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 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
28 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
29 internal keys, if they correspond to one of those built in types.
33 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
34 contain a provider side internal key.
38 * `ASN1_verify()`, `ASN1_digest()` and `ASN1_sign()` have been deprecated.
39 They are old functions that we don't use, and that you could disable with
40 the macro `NO_ASN1_OLD`. This goes all the way back to OpenSSL 0.9.7.
44 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
45 have been converted to Markdown with the goal to produce documents
46 which not only look pretty when viewed online in the browser, but
47 remain well readable inside a plain text editor.
49 To achieve this goal, a 'minimalistic' Markdown style has been applied
50 which avoids formatting elements that interfere too much with the
51 reading flow in the text file. For example, it
53 * avoids [ATX headings][] and uses [setext headings][] instead
54 (which works for `<h1>` and `<h2>` headings only).
55 * avoids [inline links][] and uses [reference links][] instead.
56 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
58 [ATX headings]: https://github.github.com/gfm/#atx-headings
59 [setext headings]: https://github.github.com/gfm/#setext-headings
60 [inline links]: https://github.github.com/gfm/#inline-link
61 [reference links]: https://github.github.com/gfm/#reference-link
62 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
63 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
68 * The test suite is changed to preserve results of each test recipe.
69 A new directory test-runs/ with subdirectories named like the
70 test recipes are created in the build tree for this purpose.
74 * The command line utilities ecparam and ec have been deprecated. Instead
75 use the pkeyparam, pkey and genpkey programs.
79 * All of the low level RSA functions have been deprecated including:
81 RSA_new_method, RSA_bits, RSA_size, RSA_security_bits,
82 RSA_get0_pss_params, RSA_get_version, RSA_get0_engine,
83 RSA_generate_key_ex, RSA_generate_multi_prime_key,
84 RSA_X931_derive_ex, RSA_X931_generate_key_ex, RSA_check_key,
85 RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
86 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
87 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
88 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
89 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING,
90 RSA_blinding_on, RSA_blinding_off, RSA_setup_blinding,
91 RSA_padding_add_PKCS1_type_1, RSA_padding_check_PKCS1_type_1,
92 RSA_padding_add_PKCS1_type_2, RSA_padding_check_PKCS1_type_2,
93 PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP, RSA_padding_check_PKCS1_OAEP,
94 RSA_padding_add_PKCS1_OAEP_mgf1, RSA_padding_check_PKCS1_OAEP_mgf1,
95 RSA_padding_add_SSLv23, RSA_padding_check_SSLv23,
96 RSA_padding_add_none, RSA_padding_check_none, RSA_padding_add_X931,
97 RSA_padding_check_X931, RSA_X931_hash_id, RSA_verify_PKCS1_PSS,
98 RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
99 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
100 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
101 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
102 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
103 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
104 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
105 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
106 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
107 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
108 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
109 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
110 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
112 Use of these low level functions has been informally discouraged for a long
113 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
114 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
115 L<EVP_PKEY_decrypt(3)>.
119 * X509 certificates signed using SHA1 are no longer allowed at security
121 In TLS/SSL the default security level is 1. It can be set either
122 using the cipher string with @SECLEVEL, or calling
123 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
124 a call to SSL_CTX_use_certificate() will fail if the security level is not
126 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
127 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
132 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
133 deprecated. Instead use the pkeyparam, pkey, genpkey and pkeyparam
134 programs respectively.
138 * All of the low level DH functions have been deprecated including:
140 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
141 DH_new_method, DH_bits, DH_size, DH_security_bits, DH_get_ex_new_index,
142 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
143 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
144 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
145 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
146 DH_KDF_X9_42, DH_get0_engine, DH_get_length, DH_set_length, DH_meth_new,
147 DH_meth_free, DH_meth_dup, DH_meth_get0_name, DH_meth_set1_name,
148 DH_meth_get_flags, DH_meth_set_flags, DH_meth_get0_app_data,
149 DH_meth_set0_app_data, DH_meth_get_generate_key,
150 DH_meth_set_generate_key, DH_meth_get_compute_key,
151 DH_meth_set_compute_key, DH_meth_get_bn_mod_exp,
152 DH_meth_set_bn_mod_exp, DH_meth_get_init, DH_meth_set_init,
153 DH_meth_get_finish, DH_meth_set_finish, DH_meth_get_generate_params
154 and DH_meth_set_generate_params.
156 Use of these low level functions has been informally discouraged for a long
157 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
158 and L<EVP_PKEY_derive(3)>.
162 * All of the low level DSA functions have been deprecated including:
164 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
165 DSA_get_default_method, DSA_set_method, DSA_get_method, DSA_new_method,
166 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
167 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
168 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
169 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
170 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
171 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
172 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
173 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
174 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
175 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
176 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
178 Use of these low level functions has been informally discouraged for a long
179 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
180 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
184 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
185 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
186 This means that applications don't have to look at the curve NID and
187 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
188 However, they still can, that EVP_PKEY_set_alias_type() call acts as
189 a no-op when the EVP_PKEY is already of the given type.
191 Parameter and key generation is also reworked to make it possible
192 to generate EVP_PKEY_SM2 parameters and keys without having to go
193 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
194 However, code that does the latter will still work as before.
198 * Deprecated low level ECDH and ECDSA functions. These include:
200 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
201 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
204 Use of these low level functions has been informally discouraged for a long
205 time. Instead applications should use the EVP_PKEY_derive(3),
206 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
210 * Deprecated the EC_KEY_METHOD functions. These include:
212 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
213 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
214 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
215 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
216 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
217 EC_KEY_METHOD_get_verify.
219 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
223 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
224 and EVP_PKEY_decrypt() instead.
225 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
226 and EVP_PKEY_encrypt() instead.
230 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
231 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
232 a new formulation to include all the things it can be used for,
233 as well as words of caution.
237 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
238 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
242 * All of the low level HMAC functions have been deprecated including:
244 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
245 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
248 Use of these low level functions has been informally discouraged for a long
249 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
250 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
251 and L<EVP_MAC_final(3)>.
255 * Over two thousand fixes were made to the documentation, including:
256 - Common options (such as -rand/-writerand, TLS version control, etc)
257 were refactored and point to newly-enhanced descriptions in openssl.pod.
258 - Added style conformance for all options (with help from Richard Levitte),
259 documented all reported missing options, added a CI build to check
260 that all options are documented and that no unimplemented options
262 - Documented some internals, such as all use of environment variables.
263 - Addressed all internal broken L<> references.
267 * All of the low level CMAC functions have been deprecated including:
269 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
270 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
272 Use of these low level functions has been informally discouraged for a long
273 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
274 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
275 and L<EVP_MAC_final(3)>.
279 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
280 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
283 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
284 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
285 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
286 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
287 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
288 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
289 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
290 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
291 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
292 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
294 Use of these low level functions has been informally discouraged
295 for a long time. Applications should use the EVP_DigestInit_ex(3),
296 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
300 * Corrected the documentation of the return values from the EVP_DigestSign*
301 set of functions. The documentation mentioned negative values for some
302 errors, but this was never the case, so the mention of negative values
305 Code that followed the documentation and thereby check with something
306 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
310 * All of the low level cipher functions have been deprecated including:
312 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
313 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
314 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
315 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
316 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
317 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
318 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
319 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
320 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
321 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
322 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
323 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
324 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
325 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
326 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
327 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
328 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
329 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
330 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
331 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
332 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
333 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
334 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
335 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
336 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
337 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
338 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
339 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
340 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
342 Use of these low level functions has been informally discouraged for
343 a long time. Applications should use the high level EVP APIs, e.g.
344 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
345 equivalently named decrypt functions instead.
347 *Matt Caswell and Paul Dale*
349 * Removed include/openssl/opensslconf.h.in and replaced it with
350 include/openssl/configuration.h.in, which differs in not including
351 <openssl/macros.h>. A short header include/openssl/opensslconf.h
352 was added to include both.
354 This allows internal hacks where one might need to modify the set
355 of configured macros, for example this if deprecated symbols are
356 still supposed to be available internally:
358 #include <openssl/configuration.h>
360 #undef OPENSSL_NO_DEPRECATED
361 #define OPENSSL_SUPPRESS_DEPRECATED
363 #include <openssl/macros.h>
365 This should not be used by applications that use the exported
366 symbols, as that will lead to linking errors.
370 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
371 used in exponentiation with 512-bit moduli. No EC algorithms are
372 affected. Analysis suggests that attacks against 2-prime RSA1024,
373 3-prime RSA1536, and DSA1024 as a result of this defect would be very
374 difficult to perform and are not believed likely. Attacks against DH512
375 are considered just feasible. However, for an attack the target would
376 have to re-use the DH512 private key, which is not recommended anyway.
377 Also applications directly using the low level API BN_mod_exp may be
378 affected if they use BN_FLG_CONSTTIME.
383 * Most memory-debug features have been deprecated, and the functionality
384 replaced with no-ops.
388 * Introduced a new method type and API, OSSL_SERIALIZER, to
389 represent generic serializers. An implementation is expected to
390 be able to serialize an object associated with a given name (such
391 as an algorithm name for an asymmetric key) into forms given by
392 implementation properties.
394 Serializers are primarily used from inside libcrypto, through
395 calls to functions like EVP_PKEY_print_private(),
396 PEM_write_bio_PrivateKey() and similar.
398 Serializers are specified in such a way that they can be made to
399 directly handle the provider side portion of an object, if this
400 provider side part comes from the same provider as the serializer
401 itself, but can also be made to handle objects in parametrized
402 form (as an OSSL_PARAM array of data). This allows a provider to
403 offer generic serializers as a service for any other provider.
407 * Added a .pragma directive to the syntax of configuration files, to
408 allow varying behavior in a supported and predictable manner.
409 Currently added pragma:
413 This allows dollar signs to be a keyword character unless it's
414 followed by a opening brace or parenthesis. This is useful for
415 platforms where dollar signs are commonly used in names, such as
416 volume names and system directory names on VMS.
420 * Added functionality to create an EVP_PKEY from user data. This
421 is effectively the same as creating a RSA, DH or DSA object and
422 then assigning them to an EVP_PKEY, but directly using algorithm
423 agnostic EVP functions. A benefit is that this should be future
424 proof for public key algorithms to come.
428 * Change the interpretation of the '--api' configuration option to
429 mean that this is a desired API compatibility level with no
430 further meaning. The previous interpretation, that this would
431 also mean to remove all deprecated symbols up to and including
432 the given version, no requires that 'no-deprecated' is also used
433 in the configuration.
435 When building applications, the desired API compatibility level
436 can be set with the OPENSSL_API_COMPAT macro like before. For
437 API compatibility version below 3.0, the old style numerical
438 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
439 For version 3.0 and on, the value is expected to be the decimal
440 value calculated from the major and minor version like this:
442 MAJOR * 10000 + MINOR * 100
446 -DOPENSSL_API_COMPAT=30000 For 3.0
447 -DOPENSSL_API_COMPAT=30200 For 3.2
449 To hide declarations that are deprecated up to and including the
450 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
451 given when building the application as well.
455 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
456 access to certificate and CRL stores via URIs and OSSL_STORE
459 This adds the following functions:
461 - X509_LOOKUP_store()
462 - X509_STORE_load_file()
463 - X509_STORE_load_path()
464 - X509_STORE_load_store()
465 - SSL_add_store_cert_subjects_to_stack()
466 - SSL_CTX_set_default_verify_store()
467 - SSL_CTX_load_verify_file()
468 - SSL_CTX_load_verify_dir()
469 - SSL_CTX_load_verify_store()
471 Also, the following functions are now deprecated:
473 - X509_STORE_load_locations() (use X509_STORE_load_file(),
474 X509_STORE_load_path() or X509_STORE_load_store() instead)
475 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
476 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
480 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
481 The presence of this system service is determined at run-time.
485 * Added functionality to create an EVP_PKEY context based on data
486 for methods from providers. This takes an algorithm name and a
487 property query string and simply stores them, with the intent
488 that any operation that uses this context will use those strings
489 to fetch the needed methods implicitly, thereby making the port
490 of application written for pre-3.0 OpenSSL easier.
494 * The undocumented function NCONF_WIN32() has been deprecated; for
495 conversion details see the HISTORY section of doc/man5/config.pod
499 * Introduced the new functions EVP_DigestSignInit_ex() and
500 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
501 EVP_DigestVerifyUpdate() have been converted to functions. See the man
502 pages for further details.
506 * Over two thousand fixes were made to the documentation, including:
507 adding missing command flags, better style conformance, documentation
510 *Rich Salz, Richard Levitte*
512 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
513 X25519, X448, Ed25519 and Ed448.
517 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
522 * Deprecated the public definition of ERR_STATE as well as the function
523 ERR_get_state(). This is done in preparation of making ERR_STATE an
528 * Added ERR functionality to give callers access to the stored function
529 names that have replaced the older function code based functions.
531 New functions are ERR_get_error_func(), ERR_peek_error_func(),
532 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
533 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
534 and ERR_peek_last_error_all().
536 These functions have become deprecated: ERR_get_error_line_data(),
537 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
538 ERR_func_error_string().
542 * Extended testing to be verbose for failing tests only. The make variables
543 VERBOSE_FAILURE or VF can be used to enable this:
545 $ make VF=1 test # Unix
546 $ mms /macro=(VF=1) test ! OpenVMS
547 $ nmake VF=1 test # Windows
552 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
553 used even when parsing explicit parameters, when loading a serialized key
554 or calling `EC_GROUP_new_from_ecpkparameters()`/
555 `EC_GROUP_new_from_ecparameters()`.
556 This prevents bypass of security hardening and performance gains,
557 especially for curves with specialized EC_METHODs.
558 By default, if a key encoded with explicit parameters is loaded and later
559 serialized, the output is still encoded with explicit parameters, even if
560 internally a "named" EC_GROUP is used for computation.
564 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
565 this change, EC_GROUP_set_generator would accept order and/or cofactor as
566 NULL. After this change, only the cofactor parameter can be NULL. It also
567 does some minimal sanity checks on the passed order.
572 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
573 An attack is simple, if the first CMS_recipientInfo is valid but the
574 second CMS_recipientInfo is chosen ciphertext. If the second
575 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
576 encryption key will be replaced by garbage, and the message cannot be
577 decoded, but if the RSA decryption fails, the correct encryption key is
578 used and the recipient will not notice the attack.
579 As a work around for this potential attack the length of the decrypted
580 key must be equal to the cipher default key length, in case the
581 certifiate is not given and all recipientInfo are tried out.
582 The old behaviour can be re-enabled in the CMS code by setting the
583 CMS_DEBUG_DECRYPT flag.
587 * Early start up entropy quality from the DEVRANDOM seed source has been
588 improved for older Linux systems. The RAND subsystem will wait for
589 /dev/random to be producing output before seeding from /dev/urandom.
590 The seeded state is stored for future library initialisations using
591 a system global shared memory segment. The shared memory identifier
592 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
593 the desired value. The default identifier is 114.
597 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
598 when primes for RSA keys are computed.
599 Since we previously always generated primes == 2 (mod 3) for RSA keys,
600 the 2-prime and 3-prime RSA modules were easy to distinguish, since
601 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
602 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
603 This avoids possible fingerprinting of newly generated RSA modules.
607 * Correct the extended master secret constant on EBCDIC systems. Without this
608 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
609 negotiate EMS will fail. Unfortunately this also means that TLS connections
610 between EBCDIC systems with this fix, and EBCDIC systems without this
611 fix will fail if they negotiate EMS.
615 * Changed the library initialisation so that the config file is now loaded
616 by default. This was already the case for libssl. It now occurs for both
617 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
618 OPENSSL_init_crypto() to suppress automatic loading of a config file.
622 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
623 where the former acts as a replacement for ERR_put_error(), and the
624 latter replaces the combination ERR_put_error()+ERR_add_error_data().
625 ERR_raise_data() adds more flexibility by taking a format string and
626 an arbitrary number of arguments following it, to be processed with
631 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
632 to check if a named provider is loaded and available. When called, it
633 will also activate all fallback providers if such are still present.
637 * Enforce a minimum DH modulus size of 512 bits.
641 * Changed DH parameters to generate the order q subgroup instead of 2q.
642 Previously generated DH parameters are still accepted by DH_check
643 but DH_generate_key works around that by clearing bit 0 of the
644 private key for those. This avoids leaking bit 0 of the private key.
648 * Significantly reduce secure memory usage by the randomness pools.
652 * {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
657 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
658 algorithms. An implementation of a key exchange algorithm can be obtained
659 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
660 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
661 the older EVP_PKEY_derive_init() function. See the man pages for the new
662 functions for further details.
666 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
670 * Removed the function names from error messages and deprecated the
673 * Removed NextStep support and the macro OPENSSL_UNISTD
677 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
678 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
679 Also removed "export var as function" capability; we do not export
680 variables, only functions.
684 * RC5_32_set_key has been changed to return an int type, with 0 indicating
685 an error and 1 indicating success. In previous versions of OpenSSL this
686 was a void type. If a key was set longer than the maximum possible this
691 * Support SM2 signing and verification schemes with X509 certificate.
695 * Use SHA256 as the default digest for TS query in the ts app.
699 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
700 This checks that the salt length is at least 128 bits, the derived key
701 length is at least 112 bits, and that the iteration count is at least 1000.
702 For backwards compatibility these checks are disabled by default in the
703 default provider, but are enabled by default in the fips provider.
