set" this destroy handler in an ENGINE.
11462 * Alter all existing ENGINE implementations (except "openssl" and
11463 "openbsd") to dynamically instantiate their own error strings. This
11464 makes them more flexible to be built both as statically-linked ENGINEs
11465 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11466 Also, add stub code to each that makes building them as self-contained
11467 shared-libraries easier (see README.ENGINE).
11471 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11472 implementations into applications that are completely implemented in
11473 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11474 commands that can be used to configure what shared-library to load and
11475 to control aspects of the way it is handled. Also, made an update to
11476 the README.ENGINE file that brings its information up-to-date and
11477 provides some information and instructions on the "dynamic" ENGINE
11478 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11482 * Make it possible to unload ranges of ERR strings with a new
11483 "ERR_unload_strings" function.
11487 * Add a copy() function to EVP_MD.
11491 * Make EVP_MD routines take a context pointer instead of just the
11492 md_data void pointer.
11496 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11497 that the digest can only process a single chunk of data
11498 (typically because it is provided by a piece of
11499 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11500 is only going to provide a single chunk of data, and hence the
11501 framework needn't accumulate the data for oneshot drivers.
11505 * As with "ERR", make it possible to replace the underlying "ex_data"
11506 functions. This change also alters the storage and management of global
11507 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11508 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11509 index counters. The API functions that use this state have been changed
11510 to take a "class_index" rather than pointers to the class's local STACK
11511 and counter, and there is now an API function to dynamically create new
11512 classes. This centralisation allows us to (a) plug a lot of the
11513 thread-safety problems that existed, and (b) makes it possible to clean
11514 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11515 such data would previously have always leaked in application code and
11516 workarounds were in place to make the memory debugging turn a blind eye
11517 to it. Application code that doesn't use this new function will still
11518 leak as before, but their memory debugging output will announce it now
11519 rather than letting it slide.
11521 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11522 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11523 has a return value to indicate success or failure.
11527 * Make it possible to replace the underlying "ERR" functions such that the
11528 global state (2 LHASH tables and 2 locks) is only used by the "default"
11529 implementation. This change also adds two functions to "get" and "set"
11530 the implementation prior to it being automatically set the first time
11531 any other ERR function takes place. Ie. an application can call "get",
11532 pass the return value to a module it has just loaded, and that module
11533 can call its own "set" function using that value. This means the
11534 module's "ERR" operations will use (and modify) the error state in the
11535 application and not in its own statically linked copy of OpenSSL code.
11539 * Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
11540 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11541 the operation, and provides a more encapsulated way for external code
11542 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11543 to use these functions rather than manually incrementing the counts.
11545 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11549 * Add EVP test program.
11553 * Add symmetric cipher support to ENGINE. Expect the API to change!
11557 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11558 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11559 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11560 These allow a CRL to be built without having to access X509_CRL fields
11561 directly. Modify 'ca' application to use new functions.
11565 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11566 bug workarounds. Rollback attack detection is a security feature.
11567 The problem will only arise on OpenSSL servers when TLSv1 is not
11568 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11569 Software authors not wanting to support TLSv1 will have special reasons
11570 for their choice and can explicitly enable this option.
11572 *Bodo Moeller, Lutz Jaenicke*
11574 * Rationalise EVP so it can be extended: don't include a union of
11575 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11576 (similar to those existing for EVP_CIPHER_CTX).
11581 EVP_MD_CTX_init(&md); /* new function call */
11582 EVP_DigestInit(&md, EVP_sha1());
11583 EVP_DigestUpdate(&md, in, len);
11584 EVP_DigestFinal(&md, out, NULL);
11585 EVP_MD_CTX_cleanup(&md); /* new function call */
11590 * Make DES key schedule conform to the usual scheme, as well as
11591 correcting its structure. This means that calls to DES functions
11592 now have to pass a pointer to a des_key_schedule instead of a
11593 plain des_key_schedule (which was actually always a pointer
11596 des_key_schedule ks;
11598 des_set_key_checked(..., &ks);
11599 des_ncbc_encrypt(..., &ks, ...);
11601 (Note that a later change renames 'des_...' into 'DES_...'.)
11605 * Initial reduction of linker bloat: the use of some functions, such as
11606 PEM causes large amounts of unused functions to be linked in due to
11607 poor organisation. For example pem_all.c contains every PEM function
11608 which has a knock on effect of linking in large amounts of (unused)
11609 ASN1 code. Grouping together similar functions and splitting unrelated
11610 functions prevents this.
11614 * Cleanup of EVP macros.
11618 * Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
11619 correct _ecb suffix.
11623 * Add initial OCSP responder support to ocsp application. The
11624 revocation information is handled using the text based index
11625 use by the ca application. The responder can either handle
11626 requests generated internally, supplied in files (for example
11627 via a CGI script) or using an internal minimal server.
11631 * Add configuration choices to get zlib compression for TLS.
11635 * Changes to Kerberos SSL for RFC 2712 compliance:
11636 1. Implemented real KerberosWrapper, instead of just using
11637 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11638 2. Implemented optional authenticator field of KerberosWrapper.
11640 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11641 and authenticator structs; see crypto/krb5/.
11643 Generalized Kerberos calls to support multiple Kerberos libraries.
11644 *Vern Staats <staatsvr@asc.hpc.mil>,
11645 Jeffrey Altman <jaltman@columbia.edu>
11646 via Richard Levitte*
11648 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11649 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11650 values for each of the key sizes rather than having just
11651 parameters (and 'speed' generating keys each time).
11655 * Speed up EVP routines.
11658 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11659 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11660 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11661 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11663 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11664 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11665 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11668 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11670 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11674 * Added the OS2-EMX target.
11676 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11678 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11679 to support NCONF routines in extension code. New function CONF_set_nconf()
11680 to allow functions which take an NCONF to also handle the old LHASH
11681 structure: this means that the old CONF compatible routines can be
11682 retained (in particular wrt extensions) without having to duplicate the
11683 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11687 * Enhance the general user interface with mechanisms for inner control
11688 and with possibilities to have yes/no kind of prompts.
11692 * Change all calls to low level digest routines in the library and
11693 applications to use EVP. Add missing calls to HMAC_cleanup() and
11694 don't assume HMAC_CTX can be copied using memcpy().
11696 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11698 * Add the possibility to control engines through control names but with
11699 arbitrary arguments instead of just a string.
11700 Change the key loaders to take a UI_METHOD instead of a callback
11701 function pointer. NOTE: this breaks binary compatibility with earlier
11702 versions of OpenSSL [engine].
11703 Adapt the nCipher code for these new conditions and add a card insertion
11708 * Enhance the general user interface with mechanisms to better support
11709 dialog box interfaces, application-defined prompts, the possibility
11710 to use defaults (for example default passwords from somewhere else)
11711 and interrupts/cancellations.
11715 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11716 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11720 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11721 tidy up some unnecessarily weird code in 'sk_new()').
11723 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11725 * Change the key loading routines for ENGINEs to use the same kind
11726 callback (pem_password_cb) as all other routines that need this
11731 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11732 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11733 than this minimum value is recommended.
11737 * New random seeder for OpenVMS, using the system process statistics
11738 that are easily reachable.
11742 * Windows apparently can't transparently handle global
11743 variables defined in DLLs. Initialisations such as:
11745 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11747 won't compile. This is used by the any applications that need to
11748 declare their own ASN1 modules. This was fixed by adding the option
11749 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11750 needed for static libraries under Win32.
11754 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11755 setting of purpose and trust fields. New X509_STORE trust and
11756 purpose functions and tidy up setting in other SSL functions.
11760 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11761 structure. These are inherited by X509_STORE_CTX when it is
11762 initialised. This allows various defaults to be set in the
11763 X509_STORE structure (such as flags for CRL checking and custom
11764 purpose or trust settings) for functions which only use X509_STORE_CTX
11765 internally such as S/MIME.
11767 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11768 trust settings if they are not set in X509_STORE. This allows X509_STORE
11769 purposes and trust (in S/MIME for example) to override any set by default.
11771 Add command line options for CRL checking to smime, s_client and s_server
11776 * Initial CRL based revocation checking. If the CRL checking flag(s)
11777 are set then the CRL is looked up in the X509_STORE structure and
11778 its validity and signature checked, then if the certificate is found
11779 in the CRL the verify fails with a revoked error.
11781 Various new CRL related callbacks added to X509_STORE_CTX structure.
11783 Command line options added to 'verify' application to support this.
11785 This needs some additional work, such as being able to handle multiple
11786 CRLs with different times, extension based lookup (rather than just
11787 by subject name) and ultimately more complete V2 CRL extension
11792 * Add a general user interface API (crypto/ui/). This is designed
11793 to replace things like des_read_password and friends (backward
11794 compatibility functions using this new API are provided).
11795 The purpose is to remove prompting functions from the DES code
11796 section as well as provide for prompting through dialog boxes in
11797 a window system and the like.
11801 * Add "ex_data" support to ENGINE so implementations can add state at a
11802 per-structure level rather than having to store it globally.
11806 * Make it possible for ENGINE structures to be copied when retrieved by
11807 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11808 This causes the "original" ENGINE structure to act like a template,
11809 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11810 operational state can be localised to each ENGINE structure, despite the
11811 fact they all share the same "methods". New ENGINE structures returned in
11812 this case have no functional references and the return value is the single
11813 structural reference. This matches the single structural reference returned
11814 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11819 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11820 needs to match any other type at all we need to manually clear the
11825 * Changes to the "openssl engine" utility to include;
11826 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11827 about an ENGINE's available control commands.
11828 - executing control commands from command line arguments using the
11829 '-pre' and '-post' switches. '-post' is only used if '-t' is
11830 specified and the ENGINE is successfully initialised. The syntax for
11831 the individual commands are colon-separated, for example;
11832 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11836 * New dynamic control command support for ENGINEs. ENGINEs can now
11837 declare their own commands (numbers), names (strings), descriptions,
11838 and input types for run-time discovery by calling applications. A
11839 subset of these commands are implicitly classed as "executable"
11840 depending on their input type, and only these can be invoked through
11841 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11842 can be based on user input, config files, etc). The distinction is
11843 that "executable" commands cannot return anything other than a boolean
11844 result and can only support numeric or string input, whereas some
11845 discoverable commands may only be for direct use through
11846 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11847 pointers, or other custom uses. The "executable" commands are to
11848 support parameterisations of ENGINE behaviour that can be
11849 unambiguously defined by ENGINEs and used consistently across any
11850 OpenSSL-based application. Commands have been added to all the
11851 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
11852 control over shared-library paths without source code alterations.
11856 * Changed all ENGINE implementations to dynamically allocate their
11857 ENGINEs rather than declaring them statically. Apart from this being
11858 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
11859 this also allows the implementations to compile without using the
11860 internal engine_int.h header.
11864 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
11865 'const' value. Any code that should be able to modify a RAND_METHOD
11866 should already have non-const pointers to it (ie. they should only
11867 modify their own ones).
11871 * Made a variety of little tweaks to the ENGINE code.
11872 - "atalla" and "ubsec" string definitions were moved from header files
11873 to C code. "nuron" string definitions were placed in variables
11874 rather than hard-coded - allowing parameterisation of these values
11875 later on via ctrl() commands.
11876 - Removed unused "#if 0"'d code.
11877 - Fixed engine list iteration code so it uses ENGINE_free() to release
11878 structural references.
11879 - Constified the RAND_METHOD element of ENGINE structures.
11880 - Constified various get/set functions as appropriate and added
11881 missing functions (including a catch-all ENGINE_cpy that duplicates
11882 all ENGINE values onto a new ENGINE except reference counts/state).
11883 - Removed NULL parameter checks in get/set functions. Setting a method
11884 or function to NULL is a way of cancelling out a previously set
11885 value. Passing a NULL ENGINE parameter is just plain stupid anyway
11886 and doesn't justify the extra error symbols and code.
11887 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
11888 flags from engine_int.h to engine.h.
11889 - Changed prototypes for ENGINE handler functions (init(), finish(),
11890 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
11894 * Implement binary inversion algorithm for BN_mod_inverse in addition
11895 to the algorithm using long division. The binary algorithm can be
11896 used only if the modulus is odd. On 32-bit systems, it is faster
11897 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
11898 roughly 5-15% for 256-bit moduli), so we use it only for moduli
11899 up to 450 bits. In 64-bit environments, the binary algorithm
11900 appears to be advantageous for much longer moduli; here we use it
11901 for moduli up to 2048 bits.
11905 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
11906 could not support the combine flag in choice fields.
11910 * Add a 'copy_extensions' option to the 'ca' utility. This copies
11911 extensions from a certificate request to the certificate.
11915 * Allow multiple 'certopt' and 'nameopt' options to be separated
11916 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
11917 file: this allows the display of the certificate about to be
11918 signed to be customised, to allow certain fields to be included
11919 or excluded and extension details. The old system didn't display
11920 multicharacter strings properly, omitted fields not in the policy
11921 and couldn't display additional details such as extensions.
11925 * Function EC_POINTs_mul for multiple scalar multiplication
11926 of an arbitrary number of elliptic curve points
11927 \sum scalars[i]*points[i],
11928 optionally including the generator defined for the EC_GROUP:
11929 scalar*generator + \sum scalars[i]*points[i].
11931 EC_POINT_mul is a simple wrapper function for the typical case
11932 that the point list has just one item (besides the optional
11937 * First EC_METHODs for curves over GF(p):
11939 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
11940 operations and provides various method functions that can also
11941 operate with faster implementations of modular arithmetic.
11943 EC_GFp_mont_method() reuses most functions that are part of
11944 EC_GFp_simple_method, but uses Montgomery arithmetic.
11946 *Bodo Moeller; point addition and point doubling
11947 implementation directly derived from source code provided by
11948 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
11950 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
11951 crypto/ec/ec_lib.c):
11953 Curves are EC_GROUP objects (with an optional group generator)
11954 based on EC_METHODs that are built into the library.
11956 Points are EC_POINT objects based on EC_GROUP objects.
11958 Most of the framework would be able to handle curves over arbitrary
11959 finite fields, but as there are no obvious types for fields other
11960 than GF(p), some functions are limited to that for now.
11964 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
11965 that the file contains a complete HTTP response.
11969 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
11970 change the def and num file printf format specifier from "%-40sXXX"
11971 to "%-39s XXX". The latter will always guarantee a space after the
11972 field while the former will cause them to run together if the field
11973 is 40 of more characters long.
11977 * Constify the cipher and digest 'method' functions and structures
11978 and modify related functions to take constant EVP_MD and EVP_CIPHER
11983 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
11984 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
11988 * Modify EVP_Digest*() routines so they now return values. Although the
11989 internal software routines can never fail additional hardware versions
11994 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
11996 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
11997 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12000 ERR_R_NESTED_ASN1_ERROR
12002 ERR_R_MISSING_ASN1_EOS
12003 were 4 .. 9, conflicting with
12004 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12006 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12007 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12009 Add new error code 'ERR_R_INTERNAL_ERROR'.
12013 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12018 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12019 sets the subject name for a new request or supersedes the
12020 subject name in a given request. Formats that can be parsed are
12021 'CN=Some Name, OU=myOU, C=IT'
12023 'CN=Some Name/OU=myOU/C=IT'.
12025 Add options '-batch' and '-verbose' to 'openssl req'.
12027 *Massimiliano Pala <madwolf@hackmasters.net>*
12029 * Introduce the possibility to access global variables through
12030 functions on platform were that's the best way to handle exporting
12031 global variables in shared libraries. To enable this functionality,
12032 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12033 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12034 is normally done by Configure or something similar).
12036 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12037 in the source file (foo.c) like this:
12039 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12040 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12042 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12043 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12045 OPENSSL_DECLARE_GLOBAL(int,foo);
12046 #define foo OPENSSL_GLOBAL_REF(foo)
12047 OPENSSL_DECLARE_GLOBAL(double,bar);
12048 #define bar OPENSSL_GLOBAL_REF(bar)
12050 The #defines are very important, and therefore so is including the
12051 header file everywhere where the defined globals are used.
12053 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12054 of ASN.1 items, but that structure is a bit different.
12056 The largest change is in util/mkdef.pl which has been enhanced with
12057 better and easier to understand logic to choose which symbols should
12058 go into the Windows .def files as well as a number of fixes and code
12059 cleanup (among others, algorithm keywords are now sorted
12060 lexicographically to avoid constant rewrites).
12064 * In BN_div() keep a copy of the sign of 'num' before writing the
12065 result to 'rm' because if rm==num the value will be overwritten
12066 and produce the wrong result if 'num' is negative: this caused
12067 problems with BN_mod() and BN_nnmod().
12071 * Function OCSP_request_verify(). This checks the signature on an
12072 OCSP request and verifies the signer certificate. The signer
12073 certificate is just checked for a generic purpose and OCSP request
12078 * Add OCSP_check_validity() function to check the validity of OCSP
12079 responses. OCSP responses are prepared in real time and may only
12080 be a few seconds old. Simply checking that the current time lies
12081 between thisUpdate and nextUpdate max reject otherwise valid responses
12082 caused by either OCSP responder or client clock inaccuracy. Instead
12083 we allow thisUpdate and nextUpdate to fall within a certain period of
12084 the current time. The age of the response can also optionally be
12085 checked. Two new options -validity_period and -status_age added to
12090 * If signature or public key algorithm is unrecognized print out its
12091 OID rather that just UNKNOWN.
12095 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12096 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12097 ID to be generated from the issuer certificate alone which can then be
12098 passed to OCSP_id_issuer_cmp().
12102 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12103 ASN1 modules to export functions returning ASN1_ITEM pointers
12104 instead of the ASN1_ITEM structures themselves. This adds several
12105 new macros which allow the underlying ASN1 function/structure to
12106 be accessed transparently. As a result code should not use ASN1_ITEM
12107 references directly (such as &X509_it) but instead use the relevant
12108 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12109 use of the new ASN1 code on platforms where exporting structures
12110 is problematical (for example in shared libraries) but exporting
12111 functions returning pointers to structures is not.
12115 * Add support for overriding the generation of SSL/TLS session IDs.
12116 These callbacks can be registered either in an SSL_CTX or per SSL.
12117 The purpose of this is to allow applications to control, if they wish,
12118 the arbitrary values chosen for use as session IDs, particularly as it
12119 can be useful for session caching in multiple-server environments. A
12120 command-line switch for testing this (and any client code that wishes
12121 to use such a feature) has been added to "s_server".
12123 *Geoff Thorpe, Lutz Jaenicke*
12125 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12126 of the form '#if defined(...) || defined(...) || ...' and
12127 '#if !defined(...) && !defined(...) && ...'. This also avoids
12128 the growing number of special cases it was previously handling.
12132 * Make all configuration macros available for application by making
12133 sure they are available in opensslconf.h, by giving them names starting
12134 with "OPENSSL_" to avoid conflicts with other packages and by making
12135 sure e_os2.h will cover all platform-specific cases together with
12137 Additionally, it is now possible to define configuration/platform-
12138 specific names (called "system identities"). In the C code, these
12139 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
12140 macro with the name beginning with "OPENSSL_SYS_", which is determined
12141 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
12146 * New option -set_serial to 'req' and 'x509' this allows the serial
12147 number to use to be specified on the command line. Previously self
12148 signed certificates were hard coded with serial number 0 and the
12149 CA options of 'x509' had to use a serial number in a file which was
12154 * New options to 'ca' utility to support V2 CRL entry extensions.
12155 Currently CRL reason, invalidity date and hold instruction are
12156 supported. Add new CRL extensions to V3 code and some new objects.
12160 * New function EVP_CIPHER_CTX_set_padding() this is used to
12161 disable standard block padding (aka PKCS#5 padding) in the EVP
12162 API, which was previously mandatory. This means that the data is
12163 not padded in any way and so the total length much be a multiple
12164 of the block size, otherwise an error occurs.
12168 * Initial (incomplete) OCSP SSL support.
12172 * New function OCSP_parse_url(). This splits up a URL into its host,
12173 port and path components: primarily to parse OCSP URLs. New -url
12174 option to ocsp utility.
12178 * New nonce behavior. The return value of OCSP_check_nonce() now
12179 reflects the various checks performed. Applications can decide
12180 whether to tolerate certain situations such as an absent nonce
12181 in a response when one was present in a request: the ocsp application
12182 just prints out a warning. New function OCSP_add1_basic_nonce()
12183 this is to allow responders to include a nonce in a response even if
12184 the request is nonce-less.
12188 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12189 skipped when using openssl x509 multiple times on a single input file,
12190 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
12194 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12195 set string type: to handle setting ASN1_TIME structures. Fix ca
12196 utility to correctly initialize revocation date of CRLs.
12200 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12201 the clients preferred ciphersuites and rather use its own preferences.
12202 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12203 Internet Explorer by ensuring unchanged hash method during stepup.
12204 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12208 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12209 to aes and add a new 'exist' option to print out symbols that don't
12214 * Additional options to ocsp utility to allow flags to be set and
12215 additional certificates supplied.
12219 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12220 OCSP client a number of certificate to only verify the response
12225 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12226 handle the new API. Currently only ECB, CBC modes supported. Add new
12229 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12230 Encryption Standard (AES) Ciphersuites for Transport Layer
12231 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12232 not enabled by default and were not part of the "ALL" ciphersuite
12233 alias because they were not yet official; they could be
12234 explicitly requested by specifying the "AESdraft" ciphersuite
12235 group alias. In the final release of OpenSSL 0.9.7, the group
12236 alias is called "AES" and is part of "ALL".)