704 To enable or disable these checks use the control
705 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
709 * Default cipher lists/suites are now available via a function, the
710 #defines are deprecated.
714 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
715 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
716 for Windows Store apps easier. Also, the "no-uplink" option has been added.
720 * Join the directories crypto/x509 and crypto/x509v3
724 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
725 This changes the size when using the genpkey app when no size is given. It
726 fixes an omission in earlier changes that changed all RSA, DSA and DH
727 generation apps to use 2048 bits by default.
731 * Added command 'openssl kdf' that uses the EVP_KDF API.
735 * Added command 'openssl mac' that uses the EVP_MAC API.
739 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
740 as default directories. Also added the command 'openssl info'
741 for scripting purposes.
745 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
746 deprecated. These undocumented functions were never integrated into the EVP
747 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
748 Bi-directional IGE mode. These modes were never formally standardised and
749 usage of these functions is believed to be very small. In particular
750 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
751 is ever used. The security implications are believed to be minimal, but
752 this issue was never fixed for backwards compatibility reasons. New code
753 should not use these modes.
757 * Add prediction resistance to the DRBG reseeding process.
761 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
762 mandated by IEEE Std 1619-2018.
766 * Added newline escaping functionality to a filename when using openssl dgst.
767 This output format is to replicate the output format found in the '*sum'
768 checksum programs. This aims to preserve backward compatibility.
770 *Matt Eaton, Richard Levitte, and Paul Dale*
772 * Removed the heartbeat message in DTLS feature, as it has very
773 little usage and doesn't seem to fulfill a valuable purpose.
774 The configuration option is now deprecated.
778 * Changed the output of 'openssl {digestname} < file' to display the
779 digest name in its output.
783 * Added a new generic trace API which provides support for enabling
784 instrumentation through trace output. This feature is mainly intended
785 as an aid for developers and is disabled by default. To utilize it,
786 OpenSSL needs to be configured with the `enable-trace` option.
788 If the tracing API is enabled, the application can activate trace output
789 by registering BIOs as trace channels for a number of tracing and debugging
792 The 'openssl' application has been expanded to enable any of the types
793 available via environment variables defined by the user, and serves as
794 one possible example on how to use this functionality.
796 *Richard Levitte & Matthias St. Pierre*
798 * Added build tests for C++. These are generated files that only do one
799 thing, to include one public OpenSSL head file each. This tests that
800 the public header files can be usefully included in a C++ application.
802 This test isn't enabled by default. It can be enabled with the option
803 'enable-buildtest-c++'.
807 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
811 * Add KMAC to EVP_MAC.
815 * Added property based algorithm implementation selection framework to
820 * Added SCA hardening for modular field inversion in EC_GROUP through
821 a new dedicated field_inv() pointer in EC_METHOD.
822 This also addresses a leakage affecting conversions from projective
823 to affine coordinates.
825 *Billy Bob Brumley, Nicola Tuveri*
827 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
828 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
829 those algorithms that were already supported through the EVP_PKEY API
830 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
831 and scrypt are now wrappers that call EVP_KDF.
835 * Build devcrypto engine as a dynamic engine.
839 * Add keyed BLAKE2 to EVP_MAC.
843 * Fix a bug in the computation of the endpoint-pair shared secret used
844 by DTLS over SCTP. This breaks interoperability with older versions
845 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
846 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
847 interoperability with such broken implementations. However, enabling
848 this switch breaks interoperability with correct implementations.
850 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
851 re-used X509_PUBKEY object if the second PUBKEY is malformed.
855 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
859 * Change the license to the Apache License v2.0.
863 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
865 - Major releases (indicated by incrementing the MAJOR release number)
866 may introduce incompatible API/ABI changes.
867 - Minor releases (indicated by incrementing the MINOR release number)
868 may introduce new features but retain API/ABI compatibility.
869 - Patch releases (indicated by incrementing the PATCH number)
870 are intended for bug fixes and other improvements of existing
871 features only (like improving performance or adding documentation)
872 and retain API/ABI compatibility.
876 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
880 * Remove the 'dist' target and add a tarball building script. The
881 'dist' target has fallen out of use, and it shouldn't be
882 necessary to configure just to create a source distribution.
886 * Recreate the OS390-Unix config target. It no longer relies on a
887 special script like it did for OpenSSL pre-1.1.0.
891 * Instead of having the source directories listed in Configure, add
892 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
897 * Add GMAC to EVP_MAC.
901 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
905 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
906 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
907 to facilitate the continued use of MACs through raw private keys in
908 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
912 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
913 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
917 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
918 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
919 are retained for backwards compatibility.
923 * AES-XTS mode now enforces that its two keys are different to mitigate
924 the attacked described in "Efficient Instantiations of Tweakable
925 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
926 Details of this attack can be obtained from:
927 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
931 * Rename the object files, i.e. give them other names than in previous
932 versions. Their names now include the name of the final product, as
933 well as its type mnemonic (bin, lib, shlib).
937 * Added new option for 'openssl list', '-objects', which will display the
938 list of built in objects, i.e. OIDs with names.
942 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
943 improves application performance by removing data copies and providing
944 applications with zero-copy system calls such as sendfile and splice.
952 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx] ###
955 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020] ###
957 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
958 while reading in libssl then we would report an error back to the
959 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
960 an error to the stack (which means we instead return SSL_ERROR_SSL) and
961 therefore give a hint as to what went wrong.
965 * Check that ed25519 and ed448 are allowed by the security level. Previously
966 signature algorithms not using an MD were not being checked that they were
967 allowed by the security level.
971 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
972 was not quite right. The behaviour was not consistent between resumption
973 and normal handshakes, and also not quite consistent with historical
974 behaviour. The behaviour in various scenarios has been clarified and
975 it has been updated to make it match historical behaviour as closely as
980 * *[VMS only]* The header files that the VMS compilers include automatically,
981 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
982 that the C++ compiler doesn't understand. This is a shortcoming in the
983 compiler, but can be worked around with `__cplusplus` guards.
985 C++ applications that use OpenSSL libraries must be compiled using the
986 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
987 functions. Otherwise, only functions with symbols of less than 31
988 characters can be used, as the linker will not be able to successfully
989 resolve symbols with longer names.
993 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
994 The presence of this system service is determined at run-time.
998 * Added newline escaping functionality to a filename when using openssl dgst.
999 This output format is to replicate the output format found in the '*sum'
1000 checksum programs. This aims to preserve backward compatibility.
1002 *Matt Eaton, Richard Levitte, and Paul Dale*
1004 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1009 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019] ###
1011 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1012 number generator (RNG). This was intended to include protection in the
1013 event of a fork() system call in order to ensure that the parent and child
1014 processes did not share the same RNG state. However this protection was not
1015 being used in the default case.
1017 A partial mitigation for this issue is that the output from a high
1018 precision timer is mixed into the RNG state so the likelihood of a parent
1019 and child process sharing state is significantly reduced.
1021 If an application already calls OPENSSL_init_crypto() explicitly using
1022 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1025 *Matthias St. Pierre*
1027 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1028 used even when parsing explicit parameters, when loading a serialized key
1029 or calling `EC_GROUP_new_from_ecpkparameters()`/
1030 `EC_GROUP_new_from_ecparameters()`.
1031 This prevents bypass of security hardening and performance gains,
1032 especially for curves with specialized EC_METHODs.
1033 By default, if a key encoded with explicit parameters is loaded and later
1034 serialized, the output is still encoded with explicit parameters, even if
1035 internally a "named" EC_GROUP is used for computation.
1039 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1040 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1041 NULL. After this change, only the cofactor parameter can be NULL. It also
1042 does some minimal sanity checks on the passed order.
1047 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1048 An attack is simple, if the first CMS_recipientInfo is valid but the
1049 second CMS_recipientInfo is chosen ciphertext. If the second
1050 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1051 encryption key will be replaced by garbage, and the message cannot be
1052 decoded, but if the RSA decryption fails, the correct encryption key is
1053 used and the recipient will not notice the attack.
1054 As a work around for this potential attack the length of the decrypted
1055 key must be equal to the cipher default key length, in case the
1056 certifiate is not given and all recipientInfo are tried out.
1057 The old behaviour can be re-enabled in the CMS code by setting the
1058 CMS_DEBUG_DECRYPT flag.
1063 * Early start up entropy quality from the DEVRANDOM seed source has been
1064 improved for older Linux systems. The RAND subsystem will wait for
1065 /dev/random to be producing output before seeding from /dev/urandom.
1066 The seeded state is stored for future library initialisations using
1067 a system global shared memory segment. The shared memory identifier
1068 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1069 the desired value. The default identifier is 114.
1073 * Correct the extended master secret constant on EBCDIC systems. Without this
1074 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1075 negotiate EMS will fail. Unfortunately this also means that TLS connections
1076 between EBCDIC systems with this fix, and EBCDIC systems without this
1077 fix will fail if they negotiate EMS.
1081 * Use Windows installation paths in the mingw builds
1083 Mingw isn't a POSIX environment per se, which means that Windows
1084 paths should be used for installation.
1089 * Changed DH_check to accept parameters with order q and 2q subgroups.
1090 With order 2q subgroups the bit 0 of the private key is not secret
1091 but DH_generate_key works around that by clearing bit 0 of the
1092 private key for those. This avoids leaking bit 0 of the private key.
1096 * Significantly reduce secure memory usage by the randomness pools.
1100 * Revert the DEVRANDOM_WAIT feature for Linux systems
1102 The DEVRANDOM_WAIT feature added a select() call to wait for the
1103 /dev/random device to become readable before reading from the
1104 /dev/urandom device.
1106 It turned out that this change had negative side effects on
1107 performance which were not acceptable. After some discussion it
1108 was decided to revert this feature and leave it up to the OS
1109 resp. the platform maintainer to ensure a proper initialization
1110 during early boot time.
1112 *Matthias St. Pierre*
1114 ### Changes between 1.1.1b and 1.1.1c [28 May 2019] ###
1116 * Add build tests for C++. These are generated files that only do one
1117 thing, to include one public OpenSSL head file each. This tests that
1118 the public header files can be usefully included in a C++ application.
1120 This test isn't enabled by default. It can be enabled with the option
1121 'enable-buildtest-c++'.
1125 * Enable SHA3 pre-hashing for ECDSA and DSA.
1129 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1130 This changes the size when using the genpkey app when no size is given. It
1131 fixes an omission in earlier changes that changed all RSA, DSA and DH
1132 generation apps to use 2048 bits by default.
1136 * Reorganize the manual pages to consistently have RETURN VALUES,
1137 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1138 util/fix-doc-nits accordingly.
1140 *Paul Yang, Joshua Lock*
1142 * Add the missing accessor EVP_PKEY_get0_engine()
1146 * Have apps like 's_client' and 's_server' output the signature scheme
1147 along with other cipher suite parameters when debugging.
1151 * Make OPENSSL_config() error agnostic again.
1155 * Do the error handling in RSA decryption constant time.
1159 * Prevent over long nonces in ChaCha20-Poly1305.
1161 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1162 for every encryption operation. RFC 7539 specifies that the nonce value
1163 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1164 and front pads the nonce with 0 bytes if it is less than 12
1165 bytes. However it also incorrectly allows a nonce to be set of up to 16
1166 bytes. In this case only the last 12 bytes are significant and any
1167 additional leading bytes are ignored.
1169 It is a requirement of using this cipher that nonce values are
1170 unique. Messages encrypted using a reused nonce value are susceptible to
1171 serious confidentiality and integrity attacks. If an application changes
1172 the default nonce length to be longer than 12 bytes and then makes a
1173 change to the leading bytes of the nonce expecting the new value to be a
1174 new unique nonce then such an application could inadvertently encrypt
1175 messages with a reused nonce.
1177 Additionally the ignored bytes in a long nonce are not covered by the
1178 integrity guarantee of this cipher. Any application that relies on the
1179 integrity of these ignored leading bytes of a long nonce may be further
1180 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1181 is safe because no such use sets such a long nonce value. However user
1182 applications that use this cipher directly and set a non-default nonce
1183 length to be longer than 12 bytes may be vulnerable.
1185 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1191 * Add DEVRANDOM_WAIT feature for Linux systems
1193 On older Linux systems where the getrandom() system call is not available,
1194 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1195 Contrary to getrandom(), the /dev/urandom device will not block during
1196 early boot when the kernel CSPRNG has not been seeded yet.
1198 To mitigate this known weakness, use select() to wait for /dev/random to
1199 become readable before reading from /dev/urandom.
1201 * Ensure that SM2 only uses SM3 as digest algorithm
1205 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019] ###
1207 * Change the info callback signals for the start and end of a post-handshake
1208 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1209 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1210 confused by this and assume that a TLSv1.2 renegotiation has started. This
1211 can break KeyUpdate handling. Instead we no longer signal the start and end
1212 of a post handshake message exchange (although the messages themselves are
1213 still signalled). This could break some applications that were expecting
1214 the old signals. However without this KeyUpdate is not usable for many
1219 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018] ###
1221 * Timing vulnerability in DSA signature generation
1223 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1224 timing side channel attack. An attacker could use variations in the signing
1225 algorithm to recover the private key.
1227 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1232 * Timing vulnerability in ECDSA signature generation
1234 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1235 timing side channel attack. An attacker could use variations in the signing
1236 algorithm to recover the private key.
1238 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1243 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1244 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1245 of two gigabytes and the error handling improved.
1247 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1248 categorized as a normal bug, not a security issue, because the DRBG reseeds
1249 automatically and is fully functional even without additional randomness
1250 provided by the application.
1252 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018] ###
1254 * Add a new ClientHello callback. Provides a callback interface that gives
1255 the application the ability to adjust the nascent SSL object at the
1256 earliest stage of ClientHello processing, immediately after extensions have
1257 been collected but before they have been processed. In particular, this
1258 callback can adjust the supported TLS versions in response to the contents
1263 * Add SM2 base algorithm support.
1267 * s390x assembly pack: add (improved) hardware-support for the following
1268 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1269 aes-cfb/cfb8, aes-ecb.
1273 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1274 parameter is no longer accepted, as it leads to a corrupt table. NULL
1275 pem_str is reserved for alias entries only.
1279 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1280 step for prime curves. The new implementation is based on formulae from
1281 differential addition-and-doubling in homogeneous projective coordinates
1282 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1283 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1284 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1285 to work in projective coordinates.
1287 *Billy Bob Brumley, Nicola Tuveri*
1289 * Change generating and checking of primes so that the error rate of not
1290 being prime depends on the intended use based on the size of the input.
1291 For larger primes this will result in more rounds of Miller-Rabin.
1292 The maximal error rate for primes with more than 1080 bits is lowered
1295 *Kurt Roeckx, Annie Yousar*
1297 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1301 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1302 moving between systems, and to avoid confusion when a Windows build is
1303 done with mingw vs with MSVC. For POSIX installs, there's still a
1304 symlink or copy named 'tsget' to avoid that confusion as well.
1308 * Revert blinding in ECDSA sign and instead make problematic addition
1309 length-invariant. Switch even to fixed-length Montgomery multiplication.
1313 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1314 step for binary curves. The new implementation is based on formulae from
1315 differential addition-and-doubling in mixed Lopez-Dahab projective
1316 coordinates, modified to independently blind the operands.
1318 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1320 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1321 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1322 EC_METHODs to implement their own specialized "ladder step", to take
1323 advantage of more favorable coordinate systems or more efficient
1324 differential addition-and-doubling algorithms.
1326 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1328 * Modified the random device based seed sources to keep the relevant
1329 file descriptors open rather than reopening them on each access.
1330 This allows such sources to operate in a chroot() jail without
1331 the associated device nodes being available. This behaviour can be
1332 controlled using RAND_keep_random_devices_open().
1336 * Numerous side-channel attack mitigations have been applied. This may have
1337 performance impacts for some algorithms for the benefit of improved
1338 security. Specific changes are noted in this change log by their respective
1343 * AIX shared library support overhaul. Switch to AIX "natural" way of
1344 handling shared libraries, which means collecting shared objects of
1345 different versions and bitnesses in one common archive. This allows to
1346 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1347 doesn't affect the way 3rd party applications are linked, only how
1348 multi-version installation is managed.
1352 * Make ec_group_do_inverse_ord() more robust and available to other
1353 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1354 mitigations are applied to the fallback BN_mod_inverse().
1355 When using this function rather than BN_mod_inverse() directly, new
1356 EC cryptosystem implementations are then safer-by-default.
1360 * Add coordinate blinding for EC_POINT and implement projective
1361 coordinate blinding for generic prime curves as a countermeasure to
1362 chosen point SCA attacks.
1364 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1366 * Add blinding to ECDSA and DSA signatures to protect against side channel
1367 attacks discovered by Keegan Ryan (NCC Group).
1371 * Enforce checking in the pkeyutl command line app to ensure that the input
1372 length does not exceed the maximum supported digest length when performing
1373 a sign, verify or verifyrecover operation.
1377 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1378 I/O in combination with something like select() or poll() will hang. This
1379 can be turned off again using SSL_CTX_clear_mode().
1380 Many applications do not properly handle non-application data records, and
1381 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1382 around the problems in those applications, but can also break some.
1383 It's recommended to read the manpages about SSL_read(), SSL_write(),
1384 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1385 SSL_CTX_set_read_ahead() again.
1389 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1390 now allow empty (zero character) pass phrases.