12238 *Ben Laurie, Steve Henson, Bodo Moeller*
12240 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12241 request to response.
12245 * Functions for OCSP responders. OCSP_request_onereq_count(),
12246 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12247 extract information from a certificate request. OCSP_response_create()
12248 creates a response and optionally adds a basic response structure.
12249 OCSP_basic_add1_status() adds a complete single response to a basic
12250 response and returns the OCSP_SINGLERESP structure just added (to allow
12251 extensions to be included for example). OCSP_basic_add1_cert() adds a
12252 certificate to a basic response and OCSP_basic_sign() signs a basic
12253 response with various flags. New helper functions ASN1_TIME_check()
12254 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12255 (converts ASN1_TIME to GeneralizedTime).
12259 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12260 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12261 structure from a certificate. X509_pubkey_digest() digests the public_key
12262 contents: this is used in various key identifiers.
12266 * Make sk_sort() tolerate a NULL argument.
12268 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12270 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12271 passed by the function are trusted implicitly. If any of them signed the
12272 response then it is assumed to be valid and is not verified.
12276 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12277 to data. This was previously part of the PKCS7 ASN1 code. This
12278 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12279 *Steve Henson, reported by Kenneth R. Robinette
12280 <support@securenetterm.com>*
12282 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12283 routines: without these tracing memory leaks is very painful.
12284 Fix leaks in PKCS12 and PKCS7 routines.
12288 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12289 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12290 effectively meant GeneralizedTime would never be used. Now it
12291 is initialised to -1 but X509_time_adj() now has to check the value
12292 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12293 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12294 *Steve Henson, reported by Kenneth R. Robinette
12295 <support@securenetterm.com>*
12297 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12298 result in a zero length in the ASN1_INTEGER structure which was
12299 not consistent with the structure when d2i_ASN1_INTEGER() was used
12300 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12301 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12302 where it did not print out a minus for negative ASN1_INTEGER.
12306 * Add summary printout to ocsp utility. The various functions which
12307 convert status values to strings have been renamed to:
12308 OCSP_response_status_str(), OCSP_cert_status_str() and
12309 OCSP_crl_reason_str() and are no longer static. New options
12310 to verify nonce values and to disable verification. OCSP response
12311 printout format cleaned up.
12315 * Add additional OCSP certificate checks. These are those specified
12316 in RFC2560. This consists of two separate checks: the CA of the
12317 certificate being checked must either be the OCSP signer certificate
12318 or the issuer of the OCSP signer certificate. In the latter case the
12319 OCSP signer certificate must contain the OCSP signing extended key
12320 usage. This check is performed by attempting to match the OCSP
12321 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12322 in the OCSP_CERTID structures of the response.
12326 * Initial OCSP certificate verification added to OCSP_basic_verify()
12327 and related routines. This uses the standard OpenSSL certificate
12328 verify routines to perform initial checks (just CA validity) and
12329 to obtain the certificate chain. Then additional checks will be
12330 performed on the chain. Currently the root CA is checked to see
12331 if it is explicitly trusted for OCSP signing. This is used to set
12332 a root CA as a global signing root: that is any certificate that
12333 chains to that CA is an acceptable OCSP signing certificate.
12337 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12338 extensions from a separate configuration file.
12339 As when reading extensions from the main configuration file,
12340 the '-extensions ...' option may be used for specifying the
12343 *Massimiliano Pala <madwolf@comune.modena.it>*
12345 * New OCSP utility. Allows OCSP requests to be generated or
12346 read. The request can be sent to a responder and the output
12347 parsed, outputted or printed in text form. Not complete yet:
12348 still needs to check the OCSP response validity.
12352 * New subcommands for 'openssl ca':
12353 'openssl ca -status <serial>' prints the status of the cert with
12354 the given serial number (according to the index file).
12355 'openssl ca -updatedb' updates the expiry status of certificates
12358 *Massimiliano Pala <madwolf@comune.modena.it>*
12360 * New '-newreq-nodes' command option to CA.pl. This is like
12361 '-newreq', but calls 'openssl req' with the '-nodes' option
12362 so that the resulting key is not encrypted.
12364 *Damien Miller <djm@mindrot.org>*
12366 * New configuration for the GNU Hurd.
12368 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12370 * Initial code to implement OCSP basic response verify. This
12371 is currently incomplete. Currently just finds the signer's
12372 certificate and verifies the signature on the response.
12376 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12377 value of OPENSSLDIR. This is available via the new '-d' option
12378 to 'openssl version', and is also included in 'openssl version -a'.
12382 * Allowing defining memory allocation callbacks that will be given
12383 file name and line number information in additional arguments
12384 (a const char* and an int). The basic functionality remains, as
12385 well as the original possibility to just replace malloc(),
12386 realloc() and free() by functions that do not know about these
12387 additional arguments. To register and find out the current
12388 settings for extended allocation functions, the following
12389 functions are provided:
12391 CRYPTO_set_mem_ex_functions
12392 CRYPTO_set_locked_mem_ex_functions
12393 CRYPTO_get_mem_ex_functions
12394 CRYPTO_get_locked_mem_ex_functions
12396 These work the same way as CRYPTO_set_mem_functions and friends.
12397 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
12398 extended allocation function is enabled.
12399 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
12400 a conventional allocation function is enabled.
12402 *Richard Levitte, Bodo Moeller*
12404 * Finish off removing the remaining LHASH function pointer casts.
12405 There should no longer be any prototype-casting required when using
12406 the LHASH abstraction, and any casts that remain are "bugs". See
12407 the callback types and macros at the head of lhash.h for details
12408 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12412 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12413 If /dev/[u]random devices are not available or do not return enough
12414 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12416 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12417 /etc/entropy will be queried once each in this sequence, querying stops
12418 when enough entropy was collected without querying more sockets.
12422 * Change the Unix RAND_poll() variant to be able to poll several
12423 random devices, as specified by DEVRANDOM, until a sufficient amount
12424 of data has been collected. We spend at most 10 ms on each file
12425 (select timeout) and read in non-blocking mode. DEVRANDOM now
12426 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12427 (previously it was just the string "/dev/urandom"), so on typical
12428 platforms the 10 ms delay will never occur.
12429 Also separate out the Unix variant to its own file, rand_unix.c.
12430 For VMS, there's a currently-empty rand_vms.c.
12434 * Move OCSP client related routines to ocsp_cl.c. These
12435 provide utility functions which an application needing
12436 to issue a request to an OCSP responder and analyse the
12437 response will typically need: as opposed to those which an
12438 OCSP responder itself would need which will be added later.
12440 OCSP_request_sign() signs an OCSP request with an API similar
12441 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12442 response. OCSP_response_get1_basic() extracts basic response
12443 from response. OCSP_resp_find_status(): finds and extracts status
12444 information from an OCSP_CERTID structure (which will be created
12445 when the request structure is built). These are built from lower
12446 level functions which work on OCSP_SINGLERESP structures but
12447 won't normally be used unless the application wishes to examine
12448 extensions in the OCSP response for example.
12450 Replace nonce routines with a pair of functions.
12451 OCSP_request_add1_nonce() adds a nonce value and optionally
12452 generates a random value. OCSP_check_nonce() checks the
12453 validity of the nonce in an OCSP response.
12457 * Change function OCSP_request_add() to OCSP_request_add0_id().
12458 This doesn't copy the supplied OCSP_CERTID and avoids the
12459 need to free up the newly created id. Change return type
12460 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12461 This can then be used to add extensions to the request.
12462 Deleted OCSP_request_new(), since most of its functionality
12463 is now in OCSP_REQUEST_new() (and the case insensitive name
12464 clash) apart from the ability to set the request name which
12465 will be added elsewhere.
12469 * Update OCSP API. Remove obsolete extensions argument from
12470 various functions. Extensions are now handled using the new
12471 OCSP extension code. New simple OCSP HTTP function which
12472 can be used to send requests and parse the response.
12476 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12477 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12478 uses the special reorder version of SET OF to sort the attributes
12479 and reorder them to match the encoded order. This resolves a long
12480 standing problem: a verify on a PKCS7 structure just after signing
12481 it used to fail because the attribute order did not match the
12482 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12483 it uses the received order. This is necessary to tolerate some broken
12484 software that does not order SET OF. This is handled by encoding
12485 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12486 to produce the required SET OF.
12490 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12491 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12492 files to get correct declarations of the ASN.1 item variables.
12496 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12497 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12498 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12499 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12500 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12501 ASN1_ITEM and no wrapper functions.
12505 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12506 replace the old function pointer based I/O routines. Change most of
12507 the *_d2i_bio() and *_d2i_fp() functions to use these.
12511 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12512 lines, recognize more "algorithms" that can be deselected, and make
12513 it complain about algorithm deselection that isn't recognised.
12517 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12518 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12519 to use new functions. Add NO_ASN1_OLD which can be set to remove
12520 some old style ASN1 functions: this can be used to determine if old
12521 code will still work when these eventually go away.
12525 * New extension functions for OCSP structures, these follow the
12526 same conventions as certificates and CRLs.
12530 * New function X509V3_add1_i2d(). This automatically encodes and
12531 adds an extension. Its behaviour can be customised with various
12532 flags to append, replace or delete. Various wrappers added for
12533 certificates and CRLs.
12537 * Fix to avoid calling the underlying ASN1 print routine when
12538 an extension cannot be parsed. Correct a typo in the
12539 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12543 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12544 entries for variables.
12548 * Add functionality to apps/openssl.c for detecting locking
12549 problems: As the program is single-threaded, all we have
12550 to do is register a locking callback using an array for
12551 storing which locks are currently held by the program.
12555 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12556 SSL_get_ex_data_X509_STORE_idx(), which is used in
12557 ssl_verify_cert_chain() and thus can be called at any time
12558 during TLS/SSL handshakes so that thread-safety is essential.
12559 Unfortunately, the ex_data design is not at all suited
12560 for multi-threaded use, so it probably should be abolished.
12564 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12566 *Broadcom, tweaked and integrated by Geoff Thorpe*
12568 * Move common extension printing code to new function
12569 X509V3_print_extensions(). Reorganise OCSP print routines and
12570 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12574 * New function X509_signature_print() to remove duplication in some
12579 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12580 set (this was treated exactly the same as SET OF previously). This
12581 is used to reorder the STACK representing the structure to match the
12582 encoding. This will be used to get round a problem where a PKCS7
12583 structure which was signed could not be verified because the STACK
12584 order did not reflect the encoded order.
12588 * Reimplement the OCSP ASN1 module using the new code.
12592 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12593 for its ASN1 operations. The old style function pointers still exist
12594 for now but they will eventually go away.
12598 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12599 completely replaces the old ASN1 functionality with a table driven
12600 encoder and decoder which interprets an ASN1_ITEM structure describing
12601 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12602 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12603 has also been converted to the new form.
12607 * Change BN_mod_exp_recp so that negative moduli are tolerated
12608 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12609 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12610 for negative moduli.
12614 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12615 of not touching the result's sign bit.
12619 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12624 * Changed the LHASH code to use prototypes for callbacks, and created
12625 macros to declare and implement thin (optionally static) functions
12626 that provide type-safety and avoid function pointer casting for the
12627 type-specific callbacks.
12631 * Added Kerberos Cipher Suites to be used with TLS, as written in
12633 *Veers Staats <staatsvr@asc.hpc.mil>,
12634 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12636 * Reformat the FAQ so the different questions and answers can be divided
12637 in sections depending on the subject.
12641 * Have the zlib compression code load ZLIB.DLL dynamically under
12646 * New function BN_mod_sqrt for computing square roots modulo a prime
12647 (using the probabilistic Tonelli-Shanks algorithm unless
12648 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12649 be handled deterministically).
12651 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12653 * Make BN_mod_inverse faster by explicitly handling small quotients
12654 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12655 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12659 * New function BN_kronecker.
12663 * Fix BN_gcd so that it works on negative inputs; the result is
12664 positive unless both parameters are zero.
12665 Previously something reasonably close to an infinite loop was
12666 possible because numbers could be growing instead of shrinking
12667 in the implementation of Euclid's algorithm.
12671 * Fix BN_is_word() and BN_is_one() macros to take into account the
12672 sign of the number in question.
12674 Fix BN_is_word(a,w) to work correctly for w == 0.
12676 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12677 because its test if the absolute value of 'a' equals 'w'.
12678 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12679 it exists mostly for use in the implementations of BN_is_zero(),
12680 BN_is_one(), and BN_is_word().
12684 * New function BN_swap.
12688 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12689 the exponentiation functions are more likely to produce reasonable
12690 results on negative inputs.
12694 * Change BN_mod_mul so that the result is always non-negative.
12695 Previously, it could be negative if one of the factors was negative;
12696 I don't think anyone really wanted that behaviour.
12700 * Move BN_mod_... functions into new file crypto/bn/bn_mod.c
12701 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12702 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12703 and add new functions:
12712 BN_mod_lshift1_quick
12714 BN_mod_lshift_quick
12716 These functions always generate non-negative results.
12718 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12719 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12721 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12722 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12723 be reduced modulo m.
12725 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12728 The following entry accidentally appeared in the CHANGES file
12729 distributed with OpenSSL 0.9.7. The modifications described in
12730 it do *not* apply to OpenSSL 0.9.7.
12732 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12733 was actually never needed) and in BN_mul(). The removal in BN_mul()
12734 required a small change in bn_mul_part_recursive() and the addition
12735 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12736 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12737 bn_sub_words() and bn_add_words() except they take arrays with
12743 * In 'openssl passwd', verify passwords read from the terminal
12744 unless the '-salt' option is used (which usually means that
12745 verification would just waste user's time since the resulting
12746 hash is going to be compared with some given password hash)
12747 or the new '-noverify' option is used.
12749 This is an incompatible change, but it does not affect
12750 non-interactive use of 'openssl passwd' (passwords on the command
12751 line, '-stdin' option, '-in ...' option) and thus should not
12752 cause any problems.
12756 * Remove all references to RSAref, since there's no more need for it.
12760 * Make DSO load along a path given through an environment variable
12761 (SHLIB_PATH) with shl_load().
12765 * Constify the ENGINE code as a result of BIGNUM constification.
12766 Also constify the RSA code and most things related to it. In a
12767 few places, most notable in the depth of the ASN.1 code, ugly
12768 casts back to non-const were required (to be solved at a later
12773 * Make it so the openssl application has all engines loaded by default.
12777 * Constify the BIGNUM routines a little more.
12781 * Add the following functions:
12783 ENGINE_load_cswift()
12785 ENGINE_load_atalla()
12786 ENGINE_load_nuron()
12787 ENGINE_load_builtin_engines()
12789 That way, an application can itself choose if external engines that
12790 are built-in in OpenSSL shall ever be used or not. The benefit is
12791 that applications won't have to be linked with libdl or other dso
12792 libraries unless it's really needed.
12794 Changed 'openssl engine' to load all engines on demand.
12795 Changed the engine header files to avoid the duplication of some
12796 declarations (they differed!).
12800 * 'openssl engine' can now list capabilities.
12804 * Better error reporting in 'openssl engine'.
12808 * Never call load_dh_param(NULL) in s_server.
12812 * Add engine application. It can currently list engines by name and
12813 identity, and test if they are actually available.
12817 * Improve RPM specification file by forcing symbolic linking and making
12818 sure the installed documentation is also owned by root.root.
12820 *Damien Miller <djm@mindrot.org>*
12822 * Give the OpenSSL applications more possibilities to make use of
12823 keys (public as well as private) handled by engines.
12827 * Add OCSP code that comes from CertCo.
12831 * Add VMS support for the Rijndael code.
12835 * Added untested support for Nuron crypto accelerator.
12839 * Add support for external cryptographic devices. This code was
12840 previously distributed separately as the "engine" branch.
12842 *Geoff Thorpe, Richard Levitte*
12844 * Rework the filename-translation in the DSO code. It is now possible to
12845 have far greater control over how a "name" is turned into a filename
12846 depending on the operating environment and any oddities about the
12847 different shared library filenames on each system.
12851 * Support threads on FreeBSD-elf in Configure.
12855 * Fix for SHA1 assembly problem with MASM: it produces
12856 warnings about corrupt line number information when assembling
12857 with debugging information. This is caused by the overlapping
12860 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
12863 NCONF_get_number() has no error checking at all. As a replacement,
12864 NCONF_get_number_e() is defined (_e for "error checking") and is
12865 promoted strongly. The old NCONF_get_number is kept around for
12866 binary backward compatibility.
12867 Make it possible for methods to load from something other than a BIO,
12868 by providing a function pointer that is given a name instead of a BIO.
12869 For example, this could be used to load configuration data from an
12874 * Fix for non blocking accept BIOs. Added new I/O special reason
12875 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
12876 with non blocking I/O was not possible because no retry code was
12877 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
12882 * Added the beginnings of Rijndael support.
12886 * Fix for bug in DirectoryString mask setting. Add support for
12887 X509_NAME_print_ex() in 'req' and X509_print_ex() function
12888 to allow certificate printing to more controllable, additional
12889 'certopt' option to 'x509' to allow new printing options to be
12894 * Clean old EAY MD5 hack from e_os.h.
12898 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004] ###
12900 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12901 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
12903 *Joe Orton, Steve Henson*
12905 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003] ###
12907 * Fix additional bug revealed by the NISCC test suite:
12909 Stop bug triggering large recursion when presented with
12910 certain ASN.1 tags [CVE-2003-0851][]
12914 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003] ###
12916 * Fix various bugs revealed by running the NISCC test suite:
12918 Stop out of bounds reads in the ASN1 code when presented with
12919 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12921 If verify callback ignores invalid public key errors don't try to check
12922 certificate signature with the NULL public key.
12927 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12928 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12933 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12934 extra data after the compression methods not only for TLS 1.0
12935 but also for SSL 3.0 (as required by the specification).
12937 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12939 * Change X509_certificate_type() to mark the key as exported/exportable
12940 when it's 512 *bits* long, not 512 bytes.
12944 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003] ###
12946 * Countermeasure against the Klima-Pokorny-Rosa extension of
12947 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12948 a protocol version number mismatch like a decryption error
12949 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12953 * Turn on RSA blinding by default in the default implementation
12954 to avoid a timing attack. Applications that don't want it can call
12955 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12956 They would be ill-advised to do so in most cases.
12958 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12960 * Change RSA blinding code so that it works when the PRNG is not
12961 seeded (in this case, the secret RSA exponent is abused as
12962 an unpredictable seed -- if it is not unpredictable, there
12963 is no point in blinding anyway). Make RSA blinding thread-safe
12964 by remembering the creator's thread ID in rsa->blinding and
12965 having all other threads use local one-time blinding factors
12966 (this requires more computation than sharing rsa->blinding, but
12967 avoids excessive locking; and if an RSA object is not shared
12968 between threads, blinding will still be very fast).
12972 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003] ###
12974 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12975 via timing by performing a MAC computation even if incorrect
12976 block cipher padding has been found. This is a countermeasure
12977 against active attacks where the attacker has to distinguish
12978 between bad padding and a MAC verification error. [CVE-2003-0078][]
12980 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12981 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12982 Martin Vuagnoux (EPFL, Ilion)*
12984 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002] ###
12986 * New function OPENSSL_cleanse(), which is used to cleanse a section of
12987 memory from its contents. This is done with a counter that will
12988 place alternating values in each byte. This can be used to solve
12989 two issues: 1) the removal of calls to memset() by highly optimizing
12990 compilers, and 2) cleansing with other values than 0, since those can
12991 be read through on certain media, for example a swap space on disk.
12995 * Bugfix: client side session caching did not work with external caching,
12996 because the session->cipher setting was not restored when reloading
12997 from the external cache. This problem was masked, when
12998 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
12999 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13003 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13004 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13006 *Zeev Lieber <zeev-l@yahoo.com>*
13008 * Undo an undocumented change introduced in 0.9.6e which caused
13009 repeated calls to OpenSSL_add_all_ciphers() and
13010 OpenSSL_add_all_digests() to be ignored, even after calling
13015 * Change the default configuration reader to deal with last line not
13016 being properly terminated.
13020 * Change X509_NAME_cmp() so it applies the special rules on handling
13021 DN values that are of type PrintableString, as well as RDNs of type
13022 emailAddress where the value has the type ia5String.
13024 *stefank@valicert.com via Richard Levitte*
13026 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13027 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13028 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13029 the bitwise-OR of the two for use by the majority of applications
13030 wanting this behaviour, and update the docs. The documented
13031 behaviour and actual behaviour were inconsistent and had been
13032 changing anyway, so this is more a bug-fix than a behavioural
13035 *Geoff Thorpe, diagnosed by Nadav Har'El*
13037 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13038 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13042 * Fix initialization code race conditions in
13043 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13044 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13045 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13046 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13047 ssl2_get_cipher_by_char(),
13048 ssl3_get_cipher_by_char().
13050 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13052 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13053 the cached sessions are flushed, as the remove_cb() might use ex_data
13054 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13055 (see [openssl.org #212]).
13057 *Geoff Thorpe, Lutz Jaenicke*
13059 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13060 length, instead of the encoding length to d2i_ASN1_OBJECT.
13064 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002] ###
13066 * [In 0.9.6g-engine release:]
13067 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
13069 *Lynn Gazis <lgazis@rainbow.com>*
13071 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002] ###
13073 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13074 and get fix the header length calculation.
13075 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13076 Alon Kantor <alonk@checkpoint.com> (and others),
13079 * Use proper error handling instead of 'assertions' in buffer
13080 overflow checks added in 0.9.6e. This prevents DoS (the
13081 assertions could call abort()).
13083 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13085 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002] ###
13087 * Add various sanity checks to asn1_get_length() to reject
13088 the ASN1 length bytes if they exceed sizeof(long), will appear
13089 negative or the content length exceeds the length of the
13092 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13094 * Fix cipher selection routines: ciphers without encryption had no flags
13095 for the cipher strength set and where therefore not handled correctly
13096 by the selection routines (PR #130).
13100 * Fix EVP_dsa_sha macro.
13105 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13106 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13107 that was added in OpenSSL 0.9.6d.
13109 As the countermeasure turned out to be incompatible with some
13110 broken SSL implementations, the new option is part of SSL_OP_ALL.
13111 SSL_OP_ALL is usually employed when compatibility with weird SSL
13112 implementations is desired (e.g. '-bugs' option to 's_client' and
13113 's_server'), so the new option is automatically set in many
13118 * Changes in security patch:
13120 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13121 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13122 Air Force Materiel Command, USAF, under agreement number
13125 * Add various sanity checks to asn1_get_length() to reject
13126 the ASN1 length bytes if they exceed sizeof(long), will appear
13127 negative or the content length exceeds the length of the
13128 supplied buffer. [CVE-2002-0659][]
13130 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13132 * Assertions for various potential buffer overflows, not known to
13133 happen in practice.
13135 *Ben Laurie (CHATS)*
13137 * Various temporary buffers to hold ASCII versions of integers were
13138 too small for 64 bit platforms. [CVE-2002-0655][]
13139 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
13141 * Remote buffer overflow in SSL3 protocol - an attacker could
13142 supply an oversized session ID to a client. [CVE-2002-0656][]
13144 *Ben Laurie (CHATS)*
13146 * Remote buffer overflow in SSL2 protocol - an attacker could
13147 supply an oversized client master key. [CVE-2002-0656][]
13149 *Ben Laurie (CHATS)*
13151 ### Changes between 0.9.6c and 0.9.6d [9 May 2002] ###
13153 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13154 encoded as NULL) with id-dsa-with-sha1.
13156 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13158 * Check various X509_...() return values in apps/req.c.
13160 *Nils Larsch <nla@trustcenter.de>*
13162 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13163 an end-of-file condition would erroneously be flagged, when the CRLF
13164 was just at the end of a processed block. The bug was discovered when
13165 processing data through a buffering memory BIO handing the data to a
13166 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13167 <ptsekov@syntrex.com> and Nedelcho Stanev.
13171 * Implement a countermeasure against a vulnerability recently found
13172 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13173 before application data chunks to avoid the use of known IVs
13174 with data potentially chosen by the attacker.
13178 * Fix length checks in ssl3_get_client_hello().
13182 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13183 to prevent ssl3_read_internal() from incorrectly assuming that
13184 ssl3_read_bytes() found application data while handshake
13185 processing was enabled when in fact s->s3->in_read_app_data was
13186 merely automatically cleared during the initial handshake.
13188 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13190 * Fix object definitions for Private and Enterprise: they were not
13191 recognized in their shortname (=lowercase) representation. Extend
13192 obj_dat.pl to issue an error when using undefined keywords instead
13193 of silently ignoring the problem (Svenning Sorensen
13194 <sss@sss.dnsalias.net>).
13198 * Fix DH_generate_parameters() so that it works for 'non-standard'
13199 generators, i.e. generators other than 2 and 5. (Previously, the
13200 code did not properly initialise the 'add' and 'rem' values to
13201 BN_generate_prime().)
13203 In the new general case, we do not insist that 'generator' is
13204 actually a primitive root: This requirement is rather pointless;
13205 a generator of the order-q subgroup is just as good, if not
13210 * Map new X509 verification errors to alerts. Discovered and submitted by
13211 Tom Wu <tom@arcot.com>.
13215 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13216 returning non-zero before the data has been completely received
13217 when using non-blocking I/O.
13219 *Bodo Moeller; problem pointed out by John Hughes*
13221 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13223 *Ben Laurie, Lutz Jaenicke*
13225 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13226 Yoram Zahavi <YoramZ@gilian.com>).
13230 * Add information about CygWin 1.3 and on, and preserve proper
13231 configuration for the versions before that.
13233 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13235 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13236 check whether we deal with a copy of a session and do not delete from
13237 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13238 <izhar@checkpoint.com>.
13242 * Do not store session data into the internal session cache, if it
13243 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13244 flag is set). Proposed by Aslam <aslam@funk.com>.
13248 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13253 * [In 0.9.6d-engine release:]
13254 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13256 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13258 * Add the configuration target linux-s390x.
13260 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13262 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13263 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13264 variable as an indication that a ClientHello message has been
13265 received. As the flag value will be lost between multiple
13266 invocations of ssl3_accept when using non-blocking I/O, the
13267 function may not be aware that a handshake has actually taken
13268 place, thus preventing a new session from being added to the
13271 To avoid this problem, we now set s->new_session to 2 instead of
13272 using a local variable.
13274 *Lutz Jaenicke, Bodo Moeller*
13276 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13277 if the SSL_R_LENGTH_MISMATCH error is detected.
13279 *Geoff Thorpe, Bodo Moeller*
13281 * New 'shared_ldflag' column in Configure platform table.
13285 * Fix EVP_CIPHER_mode macro.
13287 *"Dan S. Camper" <dan@bti.net>*
13289 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13290 type, we must throw them away by setting rr->length to 0.
13292 *D P Chang <dpc@qualys.com>*
13294 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001] ###
13296 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13297 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13298 worked incorrectly for those cases where range = 10..._2 and
13299 3*range is two bits longer than range.)
13303 * Only add signing time to PKCS7 structures if it is not already
13308 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13309 OBJ_ld_ce should be OBJ_id_ce.
13310 Also some ip-pda OIDs in crypto/objects/objects.txt were
13311 incorrect (cf. RFC 3039).
13313 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13315 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13316 returns early because it has nothing to do.
13318 *Andy Schneider <andy.schneider@bjss.co.uk>*
13320 * [In 0.9.6c-engine release:]
13321 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13323 *Andy Schneider <andy.schneider@bjss.co.uk>*
13325 * [In 0.9.6c-engine release:]
13326 Add support for Cryptographic Appliance's keyserver technology.
13327 (Use engine 'keyclient')
13329 *Cryptographic Appliances and Geoff Thorpe*
13331 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13332 is called via tools/c89.sh because arguments have to be
13333 rearranged (all '-L' options must appear before the first object
13336 *Richard Shapiro <rshapiro@abinitio.com>*
13338 * [In 0.9.6c-engine release:]
13339 Add support for Broadcom crypto accelerator cards, backported
13342 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13344 * [In 0.9.6c-engine release:]
13345 Add support for SureWare crypto accelerator cards from
13346 Baltimore Technologies. (Use engine 'sureware')
13348 *Baltimore Technologies and Mark Cox*
13350 * [In 0.9.6c-engine release:]
13351 Add support for crypto accelerator cards from Accelerated
13352 Encryption Processing, www.aep.ie. (Use engine 'aep')
13354 *AEP Inc. and Mark Cox*
13356 * Add a configuration entry for gcc on UnixWare.
13358 *Gary Benson <gbenson@redhat.com>*
13360 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13361 messages are stored in a single piece (fixed-length part and
13362 variable-length part combined) and fix various bugs found on the way.
13366 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13367 instead. BIO_gethostbyname() does not know what timeouts are
13368 appropriate, so entries would stay in cache even when they have
13370 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
13372 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13373 faced with a pathologically small ClientHello fragment that does
13374 not contain client_version: Instead of aborting with an error,
13375 simply choose the highest available protocol version (i.e.,
13376 TLS 1.0 unless it is disabled). In practice, ClientHello
13377 messages are never sent like this, but this change gives us
13378 strictly correct behaviour at least for TLS.
13382 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13383 never resets s->method to s->ctx->method when called from within
13384 one of the SSL handshake functions.
13386 *Bodo Moeller; problem pointed out by Niko Baric*
13388 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13389 (sent using the client's version number) if client_version is
13390 smaller than the protocol version in use. Also change
13391 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13392 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13393 the client will at least see that alert.
13397 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13402 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13403 client receives HelloRequest while in a handshake.
13405 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13407 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13408 should end in 'break', not 'goto end' which circumvents various
13409 cleanups done in state SSL_ST_OK. But session related stuff
13410 must be disabled for SSL_ST_OK in the case that we just sent a
13413 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13414 before just sending a HelloRequest.
13416 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13418 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13419 reveal whether illegal block cipher padding was found or a MAC
13420 verification error occurred. (Neither SSLerr() codes nor alerts
13421 are directly visible to potential attackers, but the information
13422 may leak via logfiles.)
13424 Similar changes are not required for the SSL 2.0 implementation
13425 because the number of padding bytes is sent in clear for SSL 2.0,
13426 and the extra bytes are just ignored. However ssl/s2_pkt.c
13427 failed to verify that the purported number of padding bytes is in
13432 * Add OpenUNIX-8 support including shared libraries
13433 (Boyd Lynn Gerber <gerberb@zenez.com>).
13437 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13438 'wristwatch attack' using huge encoding parameters (cf.
13439 James H. Manger's CRYPTO 2001 paper). Note that the
13440 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13441 encoding parameters and hence was not vulnerable.
13445 * BN_sqr() bug fix.
13447 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13449 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13450 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13451 followed by modular reduction.
13453 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13455 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13456 equivalent based on BN_pseudo_rand() instead of BN_rand().
13460 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13461 This function was broken, as the check for a new client hello message
13462 to handle SGC did not allow these large messages.
13463 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13467 * Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
13471 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13472 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13476 * Rework the configuration and shared library support for Tru64 Unix.
13477 The configuration part makes use of modern compiler features and
13478 still retains old compiler behavior for those that run older versions
13479 of the OS. The shared library support part includes a variant that
13480 uses the RPATH feature, and is available through the special
13481 configuration target "alpha-cc-rpath", which will never be selected
13484 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13486 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13487 with the same message size as in ssl3_get_certificate_request().
13488 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13489 messages might inadvertently be reject as too long.
13491 *Petr Lampa <lampa@fee.vutbr.cz>*
13493 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13497 * Modified SSL library such that the verify_callback that has been set
13498 specifically for an SSL object with SSL_set_verify() is actually being
13499 used. Before the change, a verify_callback set with this function was
13500 ignored and the verify_callback() set in the SSL_CTX at the time of
13501 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13502 to allow the necessary settings.
13506 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13507 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13508 done automatically (in contradiction to the requirements of the C
13509 standard). This made problems when used from OpenSSH.
13513 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13514 dh->length and always used
13516 BN_rand_range(priv_key, dh->p).
13518 BN_rand_range() is not necessary for Diffie-Hellman, and this
13519 specific range makes Diffie-Hellman unnecessarily inefficient if
13520 dh->length (recommended exponent length) is much smaller than the
13521 length of dh->p. We could use BN_rand_range() if the order of
13522 the subgroup was stored in the DH structure, but we only have
13527 BN_rand(priv_key, l, ...)
13529 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13536 RSA_eay_public_encrypt
13537 RSA_eay_private_decrypt
13538 RSA_eay_private_encrypt (signing)
13539 RSA_eay_public_decrypt (signature verification)
13541 (default implementations for RSA_public_encrypt,
13542 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13543 always reject numbers >= n.
13547 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13548 to synchronize access to 'locking_thread'. This is necessary on
13549 systems where access to 'locking_thread' (an 'unsigned long'
13550 variable) is not atomic.
13554 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13555 *before* setting the 'crypto_lock_rand' flag. The previous code had
13556 a race condition if 0 is a valid thread ID.
13558 *Travis Vitek <vitek@roguewave.com>*
13560 * Add support for shared libraries under Irix.
13562 *Albert Chin-A-Young <china@thewrittenword.com>*
13564 * Add configuration option to build on Linux on both big-endian and
13565 little-endian MIPS.
13567 *Ralf Baechle <ralf@uni-koblenz.de>*
13569 * Add the possibility to create shared libraries on HP-UX.
13573 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001] ###
13575 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13576 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13577 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13578 PRNG state recovery was possible based on the output of
13579 one PRNG request appropriately sized to gain knowledge on
13580 'md' followed by enough consecutive 1-byte PRNG requests
13581 to traverse all of 'state'.
13583 1. When updating 'md_local' (the current thread's copy of 'md')
13584 during PRNG output generation, hash all of the previous
13585 'md_local' value, not just the half used for PRNG output.
13587 2. Make the number of bytes from 'state' included into the hash
13588 independent from the number of PRNG bytes requested.
13590 The first measure alone would be sufficient to avoid
13591 Markku-Juhani's attack. (Actually it had never occurred
13592 to me that the half of 'md_local' used for chaining was the
13593 half from which PRNG output bytes were taken -- I had always
13594 assumed that the secret half would be used.) The second
13595 measure makes sure that additional data from 'state' is never
13596 mixed into 'md_local' in small portions; this heuristically
13597 further strengthens the PRNG.
13601 * Fix crypto/bn/asm/mips3.s.
13605 * When only the key is given to "enc", the IV is undefined. Print out
13606 an error message in this case.
13610 * Handle special case when X509_NAME is empty in X509 printing routines.
13614 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13615 positive and less than q.
13619 * Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
13620 used: it isn't thread safe and the add_lock_callback should handle
13623 *Paul Rose <Paul.Rose@bridge.com>*
13625 * Verify that incoming data obeys the block size in
13626 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13632 *Ulf Möller, Bodo Möller*
13634 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13635 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13636 when fixing the server behaviour for backwards-compatible 'client
13637 hello' messages. (Note that the attack is impractical against
13638 SSL 3.0 and TLS 1.0 anyway because length and version checking
13639 means that the probability of guessing a valid ciphertext is
13640 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13643 Before 0.9.5, the countermeasure (hide the error by generating a
13644 random 'decryption result') did not work properly because
13645 ERR_clear_error() was missing, meaning that SSL_get_error() would
13646 detect the supposedly ignored error.
13648 Both problems are now fixed.
13652 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13653 (previously it was 1024).
13657 * Fix for compatibility mode trust settings: ignore trust settings
13658 unless some valid trust or reject settings are present.
13662 * Fix for blowfish EVP: its a variable length cipher.
13666 * Fix various bugs related to DSA S/MIME verification. Handle missing
13667 parameters in DSA public key structures and return an error in the
13668 DSA routines if parameters are absent.
13672 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13673 in the current directory if neither $RANDFILE nor $HOME was set.
13674 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13675 caused some confusion to Windows users who haven't defined $HOME.
13676 Thus RAND_file_name() is changed again: e_os.h can define a
13677 DEFAULT_HOME, which will be used if $HOME is not set.
13678 For Windows, we use "C:"; on other platforms, we still require
13679 environment variables.
13681 * Move 'if (!initialized) RAND_poll()' into regions protected by
13682 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13683 having multiple threads call RAND_poll() concurrently.
13687 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13688 combination of a flag and a thread ID variable.
13689 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13690 flag), *other* threads can enter ssleay_add_bytes without obeying
13691 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13692 that they do not hold after the first thread unsets add_do_not_lock).
13696 * Change bctest again: '-x' expressions are not available in all
13697 versions of 'test'.
13701 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001] ###
13703 * Fix a couple of memory leaks in PKCS7_dataDecode()
13705 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13707 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13708 the default extension for executables, if any. Also, make the perl
13709 scripts that use symlink() to test if it really exists and use "cp"
13710 if it doesn't. All this made OpenSSL compilable and installable in
13715 * Fix for asn1_GetSequence() for indefinite length constructed data.
13716 If SEQUENCE is length is indefinite just set c->slen to the total
13717 amount of data available.
13719 *Steve Henson, reported by shige@FreeBSD.org*
13721 *This change does not apply to 0.9.7.*
13723 * Change bctest to avoid here-documents inside command substitution
13724 (workaround for FreeBSD /bin/sh bug).
13725 For compatibility with Ultrix, avoid shell functions (introduced
13726 in the bctest version that searches along $PATH).
13730 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13731 with des_encrypt() defined on some operating systems, like Solaris
13736 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13737 On the Importance of Eliminating Errors in Cryptographic
13738 Computations, J. Cryptology 14 (2001) 2, 101-119,
13739 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
13743 * MIPS assembler BIGNUM division bug fix.
13747 * Disabled incorrect Alpha assembler code.
13751 * Fix PKCS#7 decode routines so they correctly update the length
13752 after reading an EOC for the EXPLICIT tag.
13756 *This change does not apply to 0.9.7.*
13758 * Fix bug in PKCS#12 key generation routines. This was triggered
13759 if a 3DES key was generated with a 0 initial byte. Include
13760 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13761 (but broken) behaviour.
13765 * Enhance bctest to search for a working bc along $PATH and print
13768 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13770 * Fix memory leaks in err.c: free err_data string if necessary;
13771 don't write to the wrong index in ERR_set_error_data.
13775 * Implement ssl23_peek (analogous to ssl23_read), which previously
13780 * Replace rdtsc with _emit statements for VC++ version 5.
13782 *Jeremy Cooper <jeremy@baymoo.org>*
13784 * Make it possible to reuse SSLv2 sessions.
13788 * In copy_email() check for >= 0 as a return value for
13789 X509_NAME_get_index_by_NID() since 0 is a valid index.
13791 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13793 * Avoid coredump with unsupported or invalid public keys by checking if
13794 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13795 PKCS7_verify() fails with non detached data.
13799 * Don't use getenv in library functions when run as setuid/setgid.
13800 New function OPENSSL_issetugid().
13804 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13805 due to incorrect handling of multi-threading:
13807 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13809 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13811 3. Count how many times MemCheck_off() has been called so that
13812 nested use can be treated correctly. This also avoids
13813 inband-signalling in the previous code (which relied on the
13814 assumption that thread ID 0 is impossible).
13818 * Add "-rand" option also to s_client and s_server.
13822 * Fix CPU detection on Irix 6.x.
13823 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13824 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13826 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13831 *This change does not apply to 0.9.7.*
13833 * Use the cached encoding of an X509_NAME structure rather than
13834 copying it. This is apparently the reason for the libsafe "errors"
13835 but the code is actually correct.
13839 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13840 Bleichenbacher's DSA attack.
13841 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13842 to be set and top=0 forces the highest bit to be set; top=-1 is new
13843 and leaves the highest bit random.
13845 *Ulf Moeller, Bodo Moeller*
13847 * In the NCONF_...-based implementations for CONF_... queries
13848 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13849 a temporary CONF structure with the data component set to NULL
13850 (which gives segmentation faults in lh_retrieve).
13851 Instead, use NULL for the CONF pointer in CONF_get_string and
13852 CONF_get_number (which may use environment variables) and directly
13853 return NULL from CONF_get_section.
13857 * Fix potential buffer overrun for EBCDIC.
13861 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
13862 keyUsage if basicConstraints absent for a CA.
13866 * Make SMIME_write_PKCS7() write mail header values with a format that
13867 is more generally accepted (no spaces before the semicolon), since
13868 some programs can't parse those values properly otherwise. Also make
13869 sure BIO's that break lines after each write do not create invalid
13874 * Make the CRL encoding routines work with empty SEQUENCE OF. The
13875 macros previously used would not encode an empty SEQUENCE OF
13876 and break the signature.
13880 *This change does not apply to 0.9.7.*
13882 * Zero the premaster secret after deriving the master secret in
13887 * Add some EVP_add_digest_alias registrations (as found in
13888 OpenSSL_add_all_digests()) to SSL_library_init()
13889 aka OpenSSL_add_ssl_algorithms(). This provides improved
13890 compatibility with peers using X.509 certificates
13891 with unconventional AlgorithmIdentifier OIDs.
13895 * Fix for Irix with NO_ASM.
13897 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13899 * ./config script fixes.
13901 *Ulf Moeller, Richard Levitte*
13903 * Fix 'openssl passwd -1'.
13907 * Change PKCS12_key_gen_asc() so it can cope with non null
13908 terminated strings whose length is passed in the passlen
13909 parameter, for example from PEM callbacks. This was done
13910 by adding an extra length parameter to asc2uni().
13912 *Steve Henson, reported by <oddissey@samsung.co.kr>*
13914 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
13915 call failed, free the DSA structure.
13919 * Fix to uni2asc() to cope with zero length Unicode strings.
13920 These are present in some PKCS#12 files.
13924 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
13925 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
13926 when writing a 32767 byte record.
13928 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
13930 * In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
13931 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
13933 (RSA objects have a reference count access to which is protected
13934 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
13935 so they are meant to be shared between threads.)
13936 *Bodo Moeller, Geoff Thorpe; original patch submitted by
13937 "Reddie, Steven" <Steven.Reddie@ca.com>*
13939 * Fix a deadlock in CRYPTO_mem_leaks().