1394 * Apply blinding to binary field modular inversion and remove patent
1395 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1399 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1400 binary and prime elliptic curves.
1404 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1405 constant time fixed point multiplication.
1409 * Revise elliptic curve scalar multiplication with timing attack
1410 defenses: ec_wNAF_mul redirects to a constant time implementation
1411 when computing fixed point and variable point multiplication (which
1412 in OpenSSL are mostly used with secret scalars in keygen, sign,
1413 ECDH derive operations).
1414 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1417 * Updated CONTRIBUTING
1421 * Updated DRBG / RAND to request nonce and additional low entropy
1422 randomness from the system.
1424 *Matthias St. Pierre*
1426 * Updated 'openssl rehash' to use OpenSSL consistent default.
1430 * Moved the load of the ssl_conf module to libcrypto, which helps
1431 loading engines that libssl uses before libssl is initialised.
1435 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1439 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1441 *Ingo Schwarze, Rich Salz*
1443 * Added output of accepting IP address and port for 'openssl s_server'
1447 * Added a new API for TLSv1.3 ciphersuites:
1448 SSL_CTX_set_ciphersuites()
1449 SSL_set_ciphersuites()
1453 * Memory allocation failures consistently add an error to the error
1458 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1459 in libcrypto when run as setuid/setgid.
1463 * Load any config file by default when libssl is used.
1467 * Added new public header file <openssl/rand_drbg.h> and documentation
1468 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1470 *Matthias St. Pierre*
1472 * QNX support removed (cannot find contributors to get their approval
1473 for the license change).
1477 * TLSv1.3 replay protection for early data has been implemented. See the
1478 SSL_read_early_data() man page for further details.
1482 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1483 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1484 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1485 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1486 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1487 configuration has been separated out. See the ciphers man page or the
1488 SSL_CTX_set_ciphersuites() man page for more information.
1492 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1493 in responder mode now supports the new "-multi" option, which
1494 spawns the specified number of child processes to handle OCSP
1495 requests. The "-timeout" option now also limits the OCSP
1496 responder's patience to wait to receive the full client request
1497 on a newly accepted connection. Child processes are respawned
1498 as needed, and the CA index file is automatically reloaded
1499 when changed. This makes it possible to run the "ocsp" responder
1500 as a long-running service, making the OpenSSL CA somewhat more
1501 feature-complete. In this mode, most diagnostic messages logged
1502 after entering the event loop are logged via syslog(3) rather than
1507 * Added support for X448 and Ed448. Heavily based on original work by
1512 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1513 objects loaded. This adds the functions OSSL_STORE_expect() and
1514 OSSL_STORE_find() as well as needed tools to construct searches and
1515 get the search data out of them.
1519 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1520 version of OpenSSL should review their configuration settings to ensure
1521 that they are still appropriate for TLSv1.3. For further information see:
1522 https://wiki.openssl.org/index.php/TLS1.3
1526 * Grand redesign of the OpenSSL random generator
1528 The default RAND method now utilizes an AES-CTR DRBG according to
1529 NIST standard SP 800-90Ar1. The new random generator is essentially
1530 a port of the default random generator from the OpenSSL FIPS 2.0
1531 object module. It is a hybrid deterministic random bit generator
1532 using an AES-CTR bit stream and which seeds and reseeds itself
1533 automatically using trusted system entropy sources.
1535 Some of its new features are:
1536 - Support for multiple DRBG instances with seed chaining.
1537 - The default RAND method makes use of a DRBG.
1538 - There is a public and private DRBG instance.
1539 - The DRBG instances are fork-safe.
1540 - Keep all global DRBG instances on the secure heap if it is enabled.
1541 - The public and private DRBG instance are per thread for lock free
1544 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1546 * Changed Configure so it only says what it does and doesn't dump
1547 so much data. Instead, ./configdata.pm should be used as a script
1548 to display all sorts of configuration data.
1552 * Added processing of "make variables" to Configure.
1556 * Added SHA512/224 and SHA512/256 algorithm support.
1560 * The last traces of Netware support, first removed in 1.1.0, have
1565 * Get rid of Makefile.shared, and in the process, make the processing
1566 of certain files (rc.obj, or the .def/.map/.opt files produced from
1567 the ordinal files) more visible and hopefully easier to trace and
1568 debug (or make silent).
1572 * Make it possible to have environment variable assignments as
1573 arguments to config / Configure.
1577 * Add multi-prime RSA (RFC 8017) support.
1581 * Add SM3 implemented according to GB/T 32905-2016
1582 * Jack Lloyd <jack.lloyd@ribose.com>,
1583 Ronald Tse <ronald.tse@ribose.com>,
1584 Erick Borsboom <erick.borsboom@ribose.com> *
1586 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1587 as documented in RFC6066.
1588 Based on a patch from Tomasz Moń
1590 *Filipe Raimundo da Silva*
1592 * Add SM4 implemented according to GB/T 32907-2016.
1593 * Jack Lloyd <jack.lloyd@ribose.com>,
1594 Ronald Tse <ronald.tse@ribose.com>,
1595 Erick Borsboom <erick.borsboom@ribose.com> *
1597 * Reimplement -newreq-nodes and ERR_error_string_n; the
1598 original author does not agree with the license change.
1602 * Add ARIA AEAD TLS support.
1606 * Some macro definitions to support VS6 have been removed. Visual
1607 Studio 6 has not worked since 1.1.0
1611 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1612 without clearing the errors.
1616 * Add "atfork" functions. If building on a system that without
1617 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1618 requirements. The RAND facility now uses/requires this.
1626 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1627 not possible to disable entirely. However, it's still possible to
1628 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1631 To disable, configure with 'no-ui-console'. 'no-ui' is still
1632 possible to use as an alias. Check at compile time with the
1633 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1634 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1638 * Add a STORE module, which implements a uniform and URI based reader of
1639 stores that can contain keys, certificates, CRLs and numerous other
1640 objects. The main API is loosely based on a few stdio functions,
1641 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1642 OSSL_STORE_error and OSSL_STORE_close.
1643 The implementation uses backends called "loaders" to implement arbitrary
1644 URI schemes. There is one built in "loader" for the 'file' scheme.
1648 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1649 then adjusted to work on FreeBSD 8.4 as well.
1650 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1651 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1655 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1656 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1657 error code calls like this:
1659 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1661 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1662 that can be encoded in C. For the foreseeable future, this will only
1665 *Richard Levitte and Tim Hudson*
1667 * Removed BSD cryptodev engine.
1671 * Add a build target 'build_all_generated', to build all generated files
1672 and only that. This can be used to prepare everything that requires
1673 things like perl for a system that lacks perl and then move everything
1674 to that system and do the rest of the build there.
1678 * In the UI interface, make it possible to duplicate the user data. This
1679 can be used by engines that need to retain the data for a longer time
1680 than just the call where this user data is passed.
1684 * Ignore the '-named_curve auto' value for compatibility of applications
1687 *Tomas Mraz <tmraz@fedoraproject.org>*
1689 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1690 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1691 alerts across multiple records (some of which could be empty). In practice
1692 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1693 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1694 support this at all. Supporting it adds significant complexity to the
1695 record layer, and its removal is unlikely to cause interoperability
1700 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1701 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1702 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1707 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1708 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1710 *Richard Levitte, Andy Polyakov*
1712 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1717 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1718 platform rather than 'mingw'.
1722 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1723 success if they are asked to add an object which already exists
1724 in the store. This change cascades to other functions which load
1725 certificates and CRLs.
1729 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1730 facilitate stack unwinding even from assembly subroutines.
1734 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1735 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1739 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1740 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1741 which is the minimum version we support.
1745 * Certificate time validation (X509_cmp_time) enforces stricter
1746 compliance with RFC 5280. Fractional seconds and timezone offsets
1747 are no longer allowed.
1751 * Add support for ARIA
1755 * s_client will now send the Server Name Indication (SNI) extension by
1756 default unless the new "-noservername" option is used. The server name is
1757 based on the host provided to the "-connect" option unless overridden by
1758 using "-servername".
1762 * Add support for SipHash
1766 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1767 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1768 prevent issues where no progress is being made and the peer continually
1769 sends unrecognised record types, using up resources processing them.
1773 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1774 using the algorithm defined in
1775 https://www.akkadia.org/drepper/SHA-crypt.txt
1779 * Heartbeat support has been removed; the ABI is changed for now.
1781 *Richard Levitte, Rich Salz*
1783 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1787 * The RSA "null" method, which was partially supported to avoid patent
1788 issues, has been replaced to always returns NULL.
1796 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019] ###
1798 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1799 used even when parsing explicit parameters, when loading a serialized key
1800 or calling `EC_GROUP_new_from_ecpkparameters()`/
1801 `EC_GROUP_new_from_ecparameters()`.
1802 This prevents bypass of security hardening and performance gains,
1803 especially for curves with specialized EC_METHODs.
1804 By default, if a key encoded with explicit parameters is loaded and later
1805 serialized, the output is still encoded with explicit parameters, even if
1806 internally a "named" EC_GROUP is used for computation.
1810 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1811 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1812 NULL. After this change, only the cofactor parameter can be NULL. It also
1813 does some minimal sanity checks on the passed order.
1818 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1819 An attack is simple, if the first CMS_recipientInfo is valid but the
1820 second CMS_recipientInfo is chosen ciphertext. If the second
1821 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1822 encryption key will be replaced by garbage, and the message cannot be
1823 decoded, but if the RSA decryption fails, the correct encryption key is
1824 used and the recipient will not notice the attack.
1825 As a work around for this potential attack the length of the decrypted
1826 key must be equal to the cipher default key length, in case the
1827 certifiate is not given and all recipientInfo are tried out.
1828 The old behaviour can be re-enabled in the CMS code by setting the
1829 CMS_DEBUG_DECRYPT flag.
1834 * Use Windows installation paths in the mingw builds
1836 Mingw isn't a POSIX environment per se, which means that Windows
1837 paths should be used for installation.
1842 ### Changes between 1.1.0j and 1.1.0k [28 May 2019] ###
1844 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1845 This changes the size when using the genpkey app when no size is given. It
1846 fixes an omission in earlier changes that changed all RSA, DSA and DH
1847 generation apps to use 2048 bits by default.
1851 * Prevent over long nonces in ChaCha20-Poly1305.
1853 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1854 for every encryption operation. RFC 7539 specifies that the nonce value
1855 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1856 and front pads the nonce with 0 bytes if it is less than 12
1857 bytes. However it also incorrectly allows a nonce to be set of up to 16
1858 bytes. In this case only the last 12 bytes are significant and any
1859 additional leading bytes are ignored.
1861 It is a requirement of using this cipher that nonce values are
1862 unique. Messages encrypted using a reused nonce value are susceptible to
1863 serious confidentiality and integrity attacks. If an application changes
1864 the default nonce length to be longer than 12 bytes and then makes a
1865 change to the leading bytes of the nonce expecting the new value to be a
1866 new unique nonce then such an application could inadvertently encrypt
1867 messages with a reused nonce.
1869 Additionally the ignored bytes in a long nonce are not covered by the
1870 integrity guarantee of this cipher. Any application that relies on the
1871 integrity of these ignored leading bytes of a long nonce may be further
1872 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1873 is safe because no such use sets such a long nonce value. However user
1874 applications that use this cipher directly and set a non-default nonce
1875 length to be longer than 12 bytes may be vulnerable.
1877 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1883 * Added SCA hardening for modular field inversion in EC_GROUP through
1884 a new dedicated field_inv() pointer in EC_METHOD.
1885 This also addresses a leakage affecting conversions from projective
1886 to affine coordinates.
1888 *Billy Bob Brumley, Nicola Tuveri*
1890 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1891 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1895 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1899 * Remove the 'dist' target and add a tarball building script. The
1900 'dist' target has fallen out of use, and it shouldn't be
1901 necessary to configure just to create a source distribution.
1905 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018] ###
1907 * Timing vulnerability in DSA signature generation
1909 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1910 timing side channel attack. An attacker could use variations in the signing
1911 algorithm to recover the private key.
1913 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1918 * Timing vulnerability in ECDSA signature generation
1920 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1921 timing side channel attack. An attacker could use variations in the signing
1922 algorithm to recover the private key.
1924 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1929 * Add coordinate blinding for EC_POINT and implement projective
1930 coordinate blinding for generic prime curves as a countermeasure to
1931 chosen point SCA attacks.
1933 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1935 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018] ###
1937 * Client DoS due to large DH parameter
1939 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1940 malicious server can send a very large prime value to the client. This will
1941 cause the client to spend an unreasonably long period of time generating a
1942 key for this prime resulting in a hang until the client has finished. This
1943 could be exploited in a Denial Of Service attack.
1945 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1950 * Cache timing vulnerability in RSA Key Generation
1952 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1953 a cache timing side channel attack. An attacker with sufficient access to
1954 mount cache timing attacks during the RSA key generation process could
1955 recover the private key.
1957 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1958 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1963 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1964 parameter is no longer accepted, as it leads to a corrupt table. NULL
1965 pem_str is reserved for alias entries only.
1969 * Revert blinding in ECDSA sign and instead make problematic addition
1970 length-invariant. Switch even to fixed-length Montgomery multiplication.
1974 * Change generating and checking of primes so that the error rate of not
1975 being prime depends on the intended use based on the size of the input.
1976 For larger primes this will result in more rounds of Miller-Rabin.
1977 The maximal error rate for primes with more than 1080 bits is lowered
1980 *Kurt Roeckx, Annie Yousar*
1982 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1986 * Add blinding to ECDSA and DSA signatures to protect against side channel
1987 attacks discovered by Keegan Ryan (NCC Group).
1991 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1992 now allow empty (zero character) pass phrases.
1996 * Certificate time validation (X509_cmp_time) enforces stricter
1997 compliance with RFC 5280. Fractional seconds and timezone offsets
1998 are no longer allowed.
2002 * Fixed a text canonicalisation bug in CMS
2004 Where a CMS detached signature is used with text content the text goes
2005 through a canonicalisation process first prior to signing or verifying a
2006 signature. This process strips trailing space at the end of lines, converts
2007 line terminators to CRLF and removes additional trailing line terminators
2008 at the end of a file. A bug in the canonicalisation process meant that
2009 some characters, such as form-feed, were incorrectly treated as whitespace
2010 and removed. This is contrary to the specification (RFC5485). This fix
2011 could mean that detached text data signed with an earlier version of
2012 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2013 signed with a fixed OpenSSL may fail to verify with an earlier version of
2014 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2015 and use the "-binary" flag (for the "cms" command line application) or set
2016 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2020 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018] ###
2022 * Constructed ASN.1 types with a recursive definition could exceed the stack
2024 Constructed ASN.1 types with a recursive definition (such as can be found
2025 in PKCS7) could eventually exceed the stack given malicious input with
2026 excessive recursion. This could result in a Denial Of Service attack. There
2027 are no such structures used within SSL/TLS that come from untrusted sources
2028 so this is considered safe.
2030 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2036 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2038 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2039 effectively reduced to only comparing the least significant bit of each
2040 byte. This allows an attacker to forge messages that would be considered as
2041 authenticated in an amount of tries lower than that guaranteed by the
2042 security claims of the scheme. The module can only be compiled by the
2043 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2045 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2051 * Add a build target 'build_all_generated', to build all generated files
2052 and only that. This can be used to prepare everything that requires
2053 things like perl for a system that lacks perl and then move everything
2054 to that system and do the rest of the build there.
2058 * Backport SSL_OP_NO_RENGOTIATION
2060 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2061 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2062 changes this is no longer possible in 1.1.0. Therefore the new
2063 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2064 1.1.0 to provide equivalent functionality.
2066 Note that if an application built against 1.1.0h headers (or above) is run
2067 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2068 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2072 * Removed the OS390-Unix config target. It relied on a script that doesn't
2077 * rsaz_1024_mul_avx2 overflow bug on x86_64
2079 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2080 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2081 Analysis suggests that attacks against RSA and DSA as a result of this
2082 defect would be very difficult to perform and are not believed likely.
2083 Attacks against DH1024 are considered just feasible, because most of the
2084 work necessary to deduce information about a private key may be performed
2085 offline. The amount of resources required for such an attack would be
2086 significant. However, for an attack on TLS to be meaningful, the server
2087 would have to share the DH1024 private key among multiple clients, which is
2088 no longer an option since CVE-2016-0701.
2090 This only affects processors that support the AVX2 but not ADX extensions
2091 like Intel Haswell (4th generation).
2093 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2094 was originally found via the OSS-Fuzz project.
2099 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017] ###
2101 * bn_sqrx8x_internal carry bug on x86_64
2103 There is a carry propagating bug in the x86_64 Montgomery squaring
2104 procedure. No EC algorithms are affected. Analysis suggests that attacks
2105 against RSA and DSA as a result of this defect would be very difficult to
2106 perform and are not believed likely. Attacks against DH are considered just
2107 feasible (although very difficult) because most of the work necessary to
2108 deduce information about a private key may be performed offline. The amount
2109 of resources required for such an attack would be very significant and
2110 likely only accessible to a limited number of attackers. An attacker would
2111 additionally need online access to an unpatched system using the target
2112 private key in a scenario with persistent DH parameters and a private
2113 key that is shared between multiple clients.
2115 This only affects processors that support the BMI1, BMI2 and ADX extensions
2116 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2118 This issue was reported to OpenSSL by the OSS-Fuzz project.