13943 * Use better test patterns in bntest.
13947 * rand_win.c fix for Borland C.
13951 * BN_rshift bugfix for n == 0.
13955 * Add a 'bctest' script that checks for some known 'bc' bugs
13956 so that 'make test' does not abort just because 'bc' is broken.
13960 * Store verify_result within SSL_SESSION also for client side to
13961 avoid potential security hole. (Re-used sessions on the client side
13962 always resulted in verify_result==X509_V_OK, not using the original
13963 result of the server certificate verification.)
13967 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
13968 SSL3_RT_APPLICATION_DATA, return 0.
13969 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
13974 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
13975 releases, have been re-implemented by renaming the previous
13976 implementations of ssl2_read and ssl3_read to ssl2_read_internal
13977 and ssl3_read_internal, respectively, and adding 'peek' parameters
13978 to them. The new ssl[23]_{read,peek} functions are calls to
13979 ssl[23]_read_internal with the 'peek' flag set appropriately.
13980 A 'peek' parameter has also been added to ssl3_read_bytes, which
13981 does the actual work for ssl3_read_internal.
13985 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
13986 the method-specific "init()" handler. Also clean up ex_data after
13987 calling the method-specific "finish()" handler. Previously, this was
13988 happening the other way round.
13992 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
13993 The previous value, 12, was not always sufficient for BN_mod_exp().
13997 * Make sure that shared libraries get the internal name engine with
13998 the full version number and not just 0. This should mark the
13999 shared libraries as not backward compatible. Of course, this should
14000 be changed again when we can guarantee backward binary compatibility.
14004 * Fix typo in get_cert_by_subject() in by_dir.c
14006 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14008 * Rework the system to generate shared libraries:
14010 - Make note of the expected extension for the shared libraries and
14011 if there is a need for symbolic links from for example libcrypto.so.0
14012 to libcrypto.so.0.9.7. There is extended info in Configure for
14015 - Make as few rebuilds of the shared libraries as possible.
14017 - Still avoid linking the OpenSSL programs with the shared libraries.
14019 - When installing, install the shared libraries separately from the
14024 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14026 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14027 and not in SSL_clear because the latter is also used by the
14028 accept/connect functions; previously, the settings made by
14029 SSL_set_read_ahead would be lost during the handshake.
14031 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14033 * Correct util/mkdef.pl to be selective about disabled algorithms.
14034 Previously, it would create entries for disabled algorithms no
14039 * Added several new manual pages for SSL_* function.
14043 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000] ###
14045 * In ssl23_get_client_hello, generate an error message when faced
14046 with an initial SSL 3.0/TLS record that is too small to contain the
14047 first two bytes of the ClientHello message, i.e. client_version.
14048 (Note that this is a pathologic case that probably has never happened
14049 in real life.) The previous approach was to use the version number
14050 from the record header as a substitute; but our protocol choice
14051 should not depend on that one because it is not authenticated
14052 by the Finished messages.
14056 * More robust randomness gathering functions for Windows.
14058 *Jeffrey Altman <jaltman@columbia.edu>*
14060 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14061 not set then we don't setup the error code for issuer check errors
14062 to avoid possibly overwriting other errors which the callback does
14063 handle. If an application does set the flag then we assume it knows
14064 what it is doing and can handle the new informational codes
14069 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14070 a general "ANY" type, as such it should be able to decode anything
14071 including tagged types. However it didn't check the class so it would
14072 wrongly interpret tagged types in the same way as their universal
14073 counterpart and unknown types were just rejected. Changed so that the
14074 tagged and unknown types are handled in the same way as a SEQUENCE:
14075 that is the encoding is stored intact. There is also a new type
14076 "V_ASN1_OTHER" which is used when the class is not universal, in this
14077 case we have no idea what the actual type is so we just lump them all
14082 * On VMS, stdout may very well lead to a file that is written to
14083 in a record-oriented fashion. That means that every write() will
14084 write a separate record, which will be read separately by the
14085 programs trying to read from it. This can be very confusing.
14087 The solution is to put a BIO filter in the way that will buffer
14088 text until a linefeed is reached, and then write everything a
14089 line at a time, so every record written will be an actual line,
14090 not chunks of lines and not (usually doesn't happen, but I've
14091 seen it once) several lines in one record. BIO_f_linebuffer() is
14094 Currently, it's a VMS-only method, because that's where it has
14095 been tested well enough.
14099 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14100 it can return incorrect results.
14101 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14102 but it was in 0.9.6-beta[12].)
14106 * Disable the check for content being present when verifying detached
14107 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14108 include zero length content when signing messages.
14112 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14113 BIO_ctrl (for BIO pairs).
14117 * Add DSO method for VMS.
14121 * Bug fix: Montgomery multiplication could produce results with the
14126 * Add RPM specification openssl.spec and modify it to build three
14127 packages. The default package contains applications, application
14128 documentation and run-time libraries. The devel package contains
14129 include files, static libraries and function documentation. The
14130 doc package contains the contents of the doc directory. The original
14131 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14135 * Add a large number of documentation files for many SSL routines.
14137 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14139 * Add a configuration entry for Sony News 4.
14141 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14143 * Don't set the two most significant bits to one when generating a
14144 random number < q in the DSA library.
14148 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14149 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14150 the underlying transport is blocking) if a handshake took place.
14151 (The default behaviour is needed by applications such as s_client
14152 and s_server that use select() to determine when to use SSL_read;
14153 but for applications that know in advance when to expect data, it
14154 just makes things more complicated.)
14158 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14163 * Add a few more EBCDIC conditionals that make `req' and `x509'
14164 work better on such systems.
14166 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14168 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14169 Update PKCS12_parse() so it copies the friendlyName and the
14170 keyid to the certificates aux info.
14174 * Fix bug in PKCS7_verify() which caused an infinite loop
14175 if there was more than one signature.
14177 *Sven Uszpelkat <su@celocom.de>*
14179 * Major change in util/mkdef.pl to include extra information
14180 about each symbol, as well as presenting variables as well
14181 as functions. This change means that there's n more need
14182 to rebuild the .num files when some algorithms are excluded.
14186 * Allow the verify time to be set by an application,
14187 rather than always using the current time.
14191 * Phase 2 verify code reorganisation. The certificate
14192 verify code now looks up an issuer certificate by a
14193 number of criteria: subject name, authority key id
14194 and key usage. It also verifies self signed certificates
14195 by the same criteria. The main comparison function is
14196 X509_check_issued() which performs these checks.
14198 Lot of changes were necessary in order to support this
14199 without completely rewriting the lookup code.
14201 Authority and subject key identifier are now cached.
14203 The LHASH 'certs' is X509_STORE has now been replaced
14204 by a STACK_OF(X509_OBJECT). This is mainly because an
14205 LHASH can't store or retrieve multiple objects with
14206 the same hash value.
14208 As a result various functions (which were all internal
14209 use only) have changed to handle the new X509_STORE
14210 structure. This will break anything that messed round
14211 with X509_STORE internally.
14213 The functions X509_STORE_add_cert() now checks for an
14214 exact match, rather than just subject name.
14216 The X509_STORE API doesn't directly support the retrieval
14217 of multiple certificates matching a given criteria, however
14218 this can be worked round by performing a lookup first
14219 (which will fill the cache with candidate certificates)
14220 and then examining the cache for matches. This is probably
14221 the best we can do without throwing out X509_LOOKUP
14222 entirely (maybe later...).
14224 The X509_VERIFY_CTX structure has been enhanced considerably.
14226 All certificate lookup operations now go via a get_issuer()
14227 callback. Although this currently uses an X509_STORE it
14228 can be replaced by custom lookups. This is a simple way
14229 to bypass the X509_STORE hackery necessary to make this
14230 work and makes it possible to use more efficient techniques
14231 in future. A very simple version which uses a simple
14232 STACK for its trusted certificate store is also provided
14233 using X509_STORE_CTX_trusted_stack().
14235 The verify_cb() and verify() callbacks now have equivalents
14236 in the X509_STORE_CTX structure.
14238 X509_STORE_CTX also has a 'flags' field which can be used
14239 to customise the verify behaviour.
14243 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14244 excludes S/MIME capabilities.
14248 * When a certificate request is read in keep a copy of the
14249 original encoding of the signed data and use it when outputting
14250 again. Signatures then use the original encoding rather than
14251 a decoded, encoded version which may cause problems if the
14252 request is improperly encoded.
14256 * For consistency with other BIO_puts implementations, call
14257 buffer_write(b, ...) directly in buffer_puts instead of calling
14260 In BIO_puts, increment b->num_write as in BIO_write.
14262 *Peter.Sylvester@EdelWeb.fr*
14264 * Fix BN_mul_word for the case where the word is 0. (We have to use
14265 BN_zero, we may not return a BIGNUM with an array consisting of
14266 words set to zero.)
14270 * Avoid calling abort() from within the library when problems are
14271 detected, except if preprocessor symbols have been defined
14272 (such as REF_CHECK, BN_DEBUG etc.).
14276 * New openssl application 'rsautl'. This utility can be
14277 used for low level RSA operations. DER public key
14278 BIO/fp routines also added.
14282 * New Configure entry and patches for compiling on QNX 4.
14284 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14286 * A demo state-machine implementation was sponsored by
14287 Nuron (http://www.nuron.com/) and is now available in
14288 demos/state_machine.
14292 * New options added to the 'dgst' utility for signature
14293 generation and verification.
14297 * Unrecognized PKCS#7 content types are now handled via a
14298 catch all ASN1_TYPE structure. This allows unsupported
14299 types to be stored as a "blob" and an application can
14300 encode and decode it manually.
14304 * Fix various signed/unsigned issues to make a_strex.c
14305 compile under VC++.
14307 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14309 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14310 length if passed a buffer. ASN1_INTEGER_to_BN failed
14311 if passed a NULL BN and its argument was negative.
14313 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14315 * Modification to PKCS#7 encoding routines to output definite
14316 length encoding. Since currently the whole structures are in
14317 memory there's not real point in using indefinite length
14318 constructed encoding. However if OpenSSL is compiled with
14319 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14323 * Added BIO_vprintf() and BIO_vsnprintf().
14327 * Added more prefixes to parse for in the strings written
14328 through a logging bio, to cover all the levels that are available
14329 through syslog. The prefixes are now:
14331 PANIC, EMERG, EMR => LOG_EMERG
14332 ALERT, ALR => LOG_ALERT
14333 CRIT, CRI => LOG_CRIT
14334 ERROR, ERR => LOG_ERR
14335 WARNING, WARN, WAR => LOG_WARNING
14336 NOTICE, NOTE, NOT => LOG_NOTICE
14337 INFO, INF => LOG_INFO
14338 DEBUG, DBG => LOG_DEBUG
14340 and as before, if none of those prefixes are present at the
14341 beginning of the string, LOG_ERR is chosen.
14343 On Win32, the LOG_* levels are mapped according to this:
14345 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14346 LOG_WARNING => EVENTLOG_WARNING_TYPE
14347 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14352 * Made it possible to reconfigure with just the configuration
14353 argument "reconf" or "reconfigure". The command line arguments
14354 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14355 and are retrieved from there when reconfiguring.
14361 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14363 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14367 * The obj_dat.pl script was messing up the sorting of object
14368 names. The reason was that it compared the quoted version
14369 of strings as a result "OCSP" > "OCSP Signing" because
14370 " > SPACE. Changed script to store unquoted versions of
14371 names and add quotes on output. It was also omitting some
14372 names from the lookup table if they were given a default
14373 value (that is if SN is missing it is given the same
14374 value as LN and vice versa), these are now added on the
14375 grounds that if an object has a name we should be able to
14376 look it up. Finally added warning output when duplicate
14377 short or long names are found.
14381 * Changes needed for Tandem NSK.
14383 *Scott Uroff <scott@xypro.com>*
14385 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14386 RSA_padding_check_SSLv23(), special padding was never detected
14387 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14388 version rollback attacks was not effective.
14390 In s23_clnt.c, don't use special rollback-attack detection padding
14391 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14392 client; similarly, in s23_srvr.c, don't do the rollback check if
14393 SSL 2.0 is the only protocol enabled in the server.
14397 * Make it possible to get hexdumps of unprintable data with 'openssl
14398 asn1parse'. By implication, the functions ASN1_parse_dump() and
14399 BIO_dump_indent() are added.
14403 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14404 these print out strings and name structures based on various
14405 flags including RFC2253 support and proper handling of
14406 multibyte characters. Added options to the 'x509' utility
14407 to allow the various flags to be set.
14411 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14412 Also change the functions X509_cmp_current_time() and
14413 X509_gmtime_adj() work with an ASN1_TIME structure,
14414 this will enable certificates using GeneralizedTime in validity
14415 dates to be checked.
14419 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14420 negative public key encodings) on by default,
14421 NO_NEG_PUBKEY_BUG can be set to disable it.
14425 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14426 content octets. An i2c_ASN1_OBJECT is unnecessary because
14427 the encoding can be trivially obtained from the structure.
14431 * crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
14432 not read locks (CRYPTO_r_[un]lock).
14436 * A first attempt at creating official support for shared
14437 libraries through configuration. I've kept it so the
14438 default is static libraries only, and the OpenSSL programs
14439 are always statically linked for now, but there are
14440 preparations for dynamic linking in place.
14441 This has been tested on Linux and Tru64.
14445 * Randomness polling function for Win9x, as described in:
14446 Peter Gutmann, Software Generation of Practically Strong
14451 * Fix so PRNG is seeded in req if using an already existing
14456 * New options to smime application. -inform and -outform
14457 allow alternative formats for the S/MIME message including
14458 PEM and DER. The -content option allows the content to be
14459 specified separately. This should allow things like Netscape
14460 form signing output easier to verify.
14464 * Fix the ASN1 encoding of tags using the 'long form'.
14468 * New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
14469 STRING types. These convert content octets to and from the
14470 underlying type. The actual tag and length octets are
14471 already assumed to have been read in and checked. These
14472 are needed because all other string types have virtually
14473 identical handling apart from the tag. By having versions
14474 of the ASN1 functions that just operate on content octets
14475 IMPLICIT tagging can be handled properly. It also allows
14476 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14477 and ASN1_INTEGER are identical apart from the tag.
14481 * Change the handling of OID objects as follows:
14483 - New object identifiers are inserted in objects.txt, following
14484 the syntax given in objects.README.
14485 - objects.pl is used to process obj_mac.num and create a new
14487 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14490 This is currently kind of a hack, and the perl code in objects.pl
14491 isn't very elegant, but it works as I intended. The simplest way
14492 to check that it worked correctly is to look in obj_dat.h and
14493 check the array nid_objs and make sure the objects haven't moved
14494 around (this is important!). Additions are OK, as well as
14495 consistent name changes.
14499 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14503 * Addition of the command line parameter '-rand file' to 'openssl req'.
14504 The given file adds to whatever has already been seeded into the
14505 random pool through the RANDFILE configuration file option or
14506 environment variable, or the default random state file.
14510 * mkstack.pl now sorts each macro group into lexical order.
14511 Previously the output order depended on the order the files
14512 appeared in the directory, resulting in needless rewriting
14517 * Patches to make OpenSSL compile under Win32 again. Mostly
14518 work arounds for the VC++ problem that it treats func() as
14519 func(void). Also stripped out the parts of mkdef.pl that
14520 added extra typesafe functions: these no longer exist.
14524 * Reorganisation of the stack code. The macros are now all
14525 collected in safestack.h . Each macro is defined in terms of
14526 a "stack macro" of the form SKM_<name>(type, a, b). The
14527 DEBUG_SAFESTACK is now handled in terms of function casts,
14528 this has the advantage of retaining type safety without the
14529 use of additional functions. If DEBUG_SAFESTACK is not defined
14530 then the non typesafe macros are used instead. Also modified the
14531 mkstack.pl script to handle the new form. Needs testing to see
14532 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14533 the default if no major problems. Similar behaviour for ASN1_SET_OF
14534 and PKCS12_STACK_OF.
14538 * When some versions of IIS use the 'NET' form of private key the
14539 key derivation algorithm is different. Normally MD5(password) is
14540 used as a 128 bit RC4 key. In the modified case
14541 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14542 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14543 as the old Netscape_RSA functions except they have an additional
14544 'sgckey' parameter which uses the modified algorithm. Also added
14545 an -sgckey command line option to the rsa utility. Thanks to
14546 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14547 algorithm to openssl-dev.
14551 * The evp_local.h macros were using 'c.##kname' which resulted in
14552 invalid expansion on some systems (SCO 5.0.5 for example).
14553 Corrected to 'c.kname'.
14555 *Phillip Porch <root@theporch.com>*
14557 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14558 a STACK of email addresses from a certificate or request, these look
14559 in the subject name and the subject alternative name extensions and
14560 omit any duplicate addresses.
14564 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14565 This makes DSA verification about 2 % faster.
14569 * Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
14570 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14571 plus overhead for 1024 bit moduli).
14572 This makes exponentiations about 0.5 % faster for 1024 bit
14573 exponents (as measured by "openssl speed rsa2048").
14577 * Rename memory handling macros to avoid conflicts with other
14579 Malloc => OPENSSL_malloc
14580 Malloc_locked => OPENSSL_malloc_locked
14581 Realloc => OPENSSL_realloc
14582 Free => OPENSSL_free
14586 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14587 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14591 * CygWin32 support.
14593 *John Jarvie <jjarvie@newsguy.com>*
14595 * The type-safe stack code has been rejigged. It is now only compiled
14596 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14597 by default all type-specific stack functions are "#define"d back to
14598 standard stack functions. This results in more streamlined output
14599 but retains the type-safety checking possibilities of the original
14604 * The STACK code has been cleaned up, and certain type declarations
14605 that didn't make a lot of sense have been brought in line. This has
14606 also involved a cleanup of sorts in safestack.h to more correctly
14607 map type-safe stack functions onto their plain stack counterparts.
14608 This work has also resulted in a variety of "const"ifications of
14609 lots of the code, especially "_cmp" operations which should normally
14610 be prototyped with "const" parameters anyway.
14614 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14615 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14616 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14617 where all of 'md' is used each time the PRNG is used, but 'state'
14618 is used only indexed by a cyclic counter. As entropy may not be
14619 well distributed from the beginning, 'md' is important as a
14620 chaining variable. However, the output function chains only half
14621 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14622 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14623 in all of 'state' being rewritten, with the new values depending
14624 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14628 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14629 the handshake is continued after ssl_verify_cert_chain();
14630 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14631 can lead to 'unexplainable' connection aborts later.
14633 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14635 * Major EVP API cipher revision.
14636 Add hooks for extra EVP features. This allows various cipher
14637 parameters to be set in the EVP interface. Support added for variable
14638 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14639 setting of RC2 and RC5 parameters.
14641 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14644 Remove lots of duplicated code from the EVP library. For example *every*
14645 cipher init() function handles the 'iv' in the same way according to the
14646 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14647 for CFB and OFB modes they zero ctx->num.
14649 New functionality allows removal of S/MIME code RC2 hack.
14651 Most of the routines have the same form and so can be declared in terms
14654 By shifting this to the top level EVP_CipherInit() it can be removed from
14655 all individual ciphers. If the cipher wants to handle IVs or keys
14656 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14659 Change lots of functions like EVP_EncryptUpdate() to now return a
14660 value: although software versions of the algorithms cannot fail
14661 any installed hardware versions can.
14665 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14666 this option is set, tolerate broken clients that send the negotiated
14667 protocol version number instead of the requested protocol version
14672 * Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
14673 i.e. non-zero for export ciphersuites, zero otherwise.
14674 Previous versions had this flag inverted, inconsistent with
14675 rsa_tmp_cb (..._TMP_RSA_CB).
14677 *Bodo Moeller; problem reported by Amit Chopra*
14679 * Add missing DSA library text string. Work around for some IIS
14680 key files with invalid SEQUENCE encoding.
14684 * Add a document (doc/standards.txt) that list all kinds of standards
14685 and so on that are implemented in OpenSSL.
14689 * Enhance c_rehash script. Old version would mishandle certificates
14690 with the same subject name hash and wouldn't handle CRLs at all.
14691 Added -fingerprint option to crl utility, to support new c_rehash
14696 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14700 * Fix for SSL server purpose checking. Server checking was
14701 rejecting certificates which had extended key usage present
14702 but no ssl client purpose.
14704 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14706 * Make PKCS#12 code work with no password. The PKCS#12 spec
14707 is a little unclear about how a blank password is handled.
14708 Since the password in encoded as a BMPString with terminating
14709 double NULL a zero length password would end up as just the
14710 double NULL. However no password at all is different and is
14711 handled differently in the PKCS#12 key generation code. NS
14712 treats a blank password as zero length. MSIE treats it as no
14713 password on export: but it will try both on import. We now do
14714 the same: PKCS12_parse() tries zero length and no password if
14715 the password is set to "" or NULL (NULL is now a valid password:
14716 it wasn't before) as does the pkcs12 application.
14720 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14721 perror when PEM_read_bio_X509_REQ fails, the error message must
14722 be obtained from the error queue.
14726 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14727 it in ERR_remove_state if appropriate, and change ERR_get_state
14728 accordingly to avoid race conditions (this is necessary because
14729 thread_hash is no longer constant once set).
14733 * Bugfix for linux-elf makefile.one.
14737 * RSA_get_default_method() will now cause a default
14738 RSA_METHOD to be chosen if one doesn't exist already.
14739 Previously this was only set during a call to RSA_new()
14740 or RSA_new_method(NULL) meaning it was possible for
14741 RSA_get_default_method() to return NULL.
14745 * Added native name translation to the existing DSO code
14746 that will convert (if the flag to do so is set) filenames
14747 that are sufficiently small and have no path information
14748 into a canonical native form. Eg. "blah" converted to
14749 "libblah.so" or "blah.dll" etc.
14753 * New function ERR_error_string_n(e, buf, len) which is like
14754 ERR_error_string(e, buf), but writes at most 'len' bytes
14755 including the 0 terminator. For ERR_error_string_n, 'buf'
14758 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14760 * CONF library reworked to become more general. A new CONF
14761 configuration file reader "class" is implemented as well as a
14762 new functions (NCONF_*, for "New CONF") to handle it. The now
14763 old CONF_* functions are still there, but are reimplemented to
14764 work in terms of the new functions. Also, a set of functions
14765 to handle the internal storage of the configuration data is
14766 provided to make it easier to write new configuration file
14767 reader "classes" (I can definitely see something reading a
14768 configuration file in XML format, for example), called _CONF_*,
14769 or "the configuration storage API"...
14771 The new configuration file reading functions are:
14773 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14774 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14776 NCONF_default, NCONF_WIN32
14778 NCONF_dump_fp, NCONF_dump_bio
14780 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14781 NCONF_new creates a new CONF object. This works in the same way
14782 as other interfaces in OpenSSL, like the BIO interface.
14783 NCONF_dump_* dump the internal storage of the configuration file,
14784 which is useful for debugging. All other functions take the same
14785 arguments as the old CONF_* functions with the exception of the
14786 first that must be a `CONF *' instead of a `LHASH *'.
14788 To make it easier to use the new classes with the old CONF_* functions,
14789 the function CONF_set_default_method is provided.
14793 * Add '-tls1' option to 'openssl ciphers', which was already
14794 mentioned in the documentation but had not been implemented.
14795 (This option is not yet really useful because even the additional
14796 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14800 * Initial DSO code added into libcrypto for letting OpenSSL (and
14801 OpenSSL-based applications) load shared libraries and bind to
14802 them in a portable way.
14804 *Geoff Thorpe, with contributions from Richard Levitte*
14806 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000] ###
14808 * Make sure _lrotl and _lrotr are only used with MSVC.
14810 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14811 (the default implementation of RAND_status).
14813 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14814 to '-clrext' (= clear extensions), as intended and documented.
14815 *Bodo Moeller; inconsistency pointed out by Michael Attili
14816 <attili@amaxo.com>*
14818 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14819 was larger than the MD block size.
14821 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14823 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14824 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14825 using the passed key: if the passed key was a private key the result
14826 of X509_print(), for example, would be to print out all the private key
14831 * des_quad_cksum() byte order bug fix.
14832 *Ulf Möller, using the problem description in krb4-0.9.7, where
14833 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14835 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14838 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14840 * For easily testing in shell scripts whether some command
14841 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14842 returns with exit code 0 iff no command of the given name is available.
14843 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14844 the output goes to stdout and nothing is printed to stderr.
14845 Additional arguments are always ignored.
14847 Since for each cipher there is a command of the same name,
14848 the 'no-cipher' compilation switches can be tested this way.
14850 ('openssl no-XXX' is not able to detect pseudo-commands such
14851 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
14855 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
14859 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
14860 is set; it will be thrown away anyway because each handshake creates
14862 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
14863 to parameters -- in previous versions (since OpenSSL 0.9.3) the
14864 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
14865 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
14869 * New s_client option -ign_eof: EOF at stdin is ignored, and
14870 'Q' and 'R' lose their special meanings (quit/renegotiate).
14871 This is part of what -quiet does; unlike -quiet, -ign_eof
14872 does not suppress any output.
14876 * Add compatibility options to the purpose and trust code. The
14877 purpose X509_PURPOSE_ANY is "any purpose" which automatically
14878 accepts a certificate or CA, this was the previous behaviour,
14879 with all the associated security issues.
14881 X509_TRUST_COMPAT is the old trust behaviour: only and
14882 automatically trust self signed roots in certificate store. A
14883 new trust setting X509_TRUST_DEFAULT is used to specify that
14884 a purpose has no associated trust setting and it should instead
14885 use the value in the default purpose.
14889 * Fix the PKCS#8 DSA private key code so it decodes keys again
14890 and fix a memory leak.
14894 * In util/mkerr.pl (which implements 'make errors'), preserve
14895 reason strings from the previous version of the .c file, as
14896 the default to have only downcase letters (and digits) in
14897 automatically generated reasons codes is not always appropriate.
14901 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
14902 using strerror. Previously, ERR_reason_error_string() returned
14903 library names as reason strings for SYSerr; but SYSerr is a special
14904 case where small numbers are errno values, not library numbers.
14908 * Add '-dsaparam' option to 'openssl dhparam' application. This
14909 converts DSA parameters into DH parameters. (When creating parameters,
14910 DSA_generate_parameters is used.)
14914 * Include 'length' (recommended exponent length) in C code generated
14915 by 'openssl dhparam -C'.
14919 * The second argument to set_label in perlasm was already being used
14920 so couldn't be used as a "file scope" flag. Moved to third argument
14925 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
14926 instead of RAND_bytes for encryption IVs and salts.
14930 * Include RAND_status() into RAND_METHOD instead of implementing
14931 it only for md_rand.c Otherwise replacing the PRNG by calling
14932 RAND_set_rand_method would be impossible.
14936 * Don't let DSA_generate_key() enter an infinite loop if the random
14937 number generation fails.
14941 * New 'rand' application for creating pseudo-random output.
14945 * Added configuration support for Linux/IA64
14947 *Rolf Haberrecker <rolf@suse.de>*
14949 * Assembler module support for Mingw32.
14953 * Shared library support for HPUX (in shlib/).
14955 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
14957 * Shared library support for Solaris gcc.
14959 *Lutz Behnke <behnke@trustcenter.de>*
14961 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000] ###
14963 * PKCS7_encrypt() was adding text MIME headers twice because they
14964 were added manually and by SMIME_crlf_copy().
14968 * In bntest.c don't call BN_rand with zero bits argument.
14970 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
14972 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
14973 case was implemented. This caused BN_div_recp() to fail occasionally.
14977 * Add an optional second argument to the set_label() in the perl
14978 assembly language builder. If this argument exists and is set
14979 to 1 it signals that the assembler should use a symbol whose
14980 scope is the entire file, not just the current function. This
14981 is needed with MASM which uses the format label:: for this scope.
14983 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
14985 * Change the ASN1 types so they are typedefs by default. Before
14986 almost all types were #define'd to ASN1_STRING which was causing
14987 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
14992 * Change names of new functions to the new get1/get0 naming
14993 convention: After 'get1', the caller owns a reference count
14994 and has to call ..._free; 'get0' returns a pointer to some
14995 data structure without incrementing reference counters.
14996 (Some of the existing 'get' functions increment a reference
14997 counter, some don't.)
14998 Similarly, 'set1' and 'add1' functions increase reference
14999 counters or duplicate objects.
15003 * Allow for the possibility of temp RSA key generation failure:
15004 the code used to assume it always worked and crashed on failure.
15008 * Fix potential buffer overrun problem in BIO_printf().
15009 *Ulf Möller, using public domain code by Patrick Powell; problem
15010 pointed out by David Sacerdote <das33@cornell.edu>*
15012 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15013 RAND_egd() and RAND_status(). In the command line application,
15014 the EGD socket can be specified like a seed file using RANDFILE
15019 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15020 Some CAs (e.g. Verisign) distribute certificates in this form.
15024 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15025 list to exclude them. This means that no special compilation option
15026 is needed to use anonymous DH: it just needs to be included in the
15031 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15032 EVP_MD_type. The old functionality is available in a new macro called
15033 EVP_MD_md(). Change code that uses it and update docs.
15037 * ..._ctrl functions now have corresponding ..._callback_ctrl functions
15038 where the 'void *' argument is replaced by a function pointer argument.
15039 Previously 'void *' was abused to point to functions, which works on
15040 many platforms, but is not correct. As these functions are usually
15041 called by macros defined in OpenSSL header files, most source code
15042 should work without changes.
15046 * <openssl/opensslconf.h> (which is created by Configure) now contains
15047 sections with information on -D... compiler switches used for
15048 compiling the library so that applications can see them. To enable
15049 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
15050 must be defined. E.g.,
15051 #define OPENSSL_ALGORITHM_DEFINES
15052 #include <openssl/opensslconf.h>
15053 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
15055 *Richard Levitte, Ulf and Bodo Möller*
15057 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15062 * Change the 'other' type in certificate aux info to a STACK_OF
15063 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15064 the required ASN1 format: arbitrary types determined by an OID.
15068 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15069 argument to 'req'. This is not because the function is newer or
15070 better than others it just uses the work 'NEW' in the certificate
15071 request header lines. Some software needs this.
15075 * Reorganise password command line arguments: now passwords can be
15076 obtained from various sources. Delete the PEM_cb function and make
15077 it the default behaviour: i.e. if the callback is NULL and the
15078 usrdata argument is not NULL interpret it as a null terminated pass
15079 phrase. If usrdata and the callback are NULL then the pass phrase
15080 is prompted for as usual.
15084 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15085 the support is automatically enabled. The resulting binaries will
15086 autodetect the card and use it if present.
15088 *Ben Laurie and Compaq Inc.*
15090 * Work around for Netscape hang bug. This sends certificate request
15091 and server done in one record. Since this is perfectly legal in the
15092 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15093 the bugs/SSLv3 entry for more info.
15097 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15101 * Add -rand argument to smime and pkcs12 applications and read/write
15106 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15110 * Add command line password options to the remaining applications.
15114 * Bug fix for BN_div_recp() for numerators with an even number of
15119 * More tests in bntest.c, and changed test_bn output.
15123 * ./config recognizes MacOS X now.
15127 * Bug fix for BN_div() when the first words of num and divisor are
15128 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
15132 * Add support for various broken PKCS#8 formats, and command line
15133 options to produce them.
15137 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15138 get temporary BIGNUMs from a BN_CTX.
15142 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15147 * Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
15148 include a #define from the old name to the new. The original intent
15149 was that statically linked binaries could for example just call
15150 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15151 link with digests. This never worked because SSLeay_add_all_digests()
15152 and SSLeay_add_all_ciphers() were in the same source file so calling
15153 one would link with the other. They are now in separate source files.
15157 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15161 * Use a less unusual form of the Miller-Rabin primality test (it used
15162 a binary algorithm for exponentiation integrated into the Miller-Rabin
15163 loop, our standard modexp algorithms are faster).
15167 * Support for the EBCDIC character set completed.
15169 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15171 * Source code cleanups: use const where appropriate, eliminate casts,
15172 use void * instead of char * in lhash.
15176 * Bugfix: ssl3_send_server_key_exchange was not restartable
15177 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15178 this the server could overwrite ephemeral keys that the client
15183 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15184 using 50 iterations of the Rabin-Miller test.
15186 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15187 iterations of the Rabin-Miller test as required by the appendix
15188 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15189 As BN_is_prime_fasttest includes trial division, DSA parameter
15190 generation becomes much faster.
15192 This implies a change for the callback functions in DSA_is_prime
15193 and DSA_generate_parameters: The callback function is called once
15194 for each positive witness in the Rabin-Miller test, not just
15195 occasionally in the inner loop; and the parameters to the
15196 callback function now provide an iteration count for the outer
15197 loop rather than for the current invocation of the inner loop.
15198 DSA_generate_parameters additionally can call the callback
15199 function with an 'iteration count' of -1, meaning that a
15200 candidate has passed the trial division test (when q is generated
15201 from an application-provided seed, trial division is skipped).
15205 * New function BN_is_prime_fasttest that optionally does trial
15206 division before starting the Rabin-Miller test and has
15207 an additional BN_CTX * argument (whereas BN_is_prime always
15208 has to allocate at least one BN_CTX).
15209 'callback(1, -1, cb_arg)' is called when a number has passed the
15210 trial division stage.
15214 * Fix for bug in CRL encoding. The validity dates weren't being handled
15219 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15223 * New function BN_pseudo_rand().
15227 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15228 bignum version of BN_from_montgomery() with the working code from
15229 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15234 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15235 made it impossible to use the same SSL_SESSION data structure in
15236 SSL2 clients in multiple threads.
15240 * The return value of RAND_load_file() no longer counts bytes obtained
15241 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15242 to seed the PRNG (previously an explicit byte count was required).
15244 *Ulf Möller, Bodo Möller*
15246 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15247 used (char *) instead of (void *) and had casts all over the place.
15251 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15255 * Retain source code compatibility for BN_prime_checks macro:
15256 BN_is_prime(..., BN_prime_checks, ...) now uses
15257 BN_prime_checks_for_size to determine the appropriate number of
15258 Rabin-Miller iterations.
15262 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15263 DH_CHECK_P_NOT_SAFE_PRIME.
15264 (Check if this is true? OpenPGP calls them "strong".)
15268 * Merge the functionality of "dh" and "gendh" programs into a new program
15269 "dhparam". The old programs are retained for now but will handle DH keys
15270 (instead of parameters) in future.
15274 * Make the ciphers, s_server and s_client programs check the return values
15275 when a new cipher list is set.
15279 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15280 ciphers. Before when the 56bit ciphers were enabled the sorting was
15283 The syntax for the cipher sorting has been extended to support sorting by
15284 cipher-strength (using the strength_bits hard coded in the tables).
15285 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15287 Fix a bug in the cipher-command parser: when supplying a cipher command
15288 string with an "undefined" symbol (neither command nor alphanumeric
15289 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15290 an error is flagged.
15292 Due to the strength-sorting extension, the code of the
15293 ssl_create_cipher_list() function was completely rearranged. I hope that
15294 the readability was also increased :-)
15296 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15298 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15299 for the first serial number and places 2 in the serial number file. This
15300 avoids problems when the root CA is created with serial number zero and
15301 the first user certificate has the same issuer name and serial number
15306 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15307 the new code. Add documentation for this stuff.
15311 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15312 X509_*() to X509at_*() on the grounds that they don't handle X509
15313 structures and behave in an analogous way to the X509v3 functions:
15314 they shouldn't be called directly but wrapper functions should be used
15317 So we also now have some wrapper functions that call the X509at functions
15318 when passed certificate requests. (TO DO: similar things can be done with
15319 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15320 things. Some of these need some d2i or i2d and print functionality
15321 because they handle more complex structures.)
15325 * Add missing #ifndefs that caused missing symbols when building libssl
15326 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15327 NO_RSA in ssl/s2*.c.
15329 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15331 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15332 has a return value which indicates the quality of the random data
15333 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15334 error queue. New function RAND_pseudo_bytes() generates output that is
15335 guaranteed to be unique but not unpredictable. RAND_add is like
15336 RAND_seed, but takes an extra argument for an entropy estimate
15337 (RAND_seed always assumes full entropy).
15341 * Do more iterations of Rabin-Miller probable prime test (specifically,
15342 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15343 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15344 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15345 false-positive rate of at most 2^-80 for random input.
15349 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15353 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15354 in the 0.9.5 release), this returns the chain
15355 from an X509_CTX structure with a dup of the stack and all
15356 the X509 reference counts upped: so the stack will exist
15357 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15360 Also make SSL_SESSION_print() print out the verify return
15365 * Add manpage for the pkcs12 command. Also change the default
15366 behaviour so MAC iteration counts are used unless the new
15367 -nomaciter option is used. This improves file security and
15368 only older versions of MSIE (4.0 for example) need it.
15372 * Honor the no-xxx Configure options when creating .DEF files.
15376 * Add PKCS#10 attributes to field table: challengePassword,
15377 unstructuredName and unstructuredAddress. These are taken from
15378 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15379 international characters are used.
15381 More changes to X509_ATTRIBUTE code: allow the setting of types
15382 based on strings. Remove the 'loc' parameter when adding
15383 attributes because these will be a SET OF encoding which is sorted
15388 * Initial changes to the 'req' utility to allow request generation
15389 automation. This will allow an application to just generate a template
15390 file containing all the field values and have req construct the
15393 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15394 used all over the place including certificate requests and PKCS#7
15395 structures. They are currently handled manually where necessary with
15396 some primitive wrappers for PKCS#7. The new functions behave in a
15397 manner analogous to the X509 extension functions: they allow
15398 attributes to be looked up by NID and added.
15400 Later something similar to the X509V3 code would be desirable to
15401 automatically handle the encoding, decoding and printing of the
15402 more complex types. The string types like challengePassword can
15403 be handled by the string table functions.
15405 Also modified the multi byte string table handling. Now there is
15406 a 'global mask' which masks out certain types. The table itself
15407 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15408 is useful when for example there is only one permissible type
15409 (as in countryName) and using the mask might result in no valid
15414 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15415 SSL_get_peer_finished to allow applications to obtain the latest
15416 Finished messages sent to the peer or expected from the peer,
15417 respectively. (SSL_get_peer_finished is usually the Finished message
15418 actually received from the peer, otherwise the protocol will be aborted.)
15420 As the Finished message are message digests of the complete handshake
15421 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15422 be used for external authentication procedures when the authentication
15423 provided by SSL/TLS is not desired or is not enough.
15427 * Enhanced support for Alpha Linux is added. Now ./config checks if
15428 the host supports BWX extension and if Compaq C is present on the
15429 $PATH. Just exploiting of the BWX extension results in 20-30%
15430 performance kick for some algorithms, e.g. DES and RC4 to mention
15431 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15436 * Add support for MS "fast SGC". This is arguably a violation of the
15437 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15438 weak crypto and after checking the certificate is SGC a second one
15439 with strong crypto. MS SGC stops the first handshake after receiving
15440 the server certificate message and sends a second client hello. Since
15441 a server will typically do all the time consuming operations before
15442 expecting any further messages from the client (server key exchange
15443 is the most expensive) there is little difference between the two.
15445 To get OpenSSL to support MS SGC we have to permit a second client
15446 hello message after we have sent server done. In addition we have to
15447 reset the MAC if we do get this second client hello.
15451 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15452 if a DER encoded private key is RSA or DSA traditional format. Changed
15453 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15454 format DER encoded private key. Newer code should use PKCS#8 format which
15455 has the key type encoded in the ASN1 structure. Added DER private key
15456 support to pkcs8 application.
15460 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15461 ciphersuites has been selected (as required by the SSL 3/TLS 1
15462 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15463 is set, we interpret this as a request to violate the specification
15464 (the worst that can happen is a handshake failure, and 'correct'
15465 behaviour would result in a handshake failure anyway).
15469 * In SSL_CTX_add_session, take into account that there might be multiple
15470 SSL_SESSION structures with the same session ID (e.g. when two threads
15471 concurrently obtain them from an external cache).
15472 The internal cache can handle only one SSL_SESSION with a given ID,
15473 so if there's a conflict, we now throw out the old one to achieve
15478 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15479 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15480 some routines that use cipher OIDs: some ciphers do not have OIDs
15481 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15486 * Simplify the trust setting structure and code. Now we just have
15487 two sequences of OIDs for trusted and rejected settings. These will
15488 typically have values the same as the extended key usage extension
15489 and any application specific purposes.
15491 The trust checking code now has a default behaviour: it will just
15492 check for an object with the same NID as the passed id. Functions can
15493 be provided to override either the default behaviour or the behaviour
15494 for a given id. SSL client, server and email already have functions
15495 in place for compatibility: they check the NID and also return "trusted"
15496 if the certificate is self signed.
15500 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15501 traditional format into an EVP_PKEY structure.
15505 * Add a password callback function PEM_cb() which either prompts for
15506 a password if usr_data is NULL or otherwise assumes it is a null
15507 terminated password. Allow passwords to be passed on command line
15508 environment or config files in a few more utilities.
15512 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15513 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15514 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15515 Update documentation.
15519 * Support for ASN1 "NULL" type. This could be handled before by using
15520 ASN1_TYPE but there wasn't any function that would try to read a NULL
15521 and produce an error if it couldn't. For compatibility we also have
15522 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15523 don't allocate anything because they don't need to.
15527 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15530 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15532 * Rebuild of the memory allocation routines used by OpenSSL code and
15533 possibly others as well. The purpose is to make an interface that
15534 provide hooks so anyone can build a separate set of allocation and
15535 deallocation routines to be used by OpenSSL, for example memory
15536 pool implementations, or something else, which was previously hard
15537 since Malloc(), Realloc() and Free() were defined as macros having
15538 the values malloc, realloc and free, respectively (except for Win32
15539 compilations). The same is provided for memory debugging code.
15540 OpenSSL already comes with functionality to find memory leaks, but
15541 this gives people a chance to debug other memory problems.