2123 * Malformed X.509 IPAddressFamily could cause OOB read
2125 If an X.509 certificate has a malformed IPAddressFamily extension,
2126 OpenSSL could do a one-byte buffer overread. The most likely result
2127 would be an erroneous display of the certificate in text format.
2129 This issue was reported to OpenSSL by the OSS-Fuzz project.
2134 ### Changes between 1.1.0e and 1.1.0f [25 May 2017] ###
2136 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2137 platform rather than 'mingw'.
2141 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2142 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2143 which is the minimum version we support.
2147 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017] ###
2149 * Encrypt-Then-Mac renegotiation crash
2151 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2152 negotiated where it was not in the original handshake (or vice-versa) then
2153 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2154 and servers are affected.
2156 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2161 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017] ###
2163 * Truncated packet could crash via OOB read
2165 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2166 cipher is being used, then a truncated packet can cause that host to
2167 perform an out-of-bounds read, usually resulting in a crash.
2169 This issue was reported to OpenSSL by Robert Święcki of Google.
2174 * Bad (EC)DHE parameters cause a client crash
2176 If a malicious server supplies bad parameters for a DHE or ECDHE key
2177 exchange then this can result in the client attempting to dereference a
2178 NULL pointer leading to a client crash. This could be exploited in a Denial
2181 This issue was reported to OpenSSL by Guido Vranken.
2186 * BN_mod_exp may produce incorrect results on x86_64
2188 There is a carry propagating bug in the x86_64 Montgomery squaring
2189 procedure. No EC algorithms are affected. Analysis suggests that attacks
2190 against RSA and DSA as a result of this defect would be very difficult to
2191 perform and are not believed likely. Attacks against DH are considered just
2192 feasible (although very difficult) because most of the work necessary to
2193 deduce information about a private key may be performed offline. The amount
2194 of resources required for such an attack would be very significant and
2195 likely only accessible to a limited number of attackers. An attacker would
2196 additionally need online access to an unpatched system using the target
2197 private key in a scenario with persistent DH parameters and a private
2198 key that is shared between multiple clients. For example this can occur by
2199 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2200 similar to CVE-2015-3193 but must be treated as a separate problem.
2202 This issue was reported to OpenSSL by the OSS-Fuzz project.
2207 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016] ###
2209 * ChaCha20/Poly1305 heap-buffer-overflow
2211 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
2212 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2213 crash. This issue is not considered to be exploitable beyond a DoS.
2215 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2220 * CMS Null dereference
2222 Applications parsing invalid CMS structures can crash with a NULL pointer
2223 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2224 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2225 structure callback if an attempt is made to free certain invalid encodings.
2226 Only CHOICE structures using a callback which do not handle NULL value are
2229 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2234 * Montgomery multiplication may produce incorrect results
2236 There is a carry propagating bug in the Broadwell-specific Montgomery
2237 multiplication procedure that handles input lengths divisible by, but
2238 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2239 and DH private keys are impossible. This is because the subroutine in
2240 question is not used in operations with the private key itself and an input
2241 of the attacker's direct choice. Otherwise the bug can manifest itself as
2242 transient authentication and key negotiation failures or reproducible
2243 erroneous outcome of public-key operations with specially crafted input.
2244 Among EC algorithms only Brainpool P-512 curves are affected and one
2245 presumably can attack ECDH key negotiation. Impact was not analyzed in
2246 detail, because pre-requisites for attack are considered unlikely. Namely
2247 multiple clients have to choose the curve in question and the server has to
2248 share the private key among them, neither of which is default behaviour.
2249 Even then only clients that chose the curve will be affected.
2251 This issue was publicly reported as transient failures and was not
2252 initially recognized as a security issue. Thanks to Richard Morgan for
2253 providing reproducible case.
2258 * Removed automatic addition of RPATH in shared libraries and executables,
2259 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2263 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016] ###
2265 * Fix Use After Free for large message sizes
2267 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2268 message larger than approx 16k is received then the underlying buffer to
2269 store the incoming message is reallocated and moved. Unfortunately a
2270 dangling pointer to the old location is left which results in an attempt to
2271 write to the previously freed location. This is likely to result in a
2272 crash, however it could potentially lead to execution of arbitrary code.
2274 This issue only affects OpenSSL 1.1.0a.
2276 This issue was reported to OpenSSL by Robert Święcki.
2281 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016] ###
2283 * OCSP Status Request extension unbounded memory growth
2285 A malicious client can send an excessively large OCSP Status Request
2286 extension. If that client continually requests renegotiation, sending a
2287 large OCSP Status Request extension each time, then there will be unbounded
2288 memory growth on the server. This will eventually lead to a Denial Of
2289 Service attack through memory exhaustion. Servers with a default
2290 configuration are vulnerable even if they do not support OCSP. Builds using
2291 the "no-ocsp" build time option are not affected.
2293 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2298 * SSL_peek() hang on empty record
2300 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2301 sends an empty record. This could be exploited by a malicious peer in a
2302 Denial Of Service attack.
2304 This issue was reported to OpenSSL by Alex Gaynor.
2309 * Excessive allocation of memory in tls_get_message_header() and
2310 dtls1_preprocess_fragment()
2312 A (D)TLS message includes 3 bytes for its length in the header for the
2313 message. This would allow for messages up to 16Mb in length. Messages of
2314 this length are excessive and OpenSSL includes a check to ensure that a
2315 peer is sending reasonably sized messages in order to avoid too much memory
2316 being consumed to service a connection. A flaw in the logic of version
2317 1.1.0 means that memory for the message is allocated too early, prior to
2318 the excessive message length check. Due to way memory is allocated in
2319 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2320 to service a connection. This could lead to a Denial of Service through
2321 memory exhaustion. However, the excessive message length check still takes
2322 place, and this would cause the connection to immediately fail. Assuming
2323 that the application calls SSL_free() on the failed connection in a timely
2324 manner then the 21Mb of allocated memory will then be immediately freed
2325 again. Therefore the excessive memory allocation will be transitory in
2326 nature. This then means that there is only a security impact if:
2328 1) The application does not call SSL_free() in a timely manner in the event
2329 that the connection fails
2331 2) The application is working in a constrained environment where there is
2332 very little free memory
2334 3) The attacker initiates multiple connection attempts such that there are
2335 multiple connections in a state where memory has been allocated for the
2336 connection; SSL_free() has not yet been called; and there is insufficient
2337 memory to service the multiple requests.
2339 Except in the instance of (1) above any Denial Of Service is likely to be
2340 transitory because as soon as the connection fails the memory is
2341 subsequently freed again in the SSL_free() call. However there is an
2342 increased risk during this period of application crashes due to the lack of
2343 memory - which would then mean a more serious Denial of Service.
2345 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2346 (CVE-2016-6307 and CVE-2016-6308)
2350 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2351 had to be removed. Primary reason is that vendor assembler can't
2352 assemble our modules with -KPIC flag. As result it, assembly
2353 support, was not even available as option. But its lack means
2354 lack of side-channel resistant code, which is incompatible with
2355 security by todays standards. Fortunately gcc is readily available
2356 prepackaged option, which we firmly point at...
2360 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016] ###
2362 * Windows command-line tool supports UTF-8 opt-in option for arguments
2363 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2364 (to any value) allows Windows user to access PKCS#12 file generated
2365 with Windows CryptoAPI and protected with non-ASCII password, as well
2366 as files generated under UTF-8 locale on Linux also protected with
2371 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2372 have been disabled by default and removed from DEFAULT, just like RC4.
2373 See the RC4 item below to re-enable both.
2377 * The method for finding the storage location for the Windows RAND seed file
2378 has changed. First we check %RANDFILE%. If that is not set then we check
2379 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2380 all else fails we fall back to C:\.
2384 * The EVP_EncryptUpdate() function has had its return type changed from void
2385 to int. A return of 0 indicates and error while a return of 1 indicates
2390 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2391 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2392 off the constant time implementation for RSA, DSA and DH have been made
2393 no-ops and deprecated.
2397 * Windows RAND implementation was simplified to only get entropy by
2398 calling CryptGenRandom(). Various other RAND-related tickets
2401 *Joseph Wylie Yandle, Rich Salz*
2403 * The stack and lhash API's were renamed to start with OPENSSL_SK_
2404 and OPENSSL_LH_, respectively. The old names are available
2405 with API compatibility. They new names are now completely documented.
2409 * Unify TYPE_up_ref(obj) methods signature.
2410 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2411 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2412 int (instead of void) like all others TYPE_up_ref() methods.
2413 So now these methods also check the return value of CRYPTO_atomic_add(),
2414 and the validity of object reference counter.
2416 *fdasilvayy@gmail.com*
2418 * With Windows Visual Studio builds, the .pdb files are installed
2419 alongside the installed libraries and executables. For a static
2420 library installation, ossl_static.pdb is the associate compiler
2421 generated .pdb file to be used when linking programs.
2425 * Remove openssl.spec. Packaging files belong with the packagers.
2429 * Automatic Darwin/OSX configuration has had a refresh, it will now
2430 recognise x86_64 architectures automatically. You can still decide
2431 to build for a different bitness with the environment variable
2432 KERNEL_BITS (can be 32 or 64), for example:
2434 KERNEL_BITS=32 ./config
2438 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2439 256 bit AES and HMAC with SHA256.
2443 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2447 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2451 * To enable users to have their own config files and build file templates,
2452 Configure looks in the directory indicated by the environment variable
2453 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2454 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2455 name and is used as is.
2459 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2460 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2461 X509_CERT_FILE_CTX was removed.
2465 * "shared" builds are now the default. To create only static libraries use
2466 the "no-shared" Configure option.
2470 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2471 All of these option have not worked for some while and are fundamental
2476 * Make various cleanup routines no-ops and mark them as deprecated. Most
2477 global cleanup functions are no longer required because they are handled
2478 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2479 Explicitly de-initing can cause problems (e.g. where a library that uses
2480 OpenSSL de-inits, but an application is still using it). The affected
2481 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2482 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2483 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2484 COMP_zlib_cleanup().
2488 * --strict-warnings no longer enables runtime debugging options
2489 such as REF_DEBUG. Instead, debug options are automatically
2490 enabled with '--debug' builds.
2492 *Andy Polyakov, Emilia Käsper*
2494 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2495 have been moved out of the public header files. New functions for managing
2496 these have been added.
2500 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2501 objects have been moved out of the public header files. New
2502 functions for managing these have been added.
2506 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2507 have been moved out of the public header files. New functions for managing
2508 these have been added.
2512 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2513 moved out of the public header files. New functions for managing these
2518 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2522 * Removed the mk1mf build scripts.
2526 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2527 it is always safe to #include a header now.
2531 * Removed the aged BC-32 config and all its supporting scripts
2535 * Removed support for Ultrix, Netware, and OS/2.
2539 * Add support for HKDF.
2541 *Alessandro Ghedini*
2543 * Add support for blake2b and blake2s
2547 * Added support for "pipelining". Ciphers that have the
2548 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2549 encryptions/decryptions simultaneously. There are currently no built-in
2550 ciphers with this property but the expectation is that engines will be able
2551 to offer it to significantly improve throughput. Support has been extended
2552 into libssl so that multiple records for a single connection can be
2553 processed in one go (for >=TLS 1.1).
2557 * Added the AFALG engine. This is an async capable engine which is able to
2558 offload work to the Linux kernel. In this initial version it only supports
2559 AES128-CBC. The kernel must be version 4.1.0 or greater.
2563 * OpenSSL now uses a new threading API. It is no longer necessary to
2564 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2565 are two supported threading models: pthreads and windows threads. It is
2566 also possible to configure OpenSSL at compile time for "no-threads". The
2567 old threading API should no longer be used. The functions have been
2568 replaced with "no-op" compatibility macros.
2570 *Alessandro Ghedini, Matt Caswell*
2572 * Modify behavior of ALPN to invoke callback after SNI/servername
2573 callback, such that updates to the SSL_CTX affect ALPN.
2577 * Add SSL_CIPHER queries for authentication and key-exchange.
2581 * Changes to the DEFAULT cipherlist:
2582 - Prefer (EC)DHE handshakes over plain RSA.
2583 - Prefer AEAD ciphers over legacy ciphers.
2584 - Prefer ECDSA over RSA when both certificates are available.
2585 - Prefer TLSv1.2 ciphers/PRF.
2586 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2591 * Change the ECC default curve list to be this, in order: x25519,
2592 secp256r1, secp521r1, secp384r1.
2596 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2597 disabled by default. They can be re-enabled using the
2598 enable-weak-ssl-ciphers option to Configure.
2602 * If the server has ALPN configured, but supports no protocols that the
2603 client advertises, send a fatal "no_application_protocol" alert.
2604 This behaviour is SHALL in RFC 7301, though it isn't universally
2605 implemented by other servers.
2609 * Add X25519 support.
2610 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2611 for public and private key encoding using the format documented in
2612 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2613 key generation and key derivation.
2615 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2620 * Deprecate SRP_VBASE_get_by_user.
2621 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2622 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2623 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2624 seed, even if the seed is configured.
2626 Users should use SRP_VBASE_get1_by_user instead. Note that in
2627 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2628 also that even though configuring the SRP seed attempts to hide
2629 invalid usernames by continuing the handshake with fake
2630 credentials, this behaviour is not constant time and no strong
2631 guarantees are made that the handshake is indistinguishable from
2632 that of a valid user.
2636 * Configuration change; it's now possible to build dynamic engines
2637 without having to build shared libraries and vice versa. This
2638 only applies to the engines in engines/, those in crypto/engine/
2639 will always be built into libcrypto (i.e. "static").
2641 Building dynamic engines is enabled by default; to disable, use
2642 the configuration option "disable-dynamic-engine".
2644 The only requirements for building dynamic engines are the
2645 presence of the DSO module and building with position independent
2646 code, so they will also automatically be disabled if configuring
2647 with "disable-dso" or "disable-pic".
2649 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2650 are also taken away from openssl/opensslconf.h, as they are
2655 * Configuration change; if there is a known flag to compile
2656 position independent code, it will always be applied on the
2657 libcrypto and libssl object files, and never on the application
2658 object files. This means other libraries that use routines from
2659 libcrypto / libssl can be made into shared libraries regardless
2660 of how OpenSSL was configured.
2662 If this isn't desirable, the configuration options "disable-pic"
2663 or "no-pic" can be used to disable the use of PIC. This will
2664 also disable building shared libraries and dynamic engines.
2668 * Removed JPAKE code. It was experimental and has no wide use.
2672 * The INSTALL_PREFIX Makefile variable has been renamed to
2673 DESTDIR. That makes for less confusion on what this variable
2674 is for. Also, the configuration option --install_prefix is
2679 * Heartbeat for TLS has been removed and is disabled by default
2680 for DTLS; configure with enable-heartbeats. Code that uses the
2681 old #define's might need to be updated.
2683 *Emilia Käsper, Rich Salz*
2685 * Rename REF_CHECK to REF_DEBUG.
2689 * New "unified" build system
2691 The "unified" build system is aimed to be a common system for all
2692 platforms we support. With it comes new support for VMS.
2694 This system builds supports building in a different directory tree
2695 than the source tree. It produces one Makefile (for unix family
2696 or lookalikes), or one descrip.mms (for VMS).
2698 The source of information to make the Makefile / descrip.mms is
2699 small files called 'build.info', holding the necessary
2700 information for each directory with source to compile, and a
2701 template in Configurations, like unix-Makefile.tmpl or
2704 With this change, the library names were also renamed on Windows
2705 and on VMS. They now have names that are closer to the standard
2706 on Unix, and include the major version number, and in certain
2707 cases, the architecture they are built for. See "Notes on shared
2708 libraries" in INSTALL.
2710 We rely heavily on the perl module Text::Template.
2714 * Added support for auto-initialisation and de-initialisation of the library.
2715 OpenSSL no longer requires explicit init or deinit routines to be called,
2716 except in certain circumstances. See the OPENSSL_init_crypto() and
2717 OPENSSL_init_ssl() man pages for further information.
2721 * The arguments to the DTLSv1_listen function have changed. Specifically the
2722 "peer" argument is now expected to be a BIO_ADDR object.
2724 * Rewrite of BIO networking library. The BIO library lacked consistent
2725 support of IPv6, and adding it required some more extensive
2726 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2727 which hold all types of addresses and chains of address information.
2728 It also introduces a new API, with functions like BIO_socket,
2729 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2730 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2731 have been adapted accordingly.
2735 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2740 * CRIME protection: disable compression by default, even if OpenSSL is
2741 compiled with zlib enabled. Applications can still enable compression
2742 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2743 using the SSL_CONF library to configure compression.
2747 * The signature of the session callback configured with
2748 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2749 was explicitly marked as 'const unsigned char*' instead of
2754 * Always DPURIFY. Remove the use of uninitialized memory in the
2755 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2759 * Removed many obsolete configuration items, including
2760 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2761 MD2_CHAR, MD2_INT, MD2_LONG
2763 IDEA_SHORT, IDEA_LONG
2764 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2766 *Rich Salz, with advice from Andy Polyakov*
2768 * Many BN internals have been moved to an internal header file.
2770 *Rich Salz with help from Andy Polyakov*
2772 * Configuration and writing out the results from it has changed.
2773 Files such as Makefile include/openssl/opensslconf.h and are now
2774 produced through general templates, such as Makefile.in and
2775 crypto/opensslconf.h.in and some help from the perl module
2778 Also, the center of configuration information is no longer
2779 Makefile. Instead, Configure produces a perl module in
2780 configdata.pm which holds most of the config data (in the hash
2781 table %config), the target data that comes from the target
2782 configuration in one of the Configurations/*.conf files (in
2787 * To clarify their intended purposes, the Configure options
2788 --prefix and --openssldir change their semantics, and become more
2789 straightforward and less interdependent.
2791 --prefix shall be used exclusively to give the location INSTALLTOP
2792 where programs, scripts, libraries, include files and manuals are
2793 going to be installed. The default is now /usr/local.
2795 --openssldir shall be used exclusively to give the default
2796 location OPENSSLDIR where certificates, private keys, CRLs are
2797 managed. This is also where the default openssl.cnf gets
2799 If the directory given with this option is a relative path, the
2800 values of both the --prefix value and the --openssldir value will
2801 be combined to become OPENSSLDIR.
2802 The default for --openssldir is INSTALLTOP/ssl.
2804 Anyone who uses --openssldir to specify where OpenSSL is to be
2805 installed MUST change to use --prefix instead.
2809 * The GOST engine was out of date and therefore it has been removed. An up
2810 to date GOST engine is now being maintained in an external repository.
2811 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2812 support for GOST ciphersuites (these are only activated if a GOST engine
2817 * EGD is no longer supported by default; use enable-egd when
2820 *Ben Kaduk and Rich Salz*
2822 * The distribution now has Makefile.in files, which are used to
2823 create Makefile's when Configure is run. *Configure must be run
2824 before trying to build now.*
2828 * The return value for SSL_CIPHER_description() for error conditions
2833 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2835 Obtaining and performing DNSSEC validation of TLSA records is
2836 the application's responsibility. The application provides
2837 the TLSA records of its choice to OpenSSL, and these are then
2838 used to authenticate the peer.
2840 The TLSA records need not even come from DNS. They can, for
2841 example, be used to implement local end-entity certificate or
2842 trust-anchor "pinning", where the "pin" data takes the form
2843 of TLSA records, which can augment or replace verification
2844 based on the usual WebPKI public certification authorities.
2848 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2849 continues to support deprecated interfaces in default builds.
2850 However, applications are strongly advised to compile their
2851 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2852 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2853 or the 1.1.0 releases.
2855 In environments in which all applications have been ported to
2856 not use any deprecated interfaces OpenSSL's Configure script
2857 should be used with the --api=1.1.0 option to entirely remove
2858 support for the deprecated features from the library and
2859 unconditionally disable them in the installed headers.
2860 Essentially the same effect can be achieved with the "no-deprecated"
2861 argument to Configure, except that this will always restrict
2862 the build to just the latest API, rather than a fixed API
2865 As applications are ported to future revisions of the API,
2866 they should update their compile-time OPENSSL_API_COMPAT define
2867 accordingly, but in most cases should be able to continue to
2868 compile with later releases.
2870 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2871 0x10000000L and 0x00908000L, respectively. However those
2872 versions did not support the OPENSSL_API_COMPAT feature, and
2873 so applications are not typically tested for explicit support
2874 of just the undeprecated features of either release.
2878 * Add support for setting the minimum and maximum supported protocol.
2879 It can bet set via the SSL_set_min_proto_version() and
2880 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2881 MaxProtocol. It's recommended to use the new APIs to disable
2882 protocols instead of disabling individual protocols using
2883 SSL_set_options() or SSL_CONF's Protocol. This change also
2884 removes support for disabling TLS 1.2 in the OpenSSL TLS
2885 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2889 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2893 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2894 and integrates ECDSA and ECDH functionality into EC. Implementations can
2895 now redirect key generation and no longer need to convert to or from
2898 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2899 include the ec.h header file instead.
2903 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2904 ciphers who are no longer supported and drops support the ephemeral RSA key
2905 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2909 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2910 opaque. For HMAC_CTX, the following constructors and destructors
2913 HMAC_CTX *HMAC_CTX_new(void);
2914 void HMAC_CTX_free(HMAC_CTX *ctx);
2916 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2917 destroy such methods has been added. See EVP_MD_meth_new(3) and
2918 EVP_CIPHER_meth_new(3) for documentation.
2921 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2922 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2923 EVP_MD_CTX_reset() should be called instead to reinitialise
2924 an already created structure.
2925 2) For consistency with the majority of our object creators and
2926 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2927 EVP_MD_CTX_(new|free). The old names are retained as macros
2928 for deprecated builds.
2932 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2933 cryptographic operations to be performed asynchronously as long as an
2934 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2935 further details. Libssl has also had this capability integrated with the
2936 introduction of the new mode SSL_MODE_ASYNC and associated error
2937 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2938 pages. This work was developed in partnership with Intel Corp.
2942 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2943 always enabled now. If you want to disable the support you should
2944 exclude it using the list of supported ciphers. This also means that the
2945 "-no_ecdhe" option has been removed from s_server.
2949 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2950 SSL_{CTX_}set1_curves() which can set a list.
2954 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2955 curve you want to support using SSL_{CTX_}set1_curves().
2959 * State machine rewrite. The state machine code has been significantly
2960 refactored in order to remove much duplication of code and solve issues
2961 with the old code (see ssl/statem/README for further details). This change
2962 does have some associated API changes. Notably the SSL_state() function
2963 has been removed and replaced by SSL_get_state which now returns an
2964 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2965 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2970 * All instances of the string "ssleay" in the public API were replaced
2971 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2972 Some error codes related to internal RSA_eay API's were renamed.
2976 * The demo files in crypto/threads were moved to demo/threads.
2980 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2983 *Matt Caswell, Rich Salz*
2985 * New ASN.1 embed macro.
2987 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2988 structure is not allocated: it is part of the parent. That is instead of
2996 This reduces memory fragmentation and make it impossible to accidentally
2997 set a mandatory field to NULL.
2999 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3000 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3001 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3006 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3010 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3011 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3012 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3013 DES and RC4 ciphersuites.
3017 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3018 This changes the decoding behaviour for some invalid messages,
3019 though the change is mostly in the more lenient direction, and
3020 legacy behaviour is preserved as much as possible.
3024 * Fix no-stdio build.
3025 * David Woodhouse <David.Woodhouse@intel.com> and also
3026 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3028 * New testing framework
3029 The testing framework has been largely rewritten and is now using
3030 perl and the perl modules Test::Harness and an extended variant of
3031 Test::More called OpenSSL::Test to do its work. All test scripts in
3032 test/ have been rewritten into test recipes, and all direct calls to
3033 executables in test/Makefile have become individual recipes using the
3034 simplified testing OpenSSL::Test::Simple.
3036 For documentation on our testing modules, do:
3038 perldoc test/testlib/OpenSSL/Test/Simple.pm
3039 perldoc test/testlib/OpenSSL/Test.pm
3043 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3044 are used; the latter aborts on memory leaks (usually checked on exit).
3045 Some undocumented "set malloc, etc., hooks" functions were removed
3046 and others were changed. All are now documented.
3050 * In DSA_generate_parameters_ex, if the provided seed is too short,
3053 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3055 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3056 from RFC4279, RFC4785, RFC5487, RFC5489.
3058 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3059 original RSA_PSK patch.
3063 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3064 era flag was never set throughout the codebase (only read). Also removed
3065 SSL3_FLAGS_POP_BUFFER which was only used if
3066 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3070 * Changed the default name options in the "ca", "crl", "req" and "x509"
3071 to be "oneline" instead of "compat".
3075 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3076 not aware of clients that still exhibit this bug, and the workaround
3077 hasn't been working properly for a while.
3081 * The return type of BIO_number_read() and BIO_number_written() as well as
3082 the corresponding num_read and num_write members in the BIO structure has
3083 changed from unsigned long to uint64_t. On platforms where an unsigned
3084 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3089 * Given the pervasive nature of TLS extensions it is inadvisable to run
3090 OpenSSL without support for them. It also means that maintaining
3091 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3092 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3096 * Removed support for the two export grade static DH ciphersuites
3097 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3098 were newly added (along with a number of other static DH ciphersuites) to
3099 1.0.2. However the two export ones have *never* worked since they were
3100 introduced. It seems strange in any case to be adding new export
3101 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3105 * Version negotiation has been rewritten. In particular SSLv23_method(),
3106 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3107 and turned into macros which simply call the new preferred function names
3108 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3109 should use the new names instead. Also as part of this change the ssl23.h
3110 header file has been removed.
3114 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3115 code and the associated standard is no longer considered fit-for-purpose.
3119 * RT2547 was closed. When generating a private key, try to make the
3120 output file readable only by the owner. This behavior change might
3121 be noticeable when interacting with other software.
3123 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3128 * Added HTTP GET support to the ocsp command.
3132 * Changed default digest for the dgst and enc commands from MD5 to
3137 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3141 * Added support for TLS extended master secret from
3142 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3143 initial patch which was a great help during development.
3147 * All libssl internal structures have been removed from the public header
3148 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3149 now redundant). Users should not attempt to access internal structures
3150 directly. Instead they should use the provided API functions.
3154 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3155 Access to deprecated functions can be re-enabled by running config with
3156 "enable-deprecated". In addition applications wishing to use deprecated
3157 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3158 will, by default, disable some transitive includes that previously existed
3159 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3163 * Added support for OCB mode. OpenSSL has been granted a patent license
3164 compatible with the OpenSSL license for use of OCB. Details are available
3165 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
3166 for OCB can be removed by calling config with no-ocb.
3170 * SSLv2 support has been removed. It still supports receiving a SSLv2
3171 compatible client hello.
3175 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3176 done while fixing the error code for the key-too-small case.
3178 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3180 * CA.sh has been removed; use CA.pl instead.
3184 * Removed old DES API.
3188 * Remove various unsupported platforms:
3194 Sinix/ReliantUNIX RM400
3199 16-bit platforms such as WIN16
3203 * Clean up OPENSSL_NO_xxx #define's
3204 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3205 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3206 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3207 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3208 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3209 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3210 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3211 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3212 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3213 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3217 * Cleaned up dead code
3218 Remove all but one '#ifdef undef' which is to be looked at.
3222 * Clean up calling of xxx_free routines.
3223 Just like free(), fix most of the xxx_free routines to accept
3224 NULL. Remove the non-null checks from callers. Save much code.
3228 * Add secure heap for storage of private keys (when possible).
3229 Add BIO_s_secmem(), CBIGNUM, etc.
3230 Contributed by Akamai Technologies under our Corporate CLA.
3234 * Experimental support for a new, fast, unbiased prime candidate generator,
3235 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3237 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3239 * New output format NSS in the sess_id command line tool. This allows
3240 exporting the session id and the master key in NSS keylog format.
3242 *Martin Kaiser <martin@kaiser.cx>*
3244 * Harmonize version and its documentation. -f flag is used to display
3247 *mancha <mancha1@zoho.com>*
3249 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3250 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3252 *mancha <mancha1@zoho.com>*
3254 * Fix some double frees. These are not thought to be exploitable.
3256 *mancha <mancha1@zoho.com>*
3258 * A missing bounds check in the handling of the TLS heartbeat extension
3259 can be used to reveal up to 64k of memory to a connected client or
3262 Thanks for Neel Mehta of Google Security for discovering this bug and to
3263 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3264 preparing the fix [CVE-2014-0160][]
3266 *Adam Langley, Bodo Moeller*
3268 * Fix for the attack described in the paper "Recovering OpenSSL
3269 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3270 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3271 http://eprint.iacr.org/2014/140
3273 Thanks to Yuval Yarom and Naomi Benger for discovering this
3274 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3276 *Yuval Yarom and Naomi Benger*
3278 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3279 this fixes a limitation in previous versions of OpenSSL.
3283 * Experimental encrypt-then-mac support.
3285 Experimental support for encrypt then mac from
3286 draft-gutmann-tls-encrypt-then-mac-02.txt
3288 To enable it set the appropriate extension number (0x42 for the test
3289 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3291 For non-compliant peers (i.e. just about everything) this should have no
3294 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3299 * Add EVP support for key wrapping algorithms, to avoid problems with
3300 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3301 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3302 algorithms and include tests cases.
3306 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3311 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3312 MGF1 digest and OAEP label.
3316 * Make openssl verify return errors.
3318 *Chris Palmer <palmer@google.com> and Ben Laurie*
3320 * New function ASN1_TIME_diff to calculate the difference between two
3321 ASN1_TIME structures or one structure and the current time.
3325 * Update fips_test_suite to support multiple command line options. New
3326 test to induce all self test errors in sequence and check expected
3331 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3332 sign or verify all in one operation.
3336 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3337 test programs and fips_test_suite. Includes functionality to parse
3338 the minimal script output of fipsalgest.pl directly.
3342 * Add authorisation parameter to FIPS_module_mode_set().
3346 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3350 * Use separate DRBG fields for internal and external flags. New function
3351 FIPS_drbg_health_check() to perform on demand health checking. Add
3352 generation tests to fips_test_suite with reduced health check interval to
3353 demonstrate periodic health checking. Add "nodh" option to
3354 fips_test_suite to skip very slow DH test.
3358 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3363 * More extensive health check for DRBG checking many more failure modes.
3364 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3365 combination: call this in fips_test_suite.
3369 * Add support for canonical generation of DSA parameter 'g'. See
3372 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3373 POST to handle HMAC cases.
3377 * Add functions FIPS_module_version() and FIPS_module_version_text()
3378 to return numerical and string versions of the FIPS module number.
3382 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3383 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3384 outside the validated module in the FIPS capable OpenSSL.
3388 * Minor change to DRBG entropy callback semantics. In some cases
3389 there is no multiple of the block length between min_len and
3390 max_len. Allow the callback to return more than max_len bytes
3391 of entropy but discard any extra: it is the callback's responsibility
3392 to ensure that the extra data discarded does not impact the
3393 requested amount of entropy.
3397 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3398 information in FIPS186-3, SP800-57 and SP800-131A.
3402 * CCM support via EVP. Interface is very similar to GCM case except we
3403 must supply all data in one chunk (i.e. no update, final) and the
3404 message length must be supplied if AAD is used. Add algorithm test
3409 * Initial version of POST overhaul. Add POST callback to allow the status
3410 of POST to be monitored and/or failures induced. Modify fips_test_suite
3411 to use callback. Always run all selftests even if one fails.
3415 * XTS support including algorithm test driver in the fips_gcmtest program.
3416 Note: this does increase the maximum key length from 32 to 64 bytes but
3417 there should be no binary compatibility issues as existing applications
3418 will never use XTS mode.
3422 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3423 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3424 performs algorithm blocking for unapproved PRNG types. Also do not
3425 set PRNG type in FIPS_mode_set(): leave this to the application.
3426 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3427 the standard OpenSSL PRNG: set additional data to a date time vector.
3431 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3432 This shouldn't present any incompatibility problems because applications
3433 shouldn't be using these directly and any that are will need to rethink
3434 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3438 * Extensive self tests and health checking required by SP800-90 DRBG.
3439 Remove strength parameter from FIPS_drbg_instantiate and always
3440 instantiate at maximum supported strength.
3444 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3448 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3452 * New function DH_compute_key_padded() to compute a DH key and pad with
3453 leading zeroes if needed: this complies with SP800-56A et al.
3457 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3458 anything, incomplete, subject to change and largely untested at present.
3462 * Modify fipscanisteronly build option to only build the necessary object
3463 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3467 * Add experimental option FIPSSYMS to give all symbols in
3468 fipscanister.o and FIPS or fips prefix. This will avoid
3469 conflicts with future versions of OpenSSL. Add perl script
3470 util/fipsas.pl to preprocess assembly language source files
3471 and rename any affected symbols.
3475 * Add selftest checks and algorithm block of non-fips algorithms in
3476 FIPS mode. Remove DES2 from selftests.
3480 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3481 return internal method without any ENGINE dependencies. Add new
3482 tiny fips sign and verify functions.
3486 * New build option no-ec2m to disable characteristic 2 code.
3490 * New build option "fipscanisteronly". This only builds fipscanister.o
3491 and (currently) associated fips utilities. Uses the file Makefile.fips
3492 instead of Makefile.org as the prototype.
3496 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3497 Update fips_gcmtest to use IV generator.
3501 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3502 setting output buffer to NULL. The *Final function must be
3503 called although it will not retrieve any additional data. The tag
3504 can be set or retrieved with a ctrl. The IV length is by default 12
3505 bytes (96 bits) but can be set to an alternative value. If the IV
3506 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3511 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3512 underlying do_cipher function handles all cipher semantics itself
3513 including padding and finalisation. This is useful if (for example)
3514 an ENGINE cipher handles block padding itself. The behaviour of
3515 do_cipher is subtly changed if this flag is set: the return value
3516 is the number of characters written to the output buffer (zero is
3517 no longer an error code) or a negative error code. Also if the
3518 input buffer is NULL and length 0 finalisation should be performed.
3522 * If a candidate issuer certificate is already part of the constructed
3523 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3527 * Improve forward-security support: add functions
3529 void SSL_CTX_set_not_resumable_session_callback(
3530 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3531 void SSL_set_not_resumable_session_callback(
3532 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3534 for use by SSL/TLS servers; the callback function will be called whenever a
3535 new session is created, and gets to decide whether the session may be
3536 cached to make it resumable (return 0) or not (return 1). (As by the
3537 SSL/TLS protocol specifications, the session_id sent by the server will be
3538 empty to indicate that the session is not resumable; also, the server will
3539 not generate RFC 4507 (RFC 5077) session tickets.)