15543 With these changes, a new set of functions and macros have appeared:
15545 CRYPTO_set_mem_debug_functions() [F]
15546 CRYPTO_get_mem_debug_functions() [F]
15547 CRYPTO_dbg_set_options() [F]
15548 CRYPTO_dbg_get_options() [F]
15549 CRYPTO_malloc_debug_init() [M]
15551 The memory debug functions are NULL by default, unless the library
15552 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15553 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15554 gives the standard debugging functions that come with OpenSSL) or
15555 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15556 provided by the library user) must be used. When the standard
15557 debugging functions are used, CRYPTO_dbg_set_options can be used to
15558 request additional information:
15559 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15560 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15562 Also, things like CRYPTO_set_mem_functions will always give the
15563 expected result (the new set of functions is used for allocation
15564 and deallocation) at all times, regardless of platform and compiler
15567 To finish it up, some functions that were never use in any other
15568 way than through macros have a new API and new semantic:
15570 CRYPTO_dbg_malloc()
15571 CRYPTO_dbg_realloc()
15574 All macros of value have retained their old syntax.
15576 *Richard Levitte and Bodo Moeller*
15578 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15579 ordering of SMIMECapabilities wasn't in "strength order" and there
15580 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15585 * Some ASN1 types with illegal zero length encoding (INTEGER,
15586 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15588 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15590 * Merge in my S/MIME library for OpenSSL. This provides a simple
15591 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15592 functionality to handle multipart/signed properly) and a utility
15593 called 'smime' to call all this stuff. This is based on code I
15594 originally wrote for Celo who have kindly allowed it to be
15595 included in OpenSSL.
15599 * Add variants des_set_key_checked and des_set_key_unchecked of
15600 des_set_key (aka des_key_sched). Global variable des_check_key
15601 decides which of these is called by des_set_key; this way
15602 des_check_key behaves as it always did, but applications and
15603 the library itself, which was buggy for des_check_key == 1,
15604 have a cleaner way to pick the version they need.
15608 * New function PKCS12_newpass() which changes the password of a
15613 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15614 dynamic mix. In both cases the ids can be used as an index into the
15615 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15616 functions so they accept a list of the field values and the
15617 application doesn't need to directly manipulate the X509_TRUST
15622 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15627 * Modify the way the V3 extension code looks up extensions. This now
15628 works in a similar way to the object code: we have some "standard"
15629 extensions in a static table which is searched with OBJ_bsearch()
15630 and the application can add dynamic ones if needed. The file
15631 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15632 updated whenever a new extension is added to the core code and kept
15633 in ext_nid order. There is a simple program 'tabtest.c' which checks
15634 this. New extensions are not added too often so this file can readily
15635 be maintained manually.
15637 There are two big advantages in doing things this way. The extensions
15638 can be looked up immediately and no longer need to be "added" using
15639 X509V3_add_standard_extensions(): this function now does nothing.
15640 *Side note: I get *lots* of email saying the extension code doesn't
15641 work because people forget to call this function*
15642 Also no dynamic allocation is done unless new extensions are added:
15643 so if we don't add custom extensions there is no need to call
15644 X509V3_EXT_cleanup().
15648 * Modify enc utility's salting as follows: make salting the default. Add a
15649 magic header, so unsalted files fail gracefully instead of just decrypting
15650 to garbage. This is because not salting is a big security hole, so people
15651 should be discouraged from doing it.
15655 * Fixes and enhancements to the 'x509' utility. It allowed a message
15656 digest to be passed on the command line but it only used this
15657 parameter when signing a certificate. Modified so all relevant
15658 operations are affected by the digest parameter including the
15659 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15660 DSA key was used because it didn't fix the digest.
15664 * Initial certificate chain verify code. Currently tests the untrusted
15665 certificates for consistency with the verify purpose (which is set
15666 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15668 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15669 this is because it will reject chains with invalid extensions whereas
15670 every previous version of OpenSSL and SSLeay made no checks at all.
15672 Trust code: checks the root CA for the relevant trust settings. Trust
15673 settings have an initial value consistent with the verify purpose: e.g.
15674 if the verify purpose is for SSL client use it expects the CA to be
15675 trusted for SSL client use. However the default value can be changed to
15676 permit custom trust settings: one example of this would be to only trust
15677 certificates from a specific "secure" set of CAs.
15679 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15680 which should be used for version portability: especially since the
15681 verify structure is likely to change more often now.
15683 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15684 to set them. If not set then assume SSL clients will verify SSL servers
15687 Two new options to the verify program: -untrusted allows a set of
15688 untrusted certificates to be passed in and -purpose which sets the
15689 intended purpose of the certificate. If a purpose is set then the
15690 new chain verify code is used to check extension consistency.
15694 * Support for the authority information access extension.
15698 * Modify RSA and DSA PEM read routines to transparently handle
15699 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15700 public keys in a format compatible with certificate
15701 SubjectPublicKeyInfo structures. Unfortunately there were already
15702 functions called *_PublicKey_* which used various odd formats so
15703 these are retained for compatibility: however the DSA variants were
15704 never in a public release so they have been deleted. Changed dsa/rsa
15705 utilities to handle the new format: note no releases ever handled public
15706 keys so we should be OK.
15708 The primary motivation for this change is to avoid the same fiasco
15709 that dogs private keys: there are several incompatible private key
15710 formats some of which are standard and some OpenSSL specific and
15711 require various evil hacks to allow partial transparent handling and
15712 even then it doesn't work with DER formats. Given the option anything
15713 other than PKCS#8 should be dumped: but the other formats have to
15714 stay in the name of compatibility.
15716 With public keys and the benefit of hindsight one standard format
15717 is used which works with EVP_PKEY, RSA or DSA structures: though
15718 it clearly returns an error if you try to read the wrong kind of key.
15720 Added a -pubkey option to the 'x509' utility to output the public key.
15721 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
15722 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
15723 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
15724 that do the same as the EVP_PKEY_assign_*() except they up the
15725 reference count of the added key (they don't "swallow" the
15730 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15731 CRLs would fail if the file contained no certificates or no CRLs:
15732 added a new function to read in both types and return the number
15733 read: this means that if none are read it will be an error. The
15734 DER versions of the certificate and CRL reader would always fail
15735 because it isn't possible to mix certificates and CRLs in DER format
15736 without choking one or the other routine. Changed this to just read
15737 a certificate: this is the best we can do. Also modified the code
15738 in apps/verify.c to take notice of return codes: it was previously
15739 attempting to read in certificates from NULL pointers and ignoring
15740 any errors: this is one reason why the cert and CRL reader seemed
15741 to work. It doesn't check return codes from the default certificate
15742 routines: these may well fail if the certificates aren't installed.
15746 * Code to support otherName option in GeneralName.
15750 * First update to verify code. Change the verify utility
15751 so it warns if it is passed a self signed certificate:
15752 for consistency with the normal behaviour. X509_verify
15753 has been modified to it will now verify a self signed
15754 certificate if *exactly* the same certificate appears
15755 in the store: it was previously impossible to trust a
15756 single self signed certificate. This means that:
15757 openssl verify ss.pem
15758 now gives a warning about a self signed certificate but
15759 openssl verify -CAfile ss.pem ss.pem
15764 * For servers, store verify_result in SSL_SESSION data structure
15765 (and add it to external session representation).
15766 This is needed when client certificate verifications fails,
15767 but an application-provided verification callback (set by
15768 SSL_CTX_set_cert_verify_callback) allows accepting the session
15769 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15770 but returns 1): When the session is reused, we have to set
15771 ssl->verify_result to the appropriate error code to avoid
15774 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15776 * Fix a bug in the new PKCS#7 code: it didn't consider the
15777 case in PKCS7_dataInit() where the signed PKCS7 structure
15778 didn't contain any existing data because it was being created.
15780 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15782 * Add a salt to the key derivation routines in enc.c. This
15783 forms the first 8 bytes of the encrypted file. Also add a
15784 -S option to allow a salt to be input on the command line.
15788 * New function X509_cmp(). Oddly enough there wasn't a function
15789 to compare two certificates. We do this by working out the SHA1
15790 hash and comparing that. X509_cmp() will be needed by the trust
15795 * SSL_get1_session() is like SSL_get_session(), but increments
15796 the reference count in the SSL_SESSION returned.
15798 *Geoff Thorpe <geoff@eu.c2.net>*
15800 * Fix for 'req': it was adding a null to request attributes.
15801 Also change the X509_LOOKUP and X509_INFO code to handle
15802 certificate auxiliary information.
15806 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15811 * Add the possibility to add extra information to the memory leak
15812 detecting output, to form tracebacks, showing from where each
15813 allocation was originated: CRYPTO_push_info("constant string") adds
15814 the string plus current file name and line number to a per-thread
15815 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15816 is like calling CYRPTO_pop_info() until the stack is empty.
15817 Also updated memory leak detection code to be multi-thread-safe.
15821 * Add options -text and -noout to pkcs7 utility and delete the
15822 encryption options which never did anything. Update docs.
15826 * Add options to some of the utilities to allow the pass phrase
15827 to be included on either the command line (not recommended on
15828 OSes like Unix) or read from the environment. Update the
15829 manpages and fix a few bugs.
15833 * Add a few manpages for some of the openssl commands.
15837 * Fix the -revoke option in ca. It was freeing up memory twice,
15838 leaking and not finding already revoked certificates.
15842 * Extensive changes to support certificate auxiliary information.
15843 This involves the use of X509_CERT_AUX structure and X509_AUX
15844 functions. An X509_AUX function such as PEM_read_X509_AUX()
15845 can still read in a certificate file in the usual way but it
15846 will also read in any additional "auxiliary information". By
15847 doing things this way a fair degree of compatibility can be
15848 retained: existing certificates can have this information added
15849 using the new 'x509' options.
15851 Current auxiliary information includes an "alias" and some trust
15852 settings. The trust settings will ultimately be used in enhanced
15853 certificate chain verification routines: currently a certificate
15854 can only be trusted if it is self signed and then it is trusted
15859 * Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
15860 The problem was that one of the replacement routines had not been working
15861 since SSLeay releases. For now the offending routine has been replaced
15862 with non-optimised assembler. Even so, this now gives around 95%
15863 performance improvement for 1024 bit RSA signs.
15867 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
15868 handling. Most clients have the effective key size in bits equal to
15869 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
15870 A few however don't do this and instead use the size of the decrypted key
15871 to determine the RC2 key length and the AlgorithmIdentifier to determine
15872 the effective key length. In this case the effective key length can still
15873 be 40 bits but the key length can be 168 bits for example. This is fixed
15874 by manually forcing an RC2 key into the EVP_PKEY structure because the
15875 EVP code can't currently handle unusual RC2 key sizes: it always assumes
15876 the key length and effective key length are equal.
15880 * Add a bunch of functions that should simplify the creation of
15881 X509_NAME structures. Now you should be able to do:
15882 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
15883 and have it automatically work out the correct field type and fill in
15884 the structures. The more adventurous can try:
15885 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
15886 and it will (hopefully) work out the correct multibyte encoding.
15890 * Change the 'req' utility to use the new field handling and multibyte
15891 copy routines. Before the DN field creation was handled in an ad hoc
15892 way in req, ca, and x509 which was rather broken and didn't support
15893 BMPStrings or UTF8Strings. Since some software doesn't implement
15894 BMPStrings or UTF8Strings yet, they can be enabled using the config file
15895 using the dirstring_type option. See the new comment in the default
15896 openssl.cnf for more info.
15900 * Make crypto/rand/md_rand.c more robust:
15901 - Assure unique random numbers after fork().
15902 - Make sure that concurrent threads access the global counter and
15903 md serializably so that we never lose entropy in them
15904 or use exactly the same state in multiple threads.
15905 Access to the large state is not always serializable because
15906 the additional locking could be a performance killer, and
15907 md should be large enough anyway.
15911 * New file apps/app_rand.c with commonly needed functionality
15912 for handling the random seed file.
15914 Use the random seed file in some applications that previously did not:
15916 dsaparam -genkey (which also ignored its '-rand' option),
15919 x509 (when signing).
15920 Except on systems with /dev/urandom, it is crucial to have a random
15921 seed file at least for key creation, DSA signing, and for DH exchanges;
15922 for RSA signatures we could do without one.
15924 gendh and gendsa (unlike genrsa) used to read only the first byte
15925 of each file listed in the '-rand' option. The function as previously
15926 found in genrsa is now in app_rand.c and is used by all programs
15927 that support '-rand'.
15931 * In RAND_write_file, use mode 0600 for creating files;
15932 don't just chmod when it may be too late.
15936 * Report an error from X509_STORE_load_locations
15937 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
15941 * New function ASN1_mbstring_copy() this copies a string in either
15942 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
15943 into an ASN1_STRING type. A mask of permissible types is passed
15944 and it chooses the "minimal" type to use or an error if not type
15949 * Add function equivalents to the various macros in asn1.h. The old
15950 macros are retained with an M_ prefix. Code inside the library can
15951 use the M_ macros. External code (including the openssl utility)
15952 should *NOT* in order to be "shared library friendly".
15956 * Add various functions that can check a certificate's extensions
15957 to see if it usable for various purposes such as SSL client,
15958 server or S/MIME and CAs of these types. This is currently
15959 VERY EXPERIMENTAL but will ultimately be used for certificate chain
15960 verification. Also added a -purpose flag to x509 utility to
15961 print out all the purposes.
15965 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
15970 * New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
15971 for, obtain and decode and extension and obtain its critical flag.
15972 This allows all the necessary extension code to be handled in a
15973 single function call.
15977 * RC4 tune-up featuring 30-40% performance improvement on most RISC
15978 platforms. See crypto/rc4/rc4_enc.c for further details.
15982 * New -noout option to asn1parse. This causes no output to be produced
15983 its main use is when combined with -strparse and -out to extract data
15984 from a file (which may not be in ASN.1 format).
15988 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
15989 when producing the local key id.
15991 *Richard Levitte <levitte@stacken.kth.se>*
15993 * New option -dhparam in s_server. This allows a DH parameter file to be
15994 stated explicitly. If it is not stated then it tries the first server
15995 certificate file. The previous behaviour hard coded the filename
16000 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16001 a public key to be input or output. For example:
16002 openssl rsa -in key.pem -pubout -out pubkey.pem
16003 Also added necessary DSA public key functions to handle this.
16007 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16008 in the message. This was handled by allowing
16009 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16011 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16013 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16014 to the end of the strings whereas this didn't. This would cause problems
16015 if strings read with d2i_ASN1_bytes() were later modified.
16017 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16019 * Fix for base64 decode bug. When a base64 bio reads only one line of
16020 data and it contains EOF it will end up returning an error. This is
16021 caused by input 46 bytes long. The cause is due to the way base64
16022 BIOs find the start of base64 encoded data. They do this by trying a
16023 trial decode on each line until they find one that works. When they
16024 do a flag is set and it starts again knowing it can pass all the
16025 data directly through the decoder. Unfortunately it doesn't reset
16026 the context it uses. This means that if EOF is reached an attempt
16027 is made to pass two EOFs through the context and this causes the
16028 resulting error. This can also cause other problems as well. As is
16029 usual with these problems it takes *ages* to find and the fix is
16030 trivial: move one line.
16032 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) *
16034 * Ugly workaround to get s_client and s_server working under Windows. The
16035 old code wouldn't work because it needed to select() on sockets and the
16036 tty (for keypresses and to see if data could be written). Win32 only
16037 supports select() on sockets so we select() with a 1s timeout on the
16038 sockets and then see if any characters are waiting to be read, if none
16039 are present then we retry, we also assume we can always write data to
16040 the tty. This isn't nice because the code then blocks until we've
16041 received a complete line of data and it is effectively polling the
16042 keyboard at 1s intervals: however it's quite a bit better than not
16043 working at all :-) A dedicated Windows application might handle this
16044 with an event loop for example.
16048 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16049 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16050 will be called when RSA_sign() and RSA_verify() are used. This is useful
16051 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16052 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16053 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16054 This necessitated the support of an extra signature type NID_md5_sha1
16055 for SSL signatures and modifications to the SSL library to use it instead
16056 of calling RSA_public_decrypt() and RSA_private_encrypt().
16060 * Add new -verify -CAfile and -CApath options to the crl program, these
16061 will lookup a CRL issuers certificate and verify the signature in a
16062 similar way to the verify program. Tidy up the crl program so it
16063 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16064 less strict. It will now permit CRL extensions even if it is not
16065 a V2 CRL: this will allow it to tolerate some broken CRLs.
16069 * Initialize all non-automatic variables each time one of the openssl
16070 sub-programs is started (this is necessary as they may be started
16071 multiple times from the "OpenSSL>" prompt).
16073 *Lennart Bang, Bodo Moeller*
16075 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16076 removing all other RSA functionality (this is what NO_RSA does). This
16077 is so (for example) those in the US can disable those operations covered
16078 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16083 * Non-copying interface to BIO pairs.
16084 (still largely untested)
16088 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16089 ASCII string. This was handled independently in various places before.
16093 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16094 UTF8 strings a character at a time.
16098 * Use client_version from client hello to select the protocol
16099 (s23_srvr.c) and for RSA client key exchange verification
16100 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16104 * Add various utility functions to handle SPKACs, these were previously
16105 handled by poking round in the structure internals. Added new function
16106 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16107 print, verify and generate SPKACs. Based on an original idea from
16108 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16112 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16116 * Allow the config file extension section to be overwritten on the
16117 command line. Based on an original idea from Massimiliano Pala
16118 <madwolf@comune.modena.it>. The new option is called -extensions
16119 and can be applied to ca, req and x509. Also -reqexts to override
16120 the request extensions in req and -crlexts to override the crl extensions
16125 * Add new feature to the SPKAC handling in ca. Now you can include
16126 the same field multiple times by preceding it by "XXXX." for example:
16129 this is the same syntax as used in the req config file.
16133 * Allow certificate extensions to be added to certificate requests. These
16134 are specified in a 'req_extensions' option of the req section of the
16135 config file. They can be printed out with the -text option to req but
16136 are otherwise ignored at present.
16140 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16141 data read consists of only the final block it would not decrypted because
16142 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16143 A misplaced 'break' also meant the decrypted final block might not be
16144 copied until the next read.
16148 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16149 a few extra parameters to the DH structure: these will be useful if
16150 for example we want the value of 'q' or implement X9.42 DH.
16154 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16155 provides hooks that allow the default DSA functions or functions on a
16156 "per key" basis to be replaced. This allows hardware acceleration and
16157 hardware key storage to be handled without major modification to the
16158 library. Also added low level modexp hooks and CRYPTO_EX structure and
16159 associated functions.
16163 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16164 as "read only": it can't be written to and the buffer it points to will
16165 not be freed. Reading from a read only BIO is much more efficient than
16166 a normal memory BIO. This was added because there are several times when
16167 an area of memory needs to be read from a BIO. The previous method was
16168 to create a memory BIO and write the data to it, this results in two
16169 copies of the data and an O(n^2) reading algorithm. There is a new
16170 function BIO_new_mem_buf() which creates a read only memory BIO from
16171 an area of memory. Also modified the PKCS#7 routines to use read only
16176 * Bugfix: ssl23_get_client_hello did not work properly when called in
16177 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16178 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16179 but a retry condition occurred while trying to read the rest.
16183 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16184 NID_pkcs7_encrypted by default: this was wrong since this should almost
16185 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16186 the encrypted data type: this is a more sensible place to put it and it
16187 allows the PKCS#12 code to be tidied up that duplicated this
16192 * Changed obj_dat.pl script so it takes its input and output files on
16193 the command line. This should avoid shell escape redirection problems
16198 * Initial support for certificate extension requests, these are included
16199 in things like Xenroll certificate requests. Included functions to allow
16200 extensions to be obtained and added.
16204 * -crlf option to s_client and s_server for sending newlines as
16205 CRLF (as required by many protocols).
16209 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999] ###
16211 * Install libRSAglue.a when OpenSSL is built with RSAref.
16213 *Ralf S. Engelschall*
16215 * A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
16217 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16219 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16224 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16225 DH parameters/keys (q is lost during that conversion, but the resulting
16226 DH parameters contain its length).
16228 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16229 much faster than DH_generate_parameters (which creates parameters
16230 where p = 2*q + 1), and also the smaller q makes DH computations
16231 much more efficient (160-bit exponentiation instead of 1024-bit
16232 exponentiation); so this provides a convenient way to support DHE
16233 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16234 utter importance to use
16235 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16237 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16238 when such DH parameters are used, because otherwise small subgroup
16239 attacks may become possible!
16243 * Avoid memory leak in i2d_DHparams.
16247 * Allow the -k option to be used more than once in the enc program:
16248 this allows the same encrypted message to be read by multiple recipients.
16252 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16253 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16254 it will always use the numerical form of the OID, even if it has a short
16259 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16260 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16261 otherwise bn_mod_exp was called. In the case of hardware keys for example
16262 no private key components need be present and it might store extra data
16263 in the RSA structure, which cannot be accessed from bn_mod_exp.
16264 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16265 private key operations.
16269 * Added support for SPARC Linux.
16273 * pem_password_cb function type incompatibly changed from
16274 typedef int pem_password_cb(char *buf, int size, int rwflag);
16276 ....(char *buf, int size, int rwflag, void *userdata);
16277 so that applications can pass data to their callbacks:
16278 The PEM[_ASN1]_{read,write}... functions and macros now take an
16279 additional void * argument, which is just handed through whenever
16280 the password callback is called.
16282 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16284 New function SSL_CTX_set_default_passwd_cb_userdata.