3541 A simple reasonable callback implementation is to return is_forward_secure.
3542 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3543 by the SSL/TLS server library, indicating whether it can provide forward
3546 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3548 * New -verify_name option in command line utilities to set verification
3553 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3554 Add CMAC pkey methods.
3558 * Experimental renegotiation in s_server -www mode. If the client
3559 browses /reneg connection is renegotiated. If /renegcert it is
3560 renegotiated requesting a certificate.
3564 * Add an "external" session cache for debugging purposes to s_server. This
3565 should help trace issues which normally are only apparent in deployed
3566 multi-process servers.
3570 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3571 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3572 BIO_set_cipher() and some obscure PEM functions were changed so they
3573 can now return an error. The RAND changes required a change to the
3574 RAND_METHOD structure.
3578 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3579 a gcc attribute to warn if the result of a function is ignored. This
3580 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3581 whose return value is often ignored.
3585 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3586 These allow SCTs (signed certificate timestamps) to be requested and
3587 validated when establishing a connection.
3589 *Rob Percival <robpercival@google.com>*
3594 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019] ###
3596 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3597 used even when parsing explicit parameters, when loading a serialized key
3598 or calling `EC_GROUP_new_from_ecpkparameters()`/
3599 `EC_GROUP_new_from_ecparameters()`.
3600 This prevents bypass of security hardening and performance gains,
3601 especially for curves with specialized EC_METHODs.
3602 By default, if a key encoded with explicit parameters is loaded and later
3603 serialized, the output is still encoded with explicit parameters, even if
3604 internally a "named" EC_GROUP is used for computation.
3608 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3609 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3610 NULL. After this change, only the cofactor parameter can be NULL. It also
3611 does some minimal sanity checks on the passed order.
3616 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3617 An attack is simple, if the first CMS_recipientInfo is valid but the
3618 second CMS_recipientInfo is chosen ciphertext. If the second
3619 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3620 encryption key will be replaced by garbage, and the message cannot be
3621 decoded, but if the RSA decryption fails, the correct encryption key is
3622 used and the recipient will not notice the attack.
3623 As a work around for this potential attack the length of the decrypted
3624 key must be equal to the cipher default key length, in case the
3625 certifiate is not given and all recipientInfo are tried out.
3626 The old behaviour can be re-enabled in the CMS code by setting the
3627 CMS_DEBUG_DECRYPT flag.
3632 * Document issue with installation paths in diverse Windows builds
3634 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3635 binaries and run-time config file.
3640 ### Changes between 1.0.2r and 1.0.2s [28 May 2019] ###
3642 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3643 This changes the size when using the genpkey app when no size is given. It
3644 fixes an omission in earlier changes that changed all RSA, DSA and DH
3645 generation apps to use 2048 bits by default.
3649 * Add FIPS support for Android Arm 64-bit
3651 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3652 Module in Version 2.0.10. For some reason, the corresponding target
3653 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3654 built with FIPS support on Android Arm 64-bit. This omission has been
3657 *Matthias St. Pierre*
3659 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019] ###
3661 * 0-byte record padding oracle
3663 If an application encounters a fatal protocol error and then calls
3664 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3665 then OpenSSL can respond differently to the calling application if a 0 byte
3666 record is received with invalid padding compared to if a 0 byte record is
3667 received with an invalid MAC. If the application then behaves differently
3668 based on that in a way that is detectable to the remote peer, then this
3669 amounts to a padding oracle that could be used to decrypt data.
3671 In order for this to be exploitable "non-stitched" ciphersuites must be in
3672 use. Stitched ciphersuites are optimised implementations of certain
3673 commonly used ciphersuites. Also the application must call SSL_shutdown()
3674 twice even if a protocol error has occurred (applications should not do
3675 this but some do anyway).
3677 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3678 Aviram, with additional investigation by Steven Collison and Andrew
3679 Hourselt. It was reported to OpenSSL on 10th December 2018.
3684 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3688 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018] ###
3690 * Microarchitecture timing vulnerability in ECC scalar multiplication
3692 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3693 shown to be vulnerable to a microarchitecture timing side channel attack.
3694 An attacker with sufficient access to mount local timing attacks during
3695 ECDSA signature generation could recover the private key.
3697 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3698 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3704 * Timing vulnerability in DSA signature generation
3706 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3707 timing side channel attack. An attacker could use variations in the signing
3708 algorithm to recover the private key.
3710 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3715 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3716 Module, accidentally introduced while backporting security fixes from the
3717 development branch and hindering the use of ECC in FIPS mode.
3721 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018] ###
3723 * Client DoS due to large DH parameter
3725 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3726 malicious server can send a very large prime value to the client. This will
3727 cause the client to spend an unreasonably long period of time generating a
3728 key for this prime resulting in a hang until the client has finished. This
3729 could be exploited in a Denial Of Service attack.
3731 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3736 * Cache timing vulnerability in RSA Key Generation
3738 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3739 a cache timing side channel attack. An attacker with sufficient access to
3740 mount cache timing attacks during the RSA key generation process could
3741 recover the private key.
3743 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3744 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3749 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3750 parameter is no longer accepted, as it leads to a corrupt table. NULL
3751 pem_str is reserved for alias entries only.
3755 * Revert blinding in ECDSA sign and instead make problematic addition
3756 length-invariant. Switch even to fixed-length Montgomery multiplication.
3760 * Change generating and checking of primes so that the error rate of not
3761 being prime depends on the intended use based on the size of the input.
3762 For larger primes this will result in more rounds of Miller-Rabin.
3763 The maximal error rate for primes with more than 1080 bits is lowered
3766 *Kurt Roeckx, Annie Yousar*
3768 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3772 * Add blinding to ECDSA and DSA signatures to protect against side channel
3773 attacks discovered by Keegan Ryan (NCC Group).
3777 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3778 now allow empty (zero character) pass phrases.
3782 * Certificate time validation (X509_cmp_time) enforces stricter
3783 compliance with RFC 5280. Fractional seconds and timezone offsets
3784 are no longer allowed.
3788 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018] ###
3790 * Constructed ASN.1 types with a recursive definition could exceed the stack
3792 Constructed ASN.1 types with a recursive definition (such as can be found
3793 in PKCS7) could eventually exceed the stack given malicious input with
3794 excessive recursion. This could result in a Denial Of Service attack. There
3795 are no such structures used within SSL/TLS that come from untrusted sources
3796 so this is considered safe.
3798 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3804 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017] ###
3806 * Read/write after SSL object in error state
3808 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3809 mechanism. The intent was that if a fatal error occurred during a handshake
3810 then OpenSSL would move into the error state and would immediately fail if
3811 you attempted to continue the handshake. This works as designed for the
3812 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3813 SSL_connect()), however due to a bug it does not work correctly if
3814 SSL_read() or SSL_write() is called directly. In that scenario, if the
3815 handshake fails then a fatal error will be returned in the initial function
3816 call. If SSL_read()/SSL_write() is subsequently called by the application
3817 for the same SSL object then it will succeed and the data is passed without
3818 being decrypted/encrypted directly from the SSL/TLS record layer.
3820 In order to exploit this issue an application bug would have to be present
3821 that resulted in a call to SSL_read()/SSL_write() being issued after having
3822 already received a fatal error.
3824 This issue was reported to OpenSSL by David Benjamin (Google).
3829 * rsaz_1024_mul_avx2 overflow bug on x86_64
3831 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3832 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3833 Analysis suggests that attacks against RSA and DSA as a result of this
3834 defect would be very difficult to perform and are not believed likely.
3835 Attacks against DH1024 are considered just feasible, because most of the
3836 work necessary to deduce information about a private key may be performed
3837 offline. The amount of resources required for such an attack would be
3838 significant. However, for an attack on TLS to be meaningful, the server
3839 would have to share the DH1024 private key among multiple clients, which is
3840 no longer an option since CVE-2016-0701.
3842 This only affects processors that support the AVX2 but not ADX extensions
3843 like Intel Haswell (4th generation).
3845 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3846 was originally found via the OSS-Fuzz project.
3851 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017] ###
3853 * bn_sqrx8x_internal carry bug on x86_64
3855 There is a carry propagating bug in the x86_64 Montgomery squaring
3856 procedure. No EC algorithms are affected. Analysis suggests that attacks
3857 against RSA and DSA as a result of this defect would be very difficult to
3858 perform and are not believed likely. Attacks against DH are considered just
3859 feasible (although very difficult) because most of the work necessary to
3860 deduce information about a private key may be performed offline. The amount
3861 of resources required for such an attack would be very significant and
3862 likely only accessible to a limited number of attackers. An attacker would
3863 additionally need online access to an unpatched system using the target
3864 private key in a scenario with persistent DH parameters and a private
3865 key that is shared between multiple clients.
3867 This only affects processors that support the BMI1, BMI2 and ADX extensions
3868 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3870 This issue was reported to OpenSSL by the OSS-Fuzz project.
3875 * Malformed X.509 IPAddressFamily could cause OOB read
3877 If an X.509 certificate has a malformed IPAddressFamily extension,
3878 OpenSSL could do a one-byte buffer overread. The most likely result
3879 would be an erroneous display of the certificate in text format.
3881 This issue was reported to OpenSSL by the OSS-Fuzz project.
3886 ### Changes between 1.0.2k and 1.0.2l [25 May 2017] ###
3888 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3889 platform rather than 'mingw'.
3893 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017] ###
3895 * Truncated packet could crash via OOB read
3897 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3898 cipher is being used, then a truncated packet can cause that host to
3899 perform an out-of-bounds read, usually resulting in a crash.
3901 This issue was reported to OpenSSL by Robert Święcki of Google.
3906 * BN_mod_exp may produce incorrect results on x86_64
3908 There is a carry propagating bug in the x86_64 Montgomery squaring
3909 procedure. No EC algorithms are affected. Analysis suggests that attacks
3910 against RSA and DSA as a result of this defect would be very difficult to
3911 perform and are not believed likely. Attacks against DH are considered just
3912 feasible (although very difficult) because most of the work necessary to
3913 deduce information about a private key may be performed offline. The amount
3914 of resources required for such an attack would be very significant and
3915 likely only accessible to a limited number of attackers. An attacker would
3916 additionally need online access to an unpatched system using the target
3917 private key in a scenario with persistent DH parameters and a private
3918 key that is shared between multiple clients. For example this can occur by
3919 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3920 similar to CVE-2015-3193 but must be treated as a separate problem.
3922 This issue was reported to OpenSSL by the OSS-Fuzz project.
3927 * Montgomery multiplication may produce incorrect results
3929 There is a carry propagating bug in the Broadwell-specific Montgomery
3930 multiplication procedure that handles input lengths divisible by, but
3931 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3932 and DH private keys are impossible. This is because the subroutine in
3933 question is not used in operations with the private key itself and an input
3934 of the attacker's direct choice. Otherwise the bug can manifest itself as
3935 transient authentication and key negotiation failures or reproducible
3936 erroneous outcome of public-key operations with specially crafted input.
3937 Among EC algorithms only Brainpool P-512 curves are affected and one
3938 presumably can attack ECDH key negotiation. Impact was not analyzed in
3939 detail, because pre-requisites for attack are considered unlikely. Namely
3940 multiple clients have to choose the curve in question and the server has to
3941 share the private key among them, neither of which is default behaviour.
3942 Even then only clients that chose the curve will be affected.
3944 This issue was publicly reported as transient failures and was not
3945 initially recognized as a security issue. Thanks to Richard Morgan for
3946 providing reproducible case.
3951 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3952 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3953 prevent issues where no progress is being made and the peer continually
3954 sends unrecognised record types, using up resources processing them.
3958 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016] ###
3960 * Missing CRL sanity check
3962 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
3963 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
3964 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
3966 This issue only affects the OpenSSL 1.0.2i
3971 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016] ###
3973 * OCSP Status Request extension unbounded memory growth
3975 A malicious client can send an excessively large OCSP Status Request
3976 extension. If that client continually requests renegotiation, sending a
3977 large OCSP Status Request extension each time, then there will be unbounded
3978 memory growth on the server. This will eventually lead to a Denial Of
3979 Service attack through memory exhaustion. Servers with a default
3980 configuration are vulnerable even if they do not support OCSP. Builds using
3981 the "no-ocsp" build time option are not affected.
3983 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3988 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
3991 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
3997 * OOB write in MDC2_Update()
3999 An overflow can occur in MDC2_Update() either if called directly or
4000 through the EVP_DigestUpdate() function using MDC2. If an attacker
4001 is able to supply very large amounts of input data after a previous
4002 call to EVP_EncryptUpdate() with a partial block then a length check
4003 can overflow resulting in a heap corruption.
4005 The amount of data needed is comparable to SIZE_MAX which is impractical
4008 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4013 * Malformed SHA512 ticket DoS
4015 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4016 DoS attack where a malformed ticket will result in an OOB read which will
4019 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4020 a custom server callback and ticket lookup mechanism.
4022 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4027 * OOB write in BN_bn2dec()
4029 The function BN_bn2dec() does not check the return value of BN_div_word().
4030 This can cause an OOB write if an application uses this function with an
4031 overly large BIGNUM. This could be a problem if an overly large certificate
4032 or CRL is printed out from an untrusted source. TLS is not affected because
4033 record limits will reject an oversized certificate before it is parsed.
4035 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4040 * OOB read in TS_OBJ_print_bio()
4042 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4043 the total length the OID text representation would use and not the amount
4044 of data written. This will result in OOB reads when large OIDs are
4047 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4052 * Pointer arithmetic undefined behaviour
4054 Avoid some undefined pointer arithmetic
4056 A common idiom in the codebase is to check limits in the following manner:
4059 Where "p" points to some malloc'd data of SIZE bytes and
4062 "len" here could be from some externally supplied data (e.g. from a TLS
4065 The rules of C pointer arithmetic are such that "p + len" is only well
4066 defined where len <= SIZE. Therefore the above idiom is actually
4067 undefined behaviour.
4069 For example this could cause problems if some malloc implementation
4070 provides an address for "p" such that "p + len" actually overflows for
4071 values of len that are too big and therefore p + len < limit.
4073 This issue was reported to OpenSSL by Guido Vranken
4078 * Constant time flag not preserved in DSA signing
4080 Operations in the DSA signing algorithm should run in constant time in
4081 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4082 implementation means that a non-constant time codepath is followed for
4083 certain operations. This has been demonstrated through a cache-timing
4084 attack to be sufficient for an attacker to recover the private DSA key.
4086 This issue was reported by César Pereida (Aalto University), Billy Brumley
4087 (Tampere University of Technology), and Yuval Yarom (The University of
4088 Adelaide and NICTA).
4093 * DTLS buffered message DoS
4095 In a DTLS connection where handshake messages are delivered out-of-order
4096 those messages that OpenSSL is not yet ready to process will be buffered
4097 for later use. Under certain circumstances, a flaw in the logic means that
4098 those messages do not get removed from the buffer even though the handshake
4099 has been completed. An attacker could force up to approx. 15 messages to
4100 remain in the buffer when they are no longer required. These messages will
4101 be cleared when the DTLS connection is closed. The default maximum size for
4102 a message is 100k. Therefore the attacker could force an additional 1500k
4103 to be consumed per connection. By opening many simulataneous connections an
4104 attacker could cause a DoS attack through memory exhaustion.
4106 This issue was reported to OpenSSL by Quan Luo.
4111 * DTLS replay protection DoS
4113 A flaw in the DTLS replay attack protection mechanism means that records
4114 that arrive for future epochs update the replay protection "window" before
4115 the MAC for the record has been validated. This could be exploited by an
4116 attacker by sending a record for the next epoch (which does not have to
4117 decrypt or have a valid MAC), with a very large sequence number. This means
4118 that all subsequent legitimate packets are dropped causing a denial of
4119 service for a specific DTLS connection.
4121 This issue was reported to OpenSSL by the OCAP audit team.
4126 * Certificate message OOB reads
4128 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4129 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4130 theoretical DoS risk but this has not been observed in practice on common
4133 The messages affected are client certificate, client certificate request
4134 and server certificate. As a result the attack can only be performed
4135 against a client or a server which enables client authentication.
4137 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4142 ### Changes between 1.0.2g and 1.0.2h [3 May 2016] ###
4144 * Prevent padding oracle in AES-NI CBC MAC check
4146 A MITM attacker can use a padding oracle attack to decrypt traffic
4147 when the connection uses an AES CBC cipher and the server support
4150 This issue was introduced as part of the fix for Lucky 13 padding
4151 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4152 constant time by making sure that always the same bytes are read and
4153 compared against either the MAC or padding bytes. But it no longer
4154 checked that there was enough data to have both the MAC and padding
4157 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4162 * Fix EVP_EncodeUpdate overflow
4164 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4165 Base64 encoding of binary data. If an attacker is able to supply very large
4166 amounts of input data then a length check can overflow resulting in a heap
4169 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4170 the PEM_write_bio* family of functions. These are mainly used within the
4171 OpenSSL command line applications, so any application which processes data
4172 from an untrusted source and outputs it as a PEM file should be considered
4173 vulnerable to this issue. User applications that call these APIs directly
4174 with large amounts of untrusted data may also be vulnerable.
4176 This issue was reported by Guido Vranken.
4181 * Fix EVP_EncryptUpdate overflow
4183 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4184 is able to supply very large amounts of input data after a previous call to
4185 EVP_EncryptUpdate() with a partial block then a length check can overflow
4186 resulting in a heap corruption. Following an analysis of all OpenSSL
4187 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4188 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4189 the first called function after an EVP_EncryptInit(), and therefore that
4190 specific call must be safe. The second form is where the length passed to
4191 EVP_EncryptUpdate() can be seen from the code to be some small value and
4192 therefore there is no possibility of an overflow. Since all instances are
4193 one of these two forms, it is believed that there can be no overflows in
4194 internal code due to this problem. It should be noted that
4195 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4196 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4197 of these calls have also been analysed too and it is believed there are no
4198 instances in internal usage where an overflow could occur.
4200 This issue was reported by Guido Vranken.
4205 * Prevent ASN.1 BIO excessive memory allocation
4207 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4208 a short invalid encoding can cause allocation of large amounts of memory
4209 potentially consuming excessive resources or exhausting memory.
4211 Any application parsing untrusted data through d2i BIO functions is
4212 affected. The memory based functions such as d2i_X509() are *not* affected.
4213 Since the memory based functions are used by the TLS library, TLS
4214 applications are not affected.
4216 This issue was reported by Brian Carpenter.
4223 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4224 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4225 in arbitrary stack data being returned in the buffer.
4227 This issue was reported by Guido Vranken.
4232 * Modify behavior of ALPN to invoke callback after SNI/servername
4233 callback, such that updates to the SSL_CTX affect ALPN.
4237 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4242 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4243 methods are enabled and ssl2 is disabled the methods return NULL.
4247 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016] ###
4249 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4250 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4251 provide any "EXPORT" or "LOW" strength ciphers.
4255 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4256 is by default disabled at build-time. Builds that are not configured with
4257 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4258 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4259 will need to explicitly call either of:
4261 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4263 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4265 as appropriate. Even if either of those is used, or the application
4266 explicitly uses the version-specific SSLv2_method() or its client and
4267 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4268 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4269 ciphers, and SSLv2 56-bit DES are no longer available.
4274 * Fix a double-free in DSA code
4276 A double free bug was discovered when OpenSSL parses malformed DSA private
4277 keys and could lead to a DoS attack or memory corruption for applications
4278 that receive DSA private keys from untrusted sources. This scenario is
4281 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4287 * Disable SRP fake user seed to address a server memory leak.
4289 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4291 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4292 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4293 was changed to ignore the "fake user" SRP seed, even if the seed
4296 Users should use SRP_VBASE_get1_by_user instead. Note that in
4297 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4298 also that even though configuring the SRP seed attempts to hide
4299 invalid usernames by continuing the handshake with fake
4300 credentials, this behaviour is not constant time and no strong
4301 guarantees are made that the handshake is indistinguishable from
4302 that of a valid user.
4307 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4309 In the BN_hex2bn function the number of hex digits is calculated using an
4310 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4311 large values of |i| this can result in |bn_expand| not allocating any
4312 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4313 field as NULL leading to a subsequent NULL ptr deref. For very large values
4314 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4315 In this case memory is allocated to the internal BIGNUM data field, but it
4316 is insufficiently sized leading to heap corruption. A similar issue exists
4317 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4318 is ever called by user applications with very large untrusted hex/dec data.
4319 This is anticipated to be a rare occurrence.
4321 All OpenSSL internal usage of these functions use data that is not expected
4322 to be untrusted, e.g. config file data or application command line
4323 arguments. If user developed applications generate config file data based
4324 on untrusted data then it is possible that this could also lead to security
4325 consequences. This is also anticipated to be rare.
4327 This issue was reported to OpenSSL by Guido Vranken.
4332 * Fix memory issues in BIO_*printf functions
4334 The internal |fmtstr| function used in processing a "%s" format string in
4335 the BIO_*printf functions could overflow while calculating the length of a
4336 string and cause an OOB read when printing very long strings.
4338 Additionally the internal |doapr_outch| function can attempt to write to an
4339 OOB memory location (at an offset from the NULL pointer) in the event of a
4340 memory allocation failure. In 1.0.2 and below this could be caused where
4341 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4342 could be in processing a very long "%s" format string. Memory leaks can
4345 The first issue may mask the second issue dependent on compiler behaviour.
4346 These problems could enable attacks where large amounts of untrusted data
4347 is passed to the BIO_*printf functions. If applications use these functions
4348 in this way then they could be vulnerable. OpenSSL itself uses these
4349 functions when printing out human-readable dumps of ASN.1 data. Therefore
4350 applications that print this data could be vulnerable if the data is from
4351 untrusted sources. OpenSSL command line applications could also be
4352 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4353 as command line arguments.
4355 Libssl is not considered directly vulnerable. Additionally certificates etc
4356 received via remote connections via libssl are also unlikely to be able to
4357 trigger these issues because of message size limits enforced within libssl.
4359 This issue was reported to OpenSSL Guido Vranken.
4364 * Side channel attack on modular exponentiation
4366 A side-channel attack was found which makes use of cache-bank conflicts on
4367 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4368 of RSA keys. The ability to exploit this issue is limited as it relies on
4369 an attacker who has control of code in a thread running on the same
4370 hyper-threaded core as the victim thread which is performing decryptions.
4372 This issue was reported to OpenSSL by Yuval Yarom, The University of
4373 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4374 Nadia Heninger, University of Pennsylvania with more information at
4375 http://cachebleed.info.
4380 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4381 if no keysize is specified with default_bits. This fixes an
4382 omission in an earlier change that changed all RSA/DSA key generation
4383 apps to use 2048 bits by default.
4387 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016] ###
4388 * DH small subgroups
4390 Historically OpenSSL only ever generated DH parameters based on "safe"
4391 primes. More recently (in version 1.0.2) support was provided for
4392 generating X9.42 style parameter files such as those required for RFC 5114
4393 support. The primes used in such files may not be "safe". Where an
4394 application is using DH configured with parameters based on primes that are
4395 not "safe" then an attacker could use this fact to find a peer's private
4396 DH exponent. This attack requires that the attacker complete multiple
4397 handshakes in which the peer uses the same private DH exponent. For example
4398 this could be used to discover a TLS server's private DH exponent if it's
4399 reusing the private DH exponent or it's using a static DH ciphersuite.
4401 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4402 TLS. It is not on by default. If the option is not set then the server
4403 reuses the same private DH exponent for the life of the server process and
4404 would be vulnerable to this attack. It is believed that many popular
4405 applications do set this option and would therefore not be at risk.
4407 The fix for this issue adds an additional check where a "q" parameter is
4408 available (as is the case in X9.42 based parameters). This detects the
4409 only known attack, and is the only possible defense for static DH
4410 ciphersuites. This could have some performance impact.
4412 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4413 default and cannot be disabled. This could have some performance impact.
4415 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4420 * SSLv2 doesn't block disabled ciphers
4422 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4423 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4424 been disabled, provided that the SSLv2 protocol was not also disabled via
4427 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4428 and Sebastian Schinzel.
4433 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015] ###
4435 * BN_mod_exp may produce incorrect results on x86_64
4437 There is a carry propagating bug in the x86_64 Montgomery squaring
4438 procedure. No EC algorithms are affected. Analysis suggests that attacks
4439 against RSA and DSA as a result of this defect would be very difficult to
4440 perform and are not believed likely. Attacks against DH are considered just
4441 feasible (although very difficult) because most of the work necessary to
4442 deduce information about a private key may be performed offline. The amount
4443 of resources required for such an attack would be very significant and
4444 likely only accessible to a limited number of attackers. An attacker would
4445 additionally need online access to an unpatched system using the target
4446 private key in a scenario with persistent DH parameters and a private
4447 key that is shared between multiple clients. For example this can occur by
4448 default in OpenSSL DHE based SSL/TLS ciphersuites.
4450 This issue was reported to OpenSSL by Hanno Böck.
4455 * Certificate verify crash with missing PSS parameter
4457 The signature verification routines will crash with a NULL pointer
4458 dereference if presented with an ASN.1 signature using the RSA PSS
4459 algorithm and absent mask generation function parameter. Since these
4460 routines are used to verify certificate signature algorithms this can be
4461 used to crash any certificate verification operation and exploited in a
4462 DoS attack. Any application which performs certificate verification is
4463 vulnerable including OpenSSL clients and servers which enable client
4466 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4471 * X509_ATTRIBUTE memory leak
4473 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4474 memory. This structure is used by the PKCS#7 and CMS routines so any
4475 application which reads PKCS#7 or CMS data from untrusted sources is
4476 affected. SSL/TLS is not affected.
4478 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4484 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4485 This changes the decoding behaviour for some invalid messages,
4486 though the change is mostly in the more lenient direction, and
4487 legacy behaviour is preserved as much as possible.
4491 * In DSA_generate_parameters_ex, if the provided seed is too short,
4494 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4496 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015] ###
4498 * Alternate chains certificate forgery
4500 During certificate verification, OpenSSL will attempt to find an
4501 alternative certificate chain if the first attempt to build such a chain
4502 fails. An error in the implementation of this logic can mean that an
4503 attacker could cause certain checks on untrusted certificates to be
4504 bypassed, such as the CA flag, enabling them to use a valid leaf
4505 certificate to act as a CA and "issue" an invalid certificate.
4507 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4512 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015] ###
4514 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4515 incompatibility in the handling of HMAC. The previous ABI has now been
4520 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015] ###
4522 * Malformed ECParameters causes infinite loop
4524 When processing an ECParameters structure OpenSSL enters an infinite loop
4525 if the curve specified is over a specially malformed binary polynomial
4528 This can be used to perform denial of service against any
4529 system which processes public keys, certificate requests or
4530 certificates. This includes TLS clients and TLS servers with
4531 client authentication enabled.
4533 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4538 * Exploitable out-of-bounds read in X509_cmp_time
4540 X509_cmp_time does not properly check the length of the ASN1_TIME
4541 string and can read a few bytes out of bounds. In addition,
4542 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4545 An attacker can use this to craft malformed certificates and CRLs of
4546 various sizes and potentially cause a segmentation fault, resulting in
4547 a DoS on applications that verify certificates or CRLs. TLS clients
4548 that verify CRLs are affected. TLS clients and servers with client
4549 authentication enabled may be affected if they use custom verification
4552 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4553 independently by Hanno Böck.
4558 * PKCS7 crash with missing EnvelopedContent
4560 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4561 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4562 with missing content and trigger a NULL pointer dereference on parsing.
4564 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4565 structures from untrusted sources are affected. OpenSSL clients and
4566 servers are not affected.
4568 This issue was reported to OpenSSL by Michal Zalewski (Google).
4573 * CMS verify infinite loop with unknown hash function
4575 When verifying a signedData message the CMS code can enter an infinite loop
4576 if presented with an unknown hash function OID. This can be used to perform
4577 denial of service against any system which verifies signedData messages using
4579 This issue was reported to OpenSSL by Johannes Bauer.
4584 * Race condition handling NewSessionTicket
4586 If a NewSessionTicket is received by a multi-threaded client when attempting to
4587 reuse a previous ticket then a race condition can occur potentially leading to
4588 a double free of the ticket data.
4593 * Only support 256-bit or stronger elliptic curves with the
4594 'ecdh_auto' setting (server) or by default (client). Of supported
4595 curves, prefer P-256 (both).
4599 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015] ###
4601 * ClientHello sigalgs DoS fix
4603 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4604 invalid signature algorithms extension a NULL pointer dereference will
4605 occur. This can be exploited in a DoS attack against the server.
4607 This issue was was reported to OpenSSL by David Ramos of Stanford
4611 *Stephen Henson and Matt Caswell*
4613 * Multiblock corrupted pointer fix
4615 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4616 feature only applies on 64 bit x86 architecture platforms that support AES
4617 NI instructions. A defect in the implementation of "multiblock" can cause
4618 OpenSSL's internal write buffer to become incorrectly set to NULL when
4619 using non-blocking IO. Typically, when the user application is using a
4620 socket BIO for writing, this will only result in a failed connection.
4621 However if some other BIO is used then it is likely that a segmentation
4622 fault will be triggered, thus enabling a potential DoS attack.
4624 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4629 * Segmentation fault in DTLSv1_listen fix
4631 The DTLSv1_listen function is intended to be stateless and processes the
4632 initial ClientHello from many peers. It is common for user code to loop
4633 over the call to DTLSv1_listen until a valid ClientHello is received with
4634 an associated cookie. A defect in the implementation of DTLSv1_listen means
4635 that state is preserved in the SSL object from one invocation to the next
4636 that can lead to a segmentation fault. Errors processing the initial
4637 ClientHello can trigger this scenario. An example of such an error could be
4638 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4641 This issue was reported to OpenSSL by Per Allansson.
4646 * Segmentation fault in ASN1_TYPE_cmp fix
4648 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4649 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4650 certificate signature algorithm consistency this can be used to crash any
4651 certificate verification operation and exploited in a DoS attack. Any
4652 application which performs certificate verification is vulnerable including
4653 OpenSSL clients and servers which enable client authentication.
4658 * Segmentation fault for invalid PSS parameters fix
4660 The signature verification routines will crash with a NULL pointer
4661 dereference if presented with an ASN.1 signature using the RSA PSS
4662 algorithm and invalid parameters. Since these routines are used to verify
4663 certificate signature algorithms this can be used to crash any
4664 certificate verification operation and exploited in a DoS attack. Any
4665 application which performs certificate verification is vulnerable including
4666 OpenSSL clients and servers which enable client authentication.
4668 This issue was was reported to OpenSSL by Brian Carpenter.
4673 * ASN.1 structure reuse memory corruption fix
4675 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4676 memory corruption via an invalid write. Such reuse is and has been
4677 strongly discouraged and is believed to be rare.
4679 Applications that parse structures containing CHOICE or ANY DEFINED BY
4680 components may be affected. Certificate parsing (d2i_X509 and related
4681 functions) are however not affected. OpenSSL clients and servers are
4687 * PKCS7 NULL pointer dereferences fix
4689 The PKCS#7 parsing code does not handle missing outer ContentInfo
4690 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4691 missing content and trigger a NULL pointer dereference on parsing.
4693 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4694 otherwise parse PKCS#7 structures from untrusted sources are
4695 affected. OpenSSL clients and servers are not affected.
4697 This issue was reported to OpenSSL by Michal Zalewski (Google).
4702 * DoS via reachable assert in SSLv2 servers fix
4704 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4705 servers that both support SSLv2 and enable export cipher suites by sending
4706 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4708 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4709 (OpenSSL development team).
4714 * Empty CKE with client auth and DHE fix
4716 If client auth is used then a server can seg fault in the event of a DHE
4717 ciphersuite being selected and a zero length ClientKeyExchange message
4718 being sent by the client. This could be exploited in a DoS attack.
4723 * Handshake with unseeded PRNG fix
4725 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4726 with an unseeded PRNG. The conditions are:
4727 - The client is on a platform where the PRNG has not been seeded
4728 automatically, and the user has not seeded manually
4729 - A protocol specific client method version has been used (i.e. not
4730 SSL_client_methodv23)
4731 - A ciphersuite is used that does not require additional random data from
4732 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4734 If the handshake succeeds then the client random that has been used will
4735 have been generated from a PRNG with insufficient entropy and therefore the
4736 output may be predictable.
4738 For example using the following command with an unseeded openssl will
4739 succeed on an unpatched platform:
4741 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4746 * Use After Free following d2i_ECPrivatekey error fix
4748 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4749 could cause a use after free condition. This, in turn, could cause a double
4750 free in several private key parsing functions (such as d2i_PrivateKey
4751 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4752 for applications that receive EC private keys from untrusted
4753 sources. This scenario is considered rare.
4755 This issue was discovered by the BoringSSL project and fixed in their
4761 * X509_to_X509_REQ NULL pointer deref fix
4763 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4764 the certificate key is invalid. This function is rarely used in practice.
4766 This issue was discovered by Brian Carpenter.
4771 * Removed the export ciphers from the DEFAULT ciphers
4775 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015] ###
4777 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4778 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4779 So far those who have to target multiple platforms would compromise
4780 and argue that binary targeting say ARMv5 would still execute on
4781 ARMv8. "Universal" build resolves this compromise by providing
4782 near-optimal performance even on newer platforms.
4786 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4787 (other platforms pending).
4789 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4791 * Add support for the SignedCertificateTimestampList certificate and
4792 OCSP response extensions from RFC6962.
4796 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4797 for corner cases. (Certain input points at infinity could lead to
4798 bogus results, with non-infinity inputs mapped to infinity too.)
4802 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4803 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4804 common cases are optimized and there still is room for further
4805 improvements. Vector Permutation AES for Altivec is also added.
4809 * Add support for little-endian ppc64 Linux target.
4811 *Marcelo Cerri (IBM)*
4813 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4814 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4815 are optimized and there still is room for further improvements.
4816 Both 32- and 64-bit modes are supported.
4818 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4820 * Improved ARMv7 NEON support.
4824 * Support for SPARC Architecture 2011 crypto extensions, first
4825 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4826 SHA256/512, MD5, GHASH and modular exponentiation.
4828 *Andy Polyakov, David Miller*
4830 * Accelerated modular exponentiation for Intel processors, a.k.a.
4833 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4835 * Support for new and upcoming Intel processors, including AVX2,
4836 BMI and SHA ISA extensions. This includes additional "stitched"
4837 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4840 This work was sponsored by Intel Corp.