16286 Compatibility note: As many C implementations push function arguments
16287 onto the stack in reverse order, the new library version is likely to
16288 interoperate with programs that have been compiled with the old
16289 pem_password_cb definition (PEM_whatever takes some data that
16290 happens to be on the stack as its last argument, and the callback
16291 just ignores this garbage); but there is no guarantee whatsoever that
16294 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16295 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16296 problems not only on Windows, but also on some Unix platforms.
16297 To avoid problematic command lines, these definitions are now in an
16298 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16299 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16303 * MIPS III/IV assembler module is reimplemented.
16307 * More DES library cleanups: remove references to srand/rand and
16308 delete an unused file.
16312 * Add support for the free Netwide assembler (NASM) under Win32,
16313 since not many people have MASM (ml) and it can be hard to obtain.
16314 This is currently experimental but it seems to work OK and pass all
16315 the tests. Check out INSTALL.W32 for info.
16319 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16320 without temporary keys kept an extra copy of the server key,
16321 and connections with temporary keys did not free everything in case
16326 * New function RSA_check_key and new openssl rsa option -check
16327 for verifying the consistency of RSA keys.
16329 *Ulf Moeller, Bodo Moeller*
16331 * Various changes to make Win32 compile work:
16332 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16333 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16334 comparison" warnings.
16335 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
16339 * Add a debugging option to PKCS#5 v2 key generation function: when
16340 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16341 derived keys are printed to stderr.
16345 * Copy the flags in ASN1_STRING_dup().
16347 *Roman E. Pavlov <pre@mo.msk.ru>*
16349 * The x509 application mishandled signing requests containing DSA
16350 keys when the signing key was also DSA and the parameters didn't match.
16352 It was supposed to omit the parameters when they matched the signing key:
16353 the verifying software was then supposed to automatically use the CA's
16354 parameters if they were absent from the end user certificate.
16356 Omitting parameters is no longer recommended. The test was also
16357 the wrong way round! This was probably due to unusual behaviour in
16358 EVP_cmp_parameters() which returns 1 if the parameters match.
16359 This meant that parameters were omitted when they *didn't* match and
16360 the certificate was useless. Certificates signed with 'ca' didn't have
16363 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16365 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16366 The interface is as follows:
16367 Applications can use
16368 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16369 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16370 "off" is now the default.
16371 The library internally uses
16372 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16373 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16374 to disable memory-checking temporarily.
16376 Some inconsistent states that previously were possible (and were
16377 even the default) are now avoided.
16379 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16380 with each memory chunk allocated; this is occasionally more helpful
16381 than just having a counter.
16383 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16385 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16390 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16391 which largely parallels "options", but is for changing API behaviour,
16392 whereas "options" are about protocol behaviour.
16393 Initial "mode" flags are:
16395 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16396 a single record has been written.
16397 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16398 retries use the same buffer location.
16399 (But all of the contents must be
16404 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16407 * Fix problems with no-hmac etc.
16409 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16411 * New functions RSA_get_default_method(), RSA_set_method() and
16412 RSA_get_method(). These allows replacement of RSA_METHODs without having
16413 to mess around with the internals of an RSA structure.
16417 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16418 Also really enable memory leak checks in openssl.c and in some
16421 *Chad C. Mulligan, Bodo Moeller*
16423 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16424 up the length of negative integers. This has now been simplified to just
16425 store the length when it is first determined and use it later, rather
16426 than trying to keep track of where data is copied and updating it to
16428 *Steve Henson, reported by Brien Wheeler
16429 <bwheeler@authentica-security.com>*
16431 * Add a new function PKCS7_signatureVerify. This allows the verification
16432 of a PKCS#7 signature but with the signing certificate passed to the
16433 function itself. This contrasts with PKCS7_dataVerify which assumes the
16434 certificate is present in the PKCS#7 structure. This isn't always the
16435 case: certificates can be omitted from a PKCS#7 structure and be
16436 distributed by "out of band" means (such as a certificate database).
16440 * Complete the PEM_* macros with DECLARE_PEM versions to replace the
16441 function prototypes in pem.h, also change util/mkdef.pl to add the
16442 necessary function names.
16446 * mk1mf.pl (used by Windows builds) did not properly read the
16447 options set by Configure in the top level Makefile, and Configure
16448 was not even able to write more than one option correctly.
16449 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16453 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16454 file to be loaded from a BIO or FILE pointer. The BIO version will
16455 for example allow memory BIOs to contain config info.
16459 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16460 Whoever hopes to achieve shared-library compatibility across versions
16461 must use this, not the compile-time macro.
16462 (Exercise 0.9.4: Which is the minimum library version required by
16464 Note: All this applies only to multi-threaded programs, others don't
16469 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16470 through a BIO pair triggered the default case, i.e.
16471 SSLerr(...,SSL_R_UNKNOWN_STATE).
16475 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16476 can use the SSL library even if none of the specific BIOs is
16481 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16482 for the encoded length.
16484 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16486 * Add initial documentation of the X509V3 functions.
16490 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16491 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16492 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16493 secure PKCS#8 private key format with a high iteration count.
16497 * Fix determination of Perl interpreter: A perl or perl5
16498 _directory_ in $PATH was also accepted as the interpreter.
16500 *Ralf S. Engelschall*
16502 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16503 wrong with it but it was very old and did things like calling
16504 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16505 unusual formatting.
16509 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16510 to use the new extension code.
16514 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16515 with macros. This should make it easier to change their form, add extra
16516 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16521 * Add to configuration table a new entry that can specify an alternative
16522 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16523 according to Mark Crispin <MRC@Panda.COM>.
16528 * DES CBC did not update the IV. Weird.
16532 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16533 Changing the behaviour of the former might break existing programs --
16534 where IV updating is needed, des_ncbc_encrypt can be used.
16537 * When bntest is run from "make test" it drives bc to check its
16538 calculations, as well as internally checking them. If an internal check
16539 fails, it needs to cause bc to give a non-zero result or make test carries
16540 on without noticing the failure. Fixed.
16544 * DES library cleanups.
16548 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16549 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16550 ciphers. NOTE: although the key derivation function has been verified
16551 against some published test vectors it has not been extensively tested
16552 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16557 * Instead of "mkdir -p", which is not fully portable, use new
16558 Perl script "util/mkdir-p.pl".
16562 * Rewrite the way password based encryption (PBE) is handled. It used to
16563 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16564 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16565 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16566 the 'parameter' field of the AlgorithmIdentifier is passed to the
16567 underlying key generation function so it must do its own ASN1 parsing.
16568 This has also changed the EVP_PBE_CipherInit() function which now has a
16569 'parameter' argument instead of literal salt and iteration count values
16570 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16574 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16575 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16576 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16577 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16578 value was just used as a "magic string" and not used directly its
16579 value doesn't matter.
16583 * Introduce some semblance of const correctness to BN. Shame C doesn't
16588 * "linux-sparc64" configuration (ultrapenguin).
16590 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16591 "linux-sparc" configuration.
16593 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16595 * config now generates no-xxx options for missing ciphers.
16599 * Support the EBCDIC character set (work in progress).
16600 File ebcdic.c not yet included because it has a different license.
16602 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16604 * Support BS2000/OSD-POSIX.
16606 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16608 * Make callbacks for key generation use void * instead of char *.
16612 * Make S/MIME samples compile (not yet tested).
16616 * Additional typesafe stacks.
16620 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16625 ### Changes between 0.9.3 and 0.9.3a [29 May 1999] ###
16627 * New configuration variant "sco5-gcc".
16629 * Updated some demos.
16631 *Sean O Riordain, Wade Scholine*
16633 * Add missing BIO_free at exit of pkcs12 application.
16637 * Fix memory leak in conf.c.
16641 * Updates for Win32 to assembler version of MD5.
16645 * Set #! path to perl in apps/der_chop to where we found it
16646 instead of using a fixed path.
16650 * SHA library changes for irix64-mips4-cc.
16654 * Improvements for VMS support.
16659 ### Changes between 0.9.2b and 0.9.3 [24 May 1999] ###
16661 * Bignum library bug fix. IRIX 6 passes "make test" now!
16662 This also avoids the problems with SC4.2 and unpatched SC5.
16664 *Andy Polyakov <appro@fy.chalmers.se>*
16666 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16667 These are required because of the typesafe stack would otherwise break
16668 existing code. If old code used a structure member which used to be STACK
16669 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16670 sk_num or sk_value it would produce an error because the num, data members
16671 are not present in STACK_OF. Now it just produces a warning. sk_set
16672 replaces the old method of assigning a value to sk_value
16673 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16674 that does this will no longer work (and should use sk_set instead) but
16675 this could be regarded as a "questionable" behaviour anyway.
16679 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16680 correctly handle encrypted S/MIME data.
16684 * Change type of various DES function arguments from des_cblock
16685 (which means, in function argument declarations, pointer to char)
16686 to des_cblock * (meaning pointer to array with 8 char elements),
16687 which allows the compiler to do more typechecking; it was like
16688 that back in SSLeay, but with lots of ugly casts.
16690 Introduce new type const_des_cblock.
16694 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16695 problems: find RecipientInfo structure that matches recipient certificate
16696 and initialise the ASN1 structures properly based on passed cipher.
16700 * Belatedly make the BN tests actually check the results.
16704 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16705 to and from BNs: it was completely broken. New compilation option
16706 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16707 key elements as negative integers.
16711 * Reorganize and speed up MD5.
16713 *Andy Polyakov <appro@fy.chalmers.se>*
16717 *Richard Levitte <richard@levitte.org>*
16719 * New option -out to asn1parse to allow the parsed structure to be
16720 output to a file. This is most useful when combined with the -strparse
16721 option to examine the output of things like OCTET STRINGS.
16725 * Make SSL library a little more fool-proof by not requiring any longer
16726 that SSL_set_{accept,connect}_state be called before
16727 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
16728 in many applications because usually everything *appeared* to work as
16729 intended anyway -- now it really works as intended).
16733 * Move openssl.cnf out of lib/.
16737 * Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
16738 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16739 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
16741 *Ralf S. Engelschall*
16743 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16744 handle PKCS#7 enveloped data properly.
16746 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16748 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16749 copying pointers. The cert_st handling is changed by this in
16750 various ways (and thus what used to be known as ctx->default_cert
16751 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
16752 any longer when s->cert does not give us what we need).
16753 ssl_cert_instantiate becomes obsolete by this change.
16754 As soon as we've got the new code right (possibly it already is?),
16755 we have solved a couple of bugs of the earlier code where s->cert
16756 was used as if it could not have been shared with other SSL structures.
16758 Note that using the SSL API in certain dirty ways now will result
16759 in different behaviour than observed with earlier library versions:
16760 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
16761 does not influence s as it used to.
16763 In order to clean up things more thoroughly, inside SSL_SESSION
16764 we don't use CERT any longer, but a new structure SESS_CERT
16765 that holds per-session data (if available); currently, this is
16766 the peer's certificate chain and, for clients, the server's certificate
16767 and temporary key. CERT holds only those values that can have
16768 meaningful defaults in an SSL_CTX.
16772 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16773 from the internal representation. Various PKCS#7 fixes: remove some
16774 evil casts and set the enc_dig_alg field properly based on the signing
16779 * Allow PKCS#12 password to be set from the command line or the
16780 environment. Let 'ca' get its config file name from the environment
16781 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16786 * Allow certificate policies extension to use an IA5STRING for the
16787 organization field. This is contrary to the PKIX definition but
16788 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16793 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16794 without disallowing inline assembler and the like for non-pedantic builds.
16798 * Support Borland C++ builder.
16800 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16806 * SHA-1 cleanups and performance enhancements.
16808 *Andy Polyakov <appro@fy.chalmers.se>*
16810 * Sparc v8plus assembler for the bignum library.
16812 *Andy Polyakov <appro@fy.chalmers.se>*
16814 * Accept any -xxx and +xxx compiler options in Configure.
16818 * Update HPUX configuration.
16822 * Add missing sk_<type>_unshift() function to safestack.h
16824 *Ralf S. Engelschall*
16826 * New function SSL_CTX_use_certificate_chain_file that sets the
16827 "extra_cert"s in addition to the certificate. (This makes sense
16828 only for "PEM" format files, as chains as a whole are not
16833 * Support verify_depth from the SSL API.
16834 x509_vfy.c had what can be considered an off-by-one-error:
16835 Its depth (which was not part of the external interface)
16836 was actually counting the number of certificates in a chain;
16837 now it really counts the depth.
16841 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16842 instead of X509err, which often resulted in confusing error
16843 messages since the error codes are not globally unique
16844 (e.g. an alleged error in ssl3_accept when a certificate
16845 didn't match the private key).
16847 * New function SSL_CTX_set_session_id_context that allows to set a default
16848 value (so that you don't need SSL_set_session_id_context for each
16849 connection using the SSL_CTX).
16853 * OAEP decoding bug fix.
16857 * Support INSTALL_PREFIX for package builders, as proposed by
16862 * New Configure options "threads" and "no-threads". For systems
16863 where the proper compiler options are known (currently Solaris
16864 and Linux), "threads" is the default.
16868 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
16872 * Install various scripts to $(OPENSSLDIR)/misc, not to
16873 $(INSTALLTOP)/bin -- they shouldn't clutter directories
16874 such as /usr/local/bin.
16878 * "make linux-shared" to build shared libraries.
16880 *Niels Poppe <niels@netbox.org>*
16882 * New Configure option no-<cipher> (rsa, idea, rc5, ...).
16886 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
16887 extension adding in x509 utility.
16891 * Remove NOPROTO sections and error code comments.
16895 * Partial rewrite of the DEF file generator to now parse the ANSI
16900 * New Configure options --prefix=DIR and --openssldir=DIR.
16904 * Complete rewrite of the error code script(s). It is all now handled
16905 by one script at the top level which handles error code gathering,
16906 header rewriting and C source file generation. It should be much better
16907 than the old method: it now uses a modified version of Ulf's parser to
16908 read the ANSI prototypes in all header files (thus the old K&R definitions
16909 aren't needed for error creation any more) and do a better job of
16910 translating function codes into names. The old 'ASN1 error code embedded
16911 in a comment' is no longer necessary and it doesn't use .err files which
16912 have now been deleted. Also the error code call doesn't have to appear all
16913 on one line (which resulted in some large lines...).
16917 * Change #include filenames from <foo.h> to <openssl/foo.h>.
16921 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
16922 0 (which usually indicates a closed connection), but continue reading.
16926 * Fix some race conditions.
16930 * Add support for CRL distribution points extension. Add Certificate
16931 Policies and CRL distribution points documentation.
16935 * Move the autogenerated header file parts to crypto/opensslconf.h.
16939 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
16940 8 of keying material. Merlin has also confirmed interop with this fix
16941 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
16943 *Merlin Hughes <merlin@baltimore.ie>*
16945 * Fix lots of warnings.
16947 *Richard Levitte <levitte@stacken.kth.se>*
16949 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
16950 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
16952 *Richard Levitte <levitte@stacken.kth.se>*
16954 * Fix problems with sizeof(long) == 8.
16956 *Andy Polyakov <appro@fy.chalmers.se>*
16958 * Change functions to ANSI C.
16962 * Fix typos in error codes.
16964 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
16966 * Remove defunct assembler files from Configure.
16970 * SPARC v8 assembler BIGNUM implementation.
16972 *Andy Polyakov <appro@fy.chalmers.se>*
16974 * Support for Certificate Policies extension: both print and set.
16975 Various additions to support the r2i method this uses.
16979 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
16980 return a const string when you are expecting an allocated buffer.
16984 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
16985 types DirectoryString and DisplayText.
16989 * Add code to allow r2i extensions to access the configuration database,
16990 add an LHASH database driver and add several ctx helper functions.
16994 * Fix an evil bug in bn_expand2() which caused various BN functions to
16995 fail when they extended the size of a BIGNUM.
16999 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17000 support typesafe stack.
17004 * Fix typo in SSL_[gs]et_options().
17006 *Nils Frostberg <nils@medcom.se>*
17008 * Delete various functions and files that belonged to the (now obsolete)
17009 old X509V3 handling code.
17013 * New Configure option "rsaref".
17017 * Don't auto-generate pem.h.
17021 * Introduce type-safe ASN.1 SETs.
17025 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17027 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17029 * Introduce type-safe STACKs. This will almost certainly break lots of code
17030 that links with OpenSSL (well at least cause lots of warnings), but fear
17031 not: the conversion is trivial, and it eliminates loads of evil casts. A
17032 few STACKed things have been converted already. Feel free to convert more.
17033 In the fullness of time, I'll do away with the STACK type altogether.
17037 * Add `openssl ca -revoke <certfile>' facility which revokes a certificate
17038 specified in <certfile> by updating the entry in the index.txt file.
17039 This way one no longer has to edit the index.txt file manually for
17040 revoking a certificate. The -revoke option does the gory details now.
17042 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17044 * Fix `openssl crl -noout -text' combination where `-noout' killed the
17045 `-text' option at all and this way the `-noout -text' combination was
17046 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
17048 *Ralf S. Engelschall*
17050 * Make sure a corresponding plain text error message exists for the
17051 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17052 verify callback function determined that a certificate was revoked.
17054 *Ralf S. Engelschall*
17056 * Bugfix: In test/testenc, don't test "openssl <cipher>" for
17057 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17058 all available ciphers including rc5, which was forgotten until now.
17059 In order to let the testing shell script know which algorithms
17060 are available, a new (up to now undocumented) command
17061 "openssl list-cipher-commands" is used.
17065 * Bugfix: s_client occasionally would sleep in select() when
17066 it should have checked SSL_pending() first.
17070 * New functions DSA_do_sign and DSA_do_verify to provide access to
17071 the raw DSA values prior to ASN.1 encoding.
17075 * Tweaks to Configure
17077 *Niels Poppe <niels@netbox.org>*
17079 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17084 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17088 * New config option to avoid instructions that are illegal on the 80386.
17089 The default code is faster, but requires at least a 486.
17093 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17094 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17095 same as SSL2_VERSION anyway.
17099 * New "-showcerts" option for s_client.
17103 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17104 application. Various cleanups and fixes.
17108 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17109 modify error routines to work internally. Add error codes and PBE init
17110 to library startup routines.
17114 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17115 packing functions to asn1 and evp. Changed function names and error
17116 codes along the way.
17120 * PKCS12 integration: and so it begins... First of several patches to
17121 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17122 objects to objects.h
17126 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17127 and display support for Thawte strong extranet extension.
17131 * Add LinuxPPC support.
17133 *Jeff Dubrule <igor@pobox.org>*
17135 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17136 bn_div_words in alpha.s.
17138 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17140 * Make sure the RSA OAEP test is skipped under -DRSAref because
17141 OAEP isn't supported when OpenSSL is built with RSAref.
17143 *Ulf Moeller <ulf@fitug.de>*
17145 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17146 so they no longer are missing under -DNOPROTO.
17148 *Soren S. Jorvang <soren@t.dk>*
17151 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999] ###
17153 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17154 doesn't work when the session is reused. Coming soon!
17158 * Fix a security hole, that allows sessions to be reused in the wrong
17159 context thus bypassing client cert protection! All software that uses
17160 client certs and session caches in multiple contexts NEEDS PATCHING to
17161 allow session reuse! A fuller solution is in the works.
17163 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17165 * Some more source tree cleanups (removed obsolete files
17166 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17167 permission on "config" script to be executable) and a fix for the INSTALL
17170 *Ulf Moeller <ulf@fitug.de>*
17172 * Remove some legacy and erroneous uses of malloc, free instead of
17175 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17177 * Make rsa_oaep_test return non-zero on error.
17179 *Ulf Moeller <ulf@fitug.de>*
17181 * Add support for native Solaris shared libraries. Configure
17182 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17183 if someone would make that last step automatic.
17185 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17187 * ctx_size was not built with the right compiler during "make links". Fixed.
17191 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17192 except NULL ciphers". This means the default cipher list will no longer
17193 enable NULL ciphers. They need to be specifically enabled e.g. with
17194 the string "DEFAULT:eNULL".
17198 * Fix to RSA private encryption routines: if p < q then it would
17199 occasionally produce an invalid result. This will only happen with
17200 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17204 * Be less restrictive and allow also `perl util/perlpath.pl
17205 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17206 because this way one can also use an interpreter named `perl5' (which is
17207 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17208 installed as `perl').
17210 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17212 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17214 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17216 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17217 advapi32.lib to Win32 build and change the pem test comparison
17218 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17219 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17220 and crypto/des/ede_cbcm_enc.c.
17224 * DES quad checksum was broken on big-endian architectures. Fixed.
17228 * Comment out two functions in bio.h that aren't implemented. Fix up the
17229 Win32 test batch file so it (might) work again. The Win32 test batch file
17230 is horrible: I feel ill....
17234 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17235 in e_os.h. Audit of header files to check ANSI and non ANSI
17236 sections: 10 functions were absent from non ANSI section and not exported
17237 from Windows DLLs. Fixed up libeay.num for new functions.
17241 * Make `openssl version' output lines consistent.
17243 *Ralf S. Engelschall*
17245 * Fix Win32 symbol export lists for BIO functions: Added
17246 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17247 to ms/libeay{16,32}.def.
17249 *Ralf S. Engelschall*
17251 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17252 fine under Unix and passes some trivial tests I've now added. But the
17253 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17254 added to make sure no one expects that this stuff really works in the
17255 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17256 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17259 *Ralf S. Engelschall*
17261 * Fix the generation of two part addresses in perl.
17263 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17265 * Add config entry for Linux on MIPS.
17267 *John Tobey <jtobey@channel1.com>*
17269 * Make links whenever Configure is run, unless we are on Windoze.
17273 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17274 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17279 * Add a useful kludge to allow package maintainers to specify compiler and
17280 other platforms details on the command line without having to patch the
17281 Configure script every time: One now can use ``perl Configure
17282 <id>:<details>'', i.e. platform ids are allowed to have details appended
17283 to them (separated by colons). This is treated as there would be a static
17284 pre-configured entry in Configure's %table under key <id> with value
17285 <details> and ``perl Configure <id>'' is called. So, when you want to
17286 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17287 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
17288 now, which overrides the FreeBSD-elf entry on-the-fly.
17290 *Ralf S. Engelschall*
17292 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17296 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17297 on the `perl Configure ...' command line. This way one can compile
17298 OpenSSL libraries with Position Independent Code (PIC) which is needed
17299 for linking it into DSOs.
17301 *Ralf S. Engelschall*
17303 * Remarkably, export ciphers were totally broken and no-one had noticed!
17308 * Cleaned up the LICENSE document: The official contact for any license
17309 questions now is the OpenSSL core team under openssl-core@openssl.org.
17310 And add a paragraph about the dual-license situation to make sure people
17311 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17312 to the OpenSSL toolkit.
17314 *Ralf S. Engelschall*
17316 * General source tree makefile cleanups: Made `making xxx in yyy...'
17317 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17318 Additionally cleaned up the `make links' target: Remove unnecessary
17319 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17320 to speed processing and no longer clutter the display with confusing
17321 stuff. Instead only the actually done links are displayed.
17323 *Ralf S. Engelschall*
17325 * Permit null encryption ciphersuites, used for authentication only. It used
17326 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17327 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17332 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17333 signed attributes when verifying signatures (this would break them),
17334 the detached data encoding was wrong and public keys obtained using
17335 X509_get_pubkey() weren't freed.
17339 * Add text documentation for the BUFFER functions. Also added a work around
17340 to a Win95 console bug. This was triggered by the password read stuff: the
17341 last character typed gets carried over to the next fread(). If you were
17342 generating a new cert request using 'req' for example then the last
17343 character of the passphrase would be CR which would then enter the first
17348 * Added the new `Includes OpenSSL Cryptography Software' button as
17349 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17350 button and can be used by applications based on OpenSSL to show the
17351 relationship to the OpenSSL project.
17353 *Ralf S. Engelschall*
17355 * Remove confusing variables in function signatures in files
17356 ssl/ssl_lib.c and ssl/ssl.h.
17358 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17360 * Don't install bss_file.c under PREFIX/include/
17362 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17364 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17365 functions that return function pointers and has support for NT specific
17366 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17367 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17368 unsigned to signed types: this was killing the Win32 compile.
17372 * Add new certificate file to stack functions,
17373 SSL_add_dir_cert_subjects_to_stack() and
17374 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17375 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17376 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17377 This means that Apache-SSL and similar packages don't have to mess around
17378 to add as many CAs as they want to the preferred list.
17382 * Experiment with doxygen documentation. Currently only partially applied to
17384 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
17385 openssl.doxy as the configuration file.
17389 * Get rid of remaining C++-style comments which strict C compilers hate.
17391 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17393 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17394 compiled in by default: it has problems with large keys.
17398 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17399 DH private keys and/or callback functions which directly correspond to
17400 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17401 is needed for applications which have to configure certificates on a
17402 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17404 For the RSA certificate situation is makes no difference, but
17405 for the DSA certificate situation this fixes the "no shared cipher"
17406 problem where the OpenSSL cipher selection procedure failed because the
17407 temporary keys were not overtaken from the context and the API provided
17408 no way to reconfigure them.
17409 The new functions now let applications reconfigure the stuff and they
17410 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17411 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17412 non-public-API function ssl_cert_instantiate() is used as a helper
17413 function and also to reduce code redundancy inside ssl_rsa.c.
17415 *Ralf S. Engelschall*
17417 * Move s_server -dcert and -dkey options out of the undocumented feature
17418 area because they are useful for the DSA situation and should be
17419 recognized by the users.
17421 *Ralf S. Engelschall*
17423 * Fix the cipher decision scheme for export ciphers: the export bits are
17424 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17425 SSL_EXP_MASK. So, the original variable has to be used instead of the
17426 already masked variable.
17428 *Richard Levitte <levitte@stacken.kth.se>*
17430 * Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
17432 *Richard Levitte <levitte@stacken.kth.se>*
17434 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17435 from `int' to `unsigned int' because it's a length and initialized by
17436 EVP_DigestFinal() which expects an `unsigned int *'.
17438 *Richard Levitte <levitte@stacken.kth.se>*
17440 * Don't hard-code path to Perl interpreter on shebang line of Configure
17441 script. Instead use the usual Shell->Perl transition trick.
17443 *Ralf S. Engelschall*
17445 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17446 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17447 -noout -modulus' as it's already the case for `openssl rsa -noout
17448 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17449 currently the public key is printed (a decision which was already done by
17450 `openssl dsa -modulus' in the past) which serves a similar purpose.
17451 Additionally the NO_RSA no longer completely removes the whole -modulus
17452 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17455 *Ralf S. Engelschall*
17457 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17458 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17460 *Arne Ansper <arne@ats.cyber.ee>*
17462 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17463 to be added. Now both 'req' and 'ca' can use new objects defined in the
17468 * Add cool BIO that does syslog (or event log on NT).
17470 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17472 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17473 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17474 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17475 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17479 * Add preliminary config info for new extension code.
17483 * Make RSA_NO_PADDING really use no padding.
17485 *Ulf Moeller <ulf@fitug.de>*
17487 * Generate errors when private/public key check is done.
17491 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17492 for some CRL extensions and new objects added.
17496 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17497 key usage extension and fuller support for authority key id.
17501 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17502 padding method for RSA, which is recommended for new applications in PKCS
17503 #1 v2.0 (RFC 2437, October 1998).
17504 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17505 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17506 against Bleichbacher's attack on RSA.
17507 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17510 * Updates to the new SSL compression code
17512 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17514 * Fix so that the version number in the master secret, when passed
17515 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17516 (because the server will not accept higher), that the version number
17517 is 0x03,0x01, not 0x03,0x00
17519 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17521 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17522 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17523 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17527 * Support for RAW extensions where an arbitrary extension can be
17528 created by including its DER encoding. See apps/openssl.cnf for
17533 * Make sure latest Perl versions don't interpret some generated C array
17534 code as Perl array code in the crypto/err/err_genc.pl script.
17536 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17538 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17539 not many people have the assembler. Various Win32 compilation fixes and
17540 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17541 build instructions.
17545 * Modify configure script 'Configure' to automatically create crypto/date.h
17546 file under Win32 and also build pem.h from pem.org. New script
17547 util/mkfiles.pl to create the MINFO file on environments that can't do a
17548 'make files': perl util/mkfiles.pl >MINFO should work.
17552 * Major rework of DES function declarations, in the pursuit of correctness
17553 and purity. As a result, many evil casts evaporated, and some weirdness,
17554 too. You may find this causes warnings in your code. Zapping your evil
17555 casts will probably fix them. Mostly.
17559 * Fix for a typo in asn1.h. Bug fix to object creation script
17560 obj_dat.pl. It considered a zero in an object definition to mean
17561 "end of object": none of the objects in objects.h have any zeros
17562 so it wasn't spotted.
17564 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17566 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17567 Masking (CBCM). In the absence of test vectors, the best I have been able
17568 to do is check that the decrypt undoes the encrypt, so far. Send me test
17569 vectors if you have them.
17573 * Correct calculation of key length for export ciphers (too much space was
17574 allocated for null ciphers). This has not been tested!
17578 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17579 message is now correct (it understands "crypto" and "ssl" on its
17580 command line). There is also now an "update" option. This will update
17581 the util/ssleay.num and util/libeay.num files with any new functions.
17583 perl util/mkdef.pl crypto ssl update
17584 it will update them.
17588 * Overhauled the Perl interface (perl/*):
17589 - ported BN stuff to OpenSSL's different BN library
17590 - made the perl/ source tree CVS-aware
17591 - renamed the package from SSLeay to OpenSSL (the files still contain
17592 their history because I've copied them in the repository)
17593 - removed obsolete files (the test scripts will be replaced
17594 by better Test::Harness variants in the future)
17596 *Ralf S. Engelschall*
17598 * First cut for a very conservative source tree cleanup:
17599 1. merge various obsolete readme texts into doc/ssleay.txt
17600 where we collect the old documents and readme texts.
17601 2. remove the first part of files where I'm already sure that we no
17602 longer need them because of three reasons: either they are just temporary
17603 files which were left by Eric or they are preserved original files where
17604 I've verified that the diff is also available in the CVS via "cvs diff
17605 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17606 the crypto/md/ stuff).
17608 *Ralf S. Engelschall*
17610 * More extension code. Incomplete support for subject and issuer alt
17611 name, issuer and authority key id. Change the i2v function parameters
17612 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17613 what that's for :-) Fix to ASN1 macro which messed up
17614 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17618 * Preliminary support for ENUMERATED type. This is largely copied from the
17623 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17625 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17627 * Make sure `make rehash' target really finds the `openssl' program.
17629 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17631 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17632 like to hear about it if this slows down other processors.
17636 * Add CygWin32 platform information to Configure script.
17638 *Alan Batie <batie@aahz.jf.intel.com>*
17640 * Fixed ms/32all.bat script: `no_asm' -> `no-asm'
17642 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17644 * New program nseq to manipulate netscape certificate sequences
17648 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17653 * Fixes to BN code. Previously the default was to define BN_RECURSION
17654 but the BN code had some problems that would cause failures when
17655 doing certificate verification and some other functions.
17657 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17659 * Add ASN1 and PEM code to support netscape certificate sequences.
17663 * Add ASN1 and PEM code to support netscape certificate sequences.
17667 * Add several PKIX and private extended key usage OIDs.
17671 * Modify the 'ca' program to handle the new extension code. Modify
17672 openssl.cnf for new extension format, add comments.
17676 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17677 and add a sample to openssl.cnf so req -x509 now adds appropriate
17682 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17683 error code, add initial support to X509_print() and x509 application.
17687 * Takes a deep breath and start adding X509 V3 extension support code. Add
17688 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17689 stuff is currently isolated and isn't even compiled yet.
17693 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17694 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17695 Removed the versions check from X509 routines when loading extensions:
17696 this allows certain broken certificates that don't set the version
17697 properly to be processed.
17701 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17702 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17703 can still be regenerated with "make depend".
17707 * Spelling mistake in C version of CAST-128.
17709 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17711 * Changes to the error generation code. The perl script err-code.pl
17712 now reads in the old error codes and retains the old numbers, only
17713 adding new ones if necessary. It also only changes the .err files if new
17714 codes are added. The makefiles have been modified to only insert errors
17715 when needed (to avoid needlessly modifying header files). This is done
17716 by only inserting errors if the .err file is newer than the auto generated
17717 C file. To rebuild all the error codes from scratch (the old behaviour)
17718 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17719 or delete all the .err files.
17723 * CAST-128 was incorrectly implemented for short keys. The C version has
17724 been fixed, but is untested. The assembler versions are also fixed, but
17725 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17726 to regenerate it if needed.
17727 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17728 Hagino <itojun@kame.net>*
17730 * File was opened incorrectly in randfile.c.
17732 *Ulf Möller <ulf@fitug.de>*
17734 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17735 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17736 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17737 al: it's just almost always a UTCTime. Note this patch adds new error
17738 codes so do a "make errors" if there are problems.
17742 * Correct Linux 1 recognition in config.
17744 *Ulf Möller <ulf@fitug.de>*
17746 * Remove pointless MD5 hash when using DSA keys in ca.
17748 *Anonymous <nobody@replay.com>*
17750 * Generate an error if given an empty string as a cert directory. Also
17751 generate an error if handed NULL (previously returned 0 to indicate an
17752 error, but didn't set one).
17754 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17756 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17760 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17761 parameters. This was causing a warning which killed off the Win32 compile.
17765 * Remove C++ style comments from crypto/bn/bn_local.h.
17767 *Neil Costigan <neil.costigan@celocom.com>*
17769 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17770 based on a text string, looking up short and long names and finally
17771 "dot" format. The "dot" format stuff didn't work. Added new function
17772 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17773 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17774 OID is not part of the table.
17778 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17779 X509_LOOKUP_by_alias().
17783 * Sort openssl functions by name.
17787 * Get the gendsa program working (hopefully) and add it to app list. Remove
17788 encryption from sample DSA keys (in case anyone is interested the password
17793 * Make _all_ *_free functions accept a NULL pointer.
17795 *Frans Heymans <fheymans@isaserver.be>*
17797 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17800 *Anonymous <nobody@replay.com>*
17802 * s_server should send the CAfile as acceptable CAs, not its own cert.
17804 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17806 * Don't blow it for numeric -newkey arguments to apps/req.
17808 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17810 * Temp key "for export" tests were wrong in s3_srvr.c.
17812 *Anonymous <nobody@replay.com>*
17814 * Add prototype for temp key callback functions
17815 SSL_CTX_set_tmp_{rsa,dh}_callback().
17819 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17820 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17824 * X509_name_add_entry() freed the wrong thing after an error.
17826 *Arne Ansper <arne@ats.cyber.ee>*
17828 * rsa_eay.c would attempt to free a NULL context.
17830 *Arne Ansper <arne@ats.cyber.ee>*
17832 * BIO_s_socket() had a broken should_retry() on Windoze.
17834 *Arne Ansper <arne@ats.cyber.ee>*
17836 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17838 *Arne Ansper <arne@ats.cyber.ee>*
17840 * Make sure the already existing X509_STORE->depth variable is initialized
17841 in X509_STORE_new(), but document the fact that this variable is still
17842 unused in the certificate verification process.
17844 *Ralf S. Engelschall*
17846 * Fix the various library and apps files to free up pkeys obtained from
17847 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17851 * Fix reference counting in X509_PUBKEY_get(). This makes
17852 demos/maurice/example2.c work, amongst others, probably.
17854 *Steve Henson and Ben Laurie*
17856 * First cut of a cleanup for apps/. First the `ssleay' program is now named
17857 `openssl' and second, the shortcut symlinks for the `openssl <command>'
17858 are no longer created. This way we have a single and consistent command
17859 line interface `openssl <command>', similar to `cvs <command>'.
17861 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
17863 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
17864 BIT STRING wrapper always have zero unused bits.
17868 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
17872 * Make the top-level INSTALL documentation easier to understand.
17876 * Makefiles updated to exit if an error occurs in a sub-directory
17877 make (including if user presses ^C) [Paul Sutton]
17879 * Make Montgomery context stuff explicit in RSA data structure.
17883 * Fix build order of pem and err to allow for generated pem.h.
17887 * Fix renumbering bug in X509_NAME_delete_entry().
17891 * Enhanced the err-ins.pl script so it makes the error library number
17892 global and can add a library name. This is needed for external ASN1 and
17893 other error libraries.
17897 * Fixed sk_insert which never worked properly.
17901 * Fix ASN1 macros so they can handle indefinite length constructed
17902 EXPLICIT tags. Some non standard certificates use these: they can now
17907 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
17908 into a single doc/ssleay.txt bundle. This way the information is still
17909 preserved but no longer messes up this directory. Now it's new room for
17910 the new set of documentation files.
17912 *Ralf S. Engelschall*
17914 * SETs were incorrectly DER encoded. This was a major pain, because they
17915 shared code with SEQUENCEs, which aren't coded the same. This means that
17916 almost everything to do with SETs or SEQUENCEs has either changed name or
17917 number of arguments.
17919 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
17921 * Fix test data to work with the above.
17925 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
17926 was already fixed by Eric for 0.9.1 it seems.
17928 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
17930 * Autodetect FreeBSD3.
17934 * Fix various bugs in Configure. This affects the following platforms:
17938 unixware-2.0-pentium
17943 * Eliminate generated files from CVS. Reorder tests to regenerate files
17944 before they are needed.
17948 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
17953 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998] ###
17955 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
17956 changed SSLeay to OpenSSL in version strings.
17958 *Ralf S. Engelschall*
17960 * Some fixups to the top-level documents.
17964 * Fixed the nasty bug where rsaref.h was not found under compile-time
17965 because the symlink to include/ was missing.
17967 *Ralf S. Engelschall*
17969 * Incorporated the popular no-RSA/DSA-only patches
17970 which allow to compile a RSA-free SSLeay.
17972 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
17974 * Fixed nasty rehash problem under `make -f Makefile.ssl links'
17975 when "ssleay" is still not found.
17977 *Ralf S. Engelschall*
17979 * Added more platforms to Configure: Cray T3E, HPUX 11,
17981 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
17983 * Updated the README file.
17985 *Ralf S. Engelschall*
17987 * Added various .cvsignore files in the CVS repository subdirs
17988 to make a "cvs update" really silent.
17990 *Ralf S. Engelschall*
17992 * Recompiled the error-definition header files and added
17993 missing symbols to the Win32 linker tables.
17995 *Ralf S. Engelschall*
17997 * Cleaned up the top-level documents;
17998 o new files: CHANGES and LICENSE
17999 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18000 o merged COPYRIGHT into LICENSE
18001 o removed obsolete TODO file
18002 o renamed MICROSOFT to INSTALL.W32
18004 *Ralf S. Engelschall*
18006 * Removed dummy files from the 0.9.1b source tree:
18007 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18008 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18009 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18010 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18011 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18013 *Ralf S. Engelschall*
18015 * Added various platform portability fixes.
18019 * The Genesis of the OpenSSL rpject:
18020 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18021 Young and Tim J. Hudson created while they were working for C2Net until
18024 *The OpenSSL Project*
18027 ### Changes between 0.9.0b and 0.9.1b [not released] ###
18029 * Updated a few CA certificates under certs/
18033 * Changed some BIGNUM api stuff.
18037 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18038 DGUX x86, Linux Alpha, etc.
18042 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18043 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18048 * Add -strparse option to asn1pars program which parses nested
18051 *Dr Stephen Henson <shenson@bigfoot.com>*
18053 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18057 * DSA fix for "ca" program.
18061 * Added "-genkey" option to "dsaparam" program.
18065 * Added RIPE MD160 (rmd160) message digest.
18069 * Added -a (all) option to "ssleay version" command.
18073 * Added PLATFORM define which is the id given to Configure.
18077 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18081 * Extended the ASN.1 parser routines.
18085 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18089 * Added a BN_CTX to the BN library.
18093 * Fixed the weak key values in DES library
18097 * Changed API in EVP library for cipher aliases.
18101 * Added support for RC2/64bit cipher.
18105 * Converted the lhash library to the crypto/mem.c functions.
18109 * Added more recognized ASN.1 object ids.
18113 * Added more RSA padding checks for SSL/TLS.
18117 * Added BIO proxy/filter functionality.
18121 * Added extra_certs to SSL_CTX which can be used
18122 send extra CA certificates to the client in the CA cert chain sending
18123 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18127 * Now Fortezza is denied in the authentication phase because
18128 this is key exchange mechanism is not supported by SSLeay at all.
18132 * Additional PKCS1 checks.
18136 * Support the string "TLSv1" for all TLS v1 ciphers.
18140 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18141 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18145 * Fixed a few memory leaks.
18149 * Fixed various code and comment typos.
18153 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18154 bytes sent in the client random.
18156 *Edward Bishop <ebishop@spyglass.com>*
18161 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18162 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18163 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18164 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18165 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18166 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18167 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18168 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18169 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18170 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18171 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18172 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18173 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18174 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18175 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18176 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18177 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18178 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18179 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18180 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18181 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18182 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18183 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18184 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18185 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18186 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18187 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18188 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18189 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18190 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18191 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18192 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18193 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18194 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18195 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18196 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18197 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18198 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18199 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18200 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18201 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18202 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18203 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18204 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18205 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18206 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18207 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18208 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18209 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18210 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18211 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18212 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18213 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18214 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18215 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18216 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18217 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18218 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18219 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18220 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18221 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18222 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18223 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18224 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18225 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18226 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18227 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18228 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18229 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18230 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18231 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18232 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18233 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18234 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18235 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18236 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18237 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18238 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18239 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18240 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18241 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18242 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18243 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18244 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18245 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18246 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18247 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18248 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18249 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18250 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18251 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18252 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18253 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18254 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18255 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18256 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18257 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18258 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18259 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18260 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18261 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18262 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18263 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18264 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18265 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18266 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18267 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18268 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18269 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18270 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18271 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18272 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18273 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18274 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18275 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18276 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18277 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18278 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18279 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18280 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18281 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18282 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18283 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18284 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18285 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18286 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18287 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18288 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18289 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18290 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18291 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18292 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18293 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18294 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18295 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18296 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18297 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18298 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18299 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18300 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18301 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18302 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18303 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18304 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18305 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18306 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18307 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18308 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18309 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18310 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18311 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18312 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18313 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18314 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18315 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18316 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18317 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18318 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18319 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18320 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655