4844 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4845 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4846 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4850 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4851 this fixes a limitation in previous versions of OpenSSL.
4855 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4856 MGF1 digest and OAEP label.
4860 * Add EVP support for key wrapping algorithms, to avoid problems with
4861 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4862 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4863 algorithms and include tests cases.
4867 * Add functions to allocate and set the fields of an ECDSA_METHOD
4870 *Douglas E. Engert, Steve Henson*
4872 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4873 difference in days and seconds between two tm or ASN1_TIME structures.
4877 * Add -rev test option to s_server to just reverse order of characters
4878 received by client and send back to server. Also prints an abbreviated
4879 summary of the connection parameters.
4883 * New option -brief for s_client and s_server to print out a brief summary
4884 of connection parameters.
4888 * Add callbacks for arbitrary TLS extensions.
4890 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4892 * New option -crl_download in several openssl utilities to download CRLs
4893 from CRLDP extension in certificates.
4897 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4901 * New function X509_CRL_diff to generate a delta CRL from the difference
4902 of two full CRLs. Add support to "crl" utility.
4906 * New functions to set lookup_crls function and to retrieve
4907 X509_STORE from X509_STORE_CTX.
4911 * Print out deprecated issuer and subject unique ID fields in
4916 * Extend OCSP I/O functions so they can be used for simple general purpose
4917 HTTP as well as OCSP. New wrapper function which can be used to download
4918 CRLs using the OCSP API.
4922 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
4926 * SSL_CONF* functions. These provide a common framework for application
4927 configuration using configuration files or command lines.
4931 * SSL/TLS tracing code. This parses out SSL/TLS records using the
4932 message callback and prints the results. Needs compile time option
4933 "enable-ssl-trace". New options to s_client and s_server to enable
4938 * New ctrl and macro to retrieve supported points extensions.
4939 Print out extension in s_server and s_client.
4943 * New functions to retrieve certificate signature and signature
4948 * Add functions to retrieve and manipulate the raw cipherlist sent by a
4953 * New Suite B modes for TLS code. These use and enforce the requirements
4954 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
4955 only use Suite B curves. The Suite B modes can be set by using the
4956 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
4960 * New chain verification flags for Suite B levels of security. Check
4961 algorithms are acceptable when flags are set in X509_verify_cert.
4965 * Make tls1_check_chain return a set of flags indicating checks passed
4966 by a certificate chain. Add additional tests to handle client
4967 certificates: checks for matching certificate type and issuer name
4972 * If an attempt is made to use a signature algorithm not in the peer
4973 preference list abort the handshake. If client has no suitable
4974 signature algorithms in response to a certificate request do not
4975 use the certificate.
4979 * If server EC tmp key is not in client preference list abort handshake.
4983 * Add support for certificate stores in CERT structure. This makes it
4984 possible to have different stores per SSL structure or one store in
4985 the parent SSL_CTX. Include distinct stores for certificate chain
4986 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
4987 to build and store a certificate chain in CERT structure: returning
4988 an error if the chain cannot be built: this will allow applications
4989 to test if a chain is correctly configured.
4991 Note: if the CERT based stores are not set then the parent SSL_CTX
4992 store is used to retain compatibility with existing behaviour.
4997 * New function ssl_set_client_disabled to set a ciphersuite disabled
4998 mask based on the current session, check mask when sending client
4999 hello and checking the requested ciphersuite.
5003 * New ctrls to retrieve and set certificate types in a certificate
5004 request message. Print out received values in s_client. If certificate
5005 types is not set with custom values set sensible values based on
5006 supported signature algorithms.
5010 * Support for distinct client and server supported signature algorithms.
5014 * Add certificate callback. If set this is called whenever a certificate
5015 is required by client or server. An application can decide which
5016 certificate chain to present based on arbitrary criteria: for example
5017 supported signature algorithms. Add very simple example to s_server.
5018 This fixes many of the problems and restrictions of the existing client
5019 certificate callback: for example you can now clear an existing
5020 certificate and specify the whole chain.
5024 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5025 the certificate can be used for (if anything). Set valid_flags field
5026 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5027 to have similar checks in it.
5029 Add new "cert_flags" field to CERT structure and include a "strict mode".
5030 This enforces some TLS certificate requirements (such as only permitting
5031 certificate signature algorithms contained in the supported algorithms
5032 extension) which some implementations ignore: this option should be used
5033 with caution as it could cause interoperability issues.
5037 * Update and tidy signature algorithm extension processing. Work out
5038 shared signature algorithms based on preferences and peer algorithms
5039 and print them out in s_client and s_server. Abort handshake if no
5040 shared signature algorithms.
5044 * Add new functions to allow customised supported signature algorithms
5045 for SSL and SSL_CTX structures. Add options to s_client and s_server
5050 * New function SSL_certs_clear() to delete all references to certificates
5051 from an SSL structure. Before this once a certificate had been added
5052 it couldn't be removed.
5056 * Integrate hostname, email address and IP address checking with certificate
5057 verification. New verify options supporting checking in openssl utility.
5061 * Fixes and wildcard matching support to hostname and email checking
5062 functions. Add manual page.
5064 *Florian Weimer (Red Hat Product Security Team)*
5066 * New functions to check a hostname email or IP address against a
5067 certificate. Add options x509 utility to print results of checks against
5072 * Fix OCSP checking.
5074 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5076 * Initial experimental support for explicitly trusted non-root CAs.
5077 OpenSSL still tries to build a complete chain to a root but if an
5078 intermediate CA has a trust setting included that is used. The first
5079 setting is used: whether to trust (e.g., -addtrust option to the x509
5084 * Add -trusted_first option which attempts to find certificates in the
5085 trusted store even if an untrusted chain is also supplied.
5089 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5090 platform support for Linux and Android.
5094 * Support for linux-x32, ILP32 environment in x86_64 framework.
5098 * Experimental multi-implementation support for FIPS capable OpenSSL.
5099 When in FIPS mode the approved implementations are used as normal,
5100 when not in FIPS mode the internal unapproved versions are used instead.
5101 This means that the FIPS capable OpenSSL isn't forced to use the
5102 (often lower performance) FIPS implementations outside FIPS mode.
5106 * Transparently support X9.42 DH parameters when calling
5107 PEM_read_bio_DHparameters. This means existing applications can handle
5108 the new parameter format automatically.
5112 * Initial experimental support for X9.42 DH parameter format: mainly
5113 to support use of 'q' parameter for RFC5114 parameters.
5117 * Add DH parameters from RFC5114 including test data to dhtest.
5121 * Support for automatic EC temporary key parameter selection. If enabled
5122 the most preferred EC parameters are automatically used instead of
5123 hardcoded fixed parameters. Now a server just has to call:
5124 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5125 support ECDH and use the most appropriate parameters.
5129 * Enhance and tidy EC curve and point format TLS extension code. Use
5130 static structures instead of allocation if default values are used.
5131 New ctrls to set curves we wish to support and to retrieve shared curves.
5132 Print out shared curves in s_server. New options to s_server and s_client
5133 to set list of supported curves.
5137 * New ctrls to retrieve supported signature algorithms and
5138 supported curve values as an array of NIDs. Extend openssl utility
5139 to print out received values.
5143 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5144 between NIDs and the more common NIST names such as "P-256". Enhance
5145 ecparam utility and ECC method to recognise the NIST names for curves.
5149 * Enhance SSL/TLS certificate chain handling to support different
5150 chains for each certificate instead of one chain in the parent SSL_CTX.
5154 * Support for fixed DH ciphersuite client authentication: where both
5155 server and client use DH certificates with common parameters.
5159 * Support for fixed DH ciphersuites: those requiring DH server
5164 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5166 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5167 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5168 X509_CINF_get_signature were reverted post internal team review.
5174 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016] ###
5176 * OCSP Status Request extension unbounded memory growth
5178 A malicious client can send an excessively large OCSP Status Request
5179 extension. If that client continually requests renegotiation, sending a
5180 large OCSP Status Request extension each time, then there will be unbounded
5181 memory growth on the server. This will eventually lead to a Denial Of
5182 Service attack through memory exhaustion. Servers with a default
5183 configuration are vulnerable even if they do not support OCSP. Builds using
5184 the "no-ocsp" build time option are not affected.
5186 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5191 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5194 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5200 * OOB write in MDC2_Update()
5202 An overflow can occur in MDC2_Update() either if called directly or
5203 through the EVP_DigestUpdate() function using MDC2. If an attacker
5204 is able to supply very large amounts of input data after a previous
5205 call to EVP_EncryptUpdate() with a partial block then a length check
5206 can overflow resulting in a heap corruption.
5208 The amount of data needed is comparable to SIZE_MAX which is impractical
5211 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5216 * Malformed SHA512 ticket DoS
5218 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5219 DoS attack where a malformed ticket will result in an OOB read which will
5222 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5223 a custom server callback and ticket lookup mechanism.
5225 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5230 * OOB write in BN_bn2dec()
5232 The function BN_bn2dec() does not check the return value of BN_div_word().
5233 This can cause an OOB write if an application uses this function with an
5234 overly large BIGNUM. This could be a problem if an overly large certificate
5235 or CRL is printed out from an untrusted source. TLS is not affected because
5236 record limits will reject an oversized certificate before it is parsed.
5238 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5243 * OOB read in TS_OBJ_print_bio()
5245 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5246 the total length the OID text representation would use and not the amount
5247 of data written. This will result in OOB reads when large OIDs are
5250 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5255 * Pointer arithmetic undefined behaviour
5257 Avoid some undefined pointer arithmetic
5259 A common idiom in the codebase is to check limits in the following manner:
5262 Where "p" points to some malloc'd data of SIZE bytes and
5265 "len" here could be from some externally supplied data (e.g. from a TLS
5268 The rules of C pointer arithmetic are such that "p + len" is only well
5269 defined where len <= SIZE. Therefore the above idiom is actually
5270 undefined behaviour.
5272 For example this could cause problems if some malloc implementation
5273 provides an address for "p" such that "p + len" actually overflows for
5274 values of len that are too big and therefore p + len < limit.
5276 This issue was reported to OpenSSL by Guido Vranken
5281 * Constant time flag not preserved in DSA signing
5283 Operations in the DSA signing algorithm should run in constant time in
5284 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5285 implementation means that a non-constant time codepath is followed for
5286 certain operations. This has been demonstrated through a cache-timing
5287 attack to be sufficient for an attacker to recover the private DSA key.
5289 This issue was reported by César Pereida (Aalto University), Billy Brumley
5290 (Tampere University of Technology), and Yuval Yarom (The University of
5291 Adelaide and NICTA).
5296 * DTLS buffered message DoS
5298 In a DTLS connection where handshake messages are delivered out-of-order
5299 those messages that OpenSSL is not yet ready to process will be buffered
5300 for later use. Under certain circumstances, a flaw in the logic means that
5301 those messages do not get removed from the buffer even though the handshake
5302 has been completed. An attacker could force up to approx. 15 messages to
5303 remain in the buffer when they are no longer required. These messages will
5304 be cleared when the DTLS connection is closed. The default maximum size for
5305 a message is 100k. Therefore the attacker could force an additional 1500k
5306 to be consumed per connection. By opening many simulataneous connections an
5307 attacker could cause a DoS attack through memory exhaustion.
5309 This issue was reported to OpenSSL by Quan Luo.
5314 * DTLS replay protection DoS
5316 A flaw in the DTLS replay attack protection mechanism means that records
5317 that arrive for future epochs update the replay protection "window" before
5318 the MAC for the record has been validated. This could be exploited by an
5319 attacker by sending a record for the next epoch (which does not have to
5320 decrypt or have a valid MAC), with a very large sequence number. This means
5321 that all subsequent legitimate packets are dropped causing a denial of
5322 service for a specific DTLS connection.
5324 This issue was reported to OpenSSL by the OCAP audit team.
5329 * Certificate message OOB reads
5331 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5332 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5333 theoretical DoS risk but this has not been observed in practice on common
5336 The messages affected are client certificate, client certificate request
5337 and server certificate. As a result the attack can only be performed
5338 against a client or a server which enables client authentication.
5340 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5345 ### Changes between 1.0.1s and 1.0.1t [3 May 2016] ###
5347 * Prevent padding oracle in AES-NI CBC MAC check
5349 A MITM attacker can use a padding oracle attack to decrypt traffic
5350 when the connection uses an AES CBC cipher and the server support
5353 This issue was introduced as part of the fix for Lucky 13 padding
5354 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5355 constant time by making sure that always the same bytes are read and
5356 compared against either the MAC or padding bytes. But it no longer
5357 checked that there was enough data to have both the MAC and padding
5360 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5365 * Fix EVP_EncodeUpdate overflow
5367 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5368 Base64 encoding of binary data. If an attacker is able to supply very large
5369 amounts of input data then a length check can overflow resulting in a heap
5372 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5373 the PEM_write_bio* family of functions. These are mainly used within the
5374 OpenSSL command line applications, so any application which processes data
5375 from an untrusted source and outputs it as a PEM file should be considered
5376 vulnerable to this issue. User applications that call these APIs directly
5377 with large amounts of untrusted data may also be vulnerable.
5379 This issue was reported by Guido Vranken.
5384 * Fix EVP_EncryptUpdate overflow
5386 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5387 is able to supply very large amounts of input data after a previous call to
5388 EVP_EncryptUpdate() with a partial block then a length check can overflow
5389 resulting in a heap corruption. Following an analysis of all OpenSSL
5390 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5391 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5392 the first called function after an EVP_EncryptInit(), and therefore that
5393 specific call must be safe. The second form is where the length passed to
5394 EVP_EncryptUpdate() can be seen from the code to be some small value and
5395 therefore there is no possibility of an overflow. Since all instances are
5396 one of these two forms, it is believed that there can be no overflows in
5397 internal code due to this problem. It should be noted that
5398 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5399 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5400 of these calls have also been analysed too and it is believed there are no
5401 instances in internal usage where an overflow could occur.
5403 This issue was reported by Guido Vranken.
5408 * Prevent ASN.1 BIO excessive memory allocation
5410 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5411 a short invalid encoding can casuse allocation of large amounts of memory
5412 potentially consuming excessive resources or exhausting memory.
5414 Any application parsing untrusted data through d2i BIO functions is
5415 affected. The memory based functions such as d2i_X509() are *not* affected.
5416 Since the memory based functions are used by the TLS library, TLS
5417 applications are not affected.
5419 This issue was reported by Brian Carpenter.
5426 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5427 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5428 in arbitrary stack data being returned in the buffer.
5430 This issue was reported by Guido Vranken.
5435 * Modify behavior of ALPN to invoke callback after SNI/servername
5436 callback, such that updates to the SSL_CTX affect ALPN.
5440 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5445 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5446 methods are enabled and ssl2 is disabled the methods return NULL.
5450 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016] ###
5452 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5453 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5454 provide any "EXPORT" or "LOW" strength ciphers.
5458 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5459 is by default disabled at build-time. Builds that are not configured with
5460 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5461 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5462 will need to explicitly call either of:
5464 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5466 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5468 as appropriate. Even if either of those is used, or the application
5469 explicitly uses the version-specific SSLv2_method() or its client and
5470 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5471 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5472 ciphers, and SSLv2 56-bit DES are no longer available.
5477 * Fix a double-free in DSA code
5479 A double free bug was discovered when OpenSSL parses malformed DSA private
5480 keys and could lead to a DoS attack or memory corruption for applications
5481 that receive DSA private keys from untrusted sources. This scenario is
5484 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5490 * Disable SRP fake user seed to address a server memory leak.
5492 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5494 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5495 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5496 was changed to ignore the "fake user" SRP seed, even if the seed
5499 Users should use SRP_VBASE_get1_by_user instead. Note that in
5500 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5501 also that even though configuring the SRP seed attempts to hide
5502 invalid usernames by continuing the handshake with fake
5503 credentials, this behaviour is not constant time and no strong
5504 guarantees are made that the handshake is indistinguishable from
5505 that of a valid user.
5510 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5512 In the BN_hex2bn function the number of hex digits is calculated using an
5513 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5514 large values of |i| this can result in |bn_expand| not allocating any
5515 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5516 field as NULL leading to a subsequent NULL ptr deref. For very large values
5517 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5518 In this case memory is allocated to the internal BIGNUM data field, but it
5519 is insufficiently sized leading to heap corruption. A similar issue exists
5520 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5521 is ever called by user applications with very large untrusted hex/dec data.
5522 This is anticipated to be a rare occurrence.
5524 All OpenSSL internal usage of these functions use data that is not expected
5525 to be untrusted, e.g. config file data or application command line
5526 arguments. If user developed applications generate config file data based
5527 on untrusted data then it is possible that this could also lead to security
5528 consequences. This is also anticipated to be rare.
5530 This issue was reported to OpenSSL by Guido Vranken.
5535 * Fix memory issues in BIO_*printf functions
5537 The internal |fmtstr| function used in processing a "%s" format string in
5538 the BIO_*printf functions could overflow while calculating the length of a
5539 string and cause an OOB read when printing very long strings.
5541 Additionally the internal |doapr_outch| function can attempt to write to an
5542 OOB memory location (at an offset from the NULL pointer) in the event of a
5543 memory allocation failure. In 1.0.2 and below this could be caused where
5544 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5545 could be in processing a very long "%s" format string. Memory leaks can
5548 The first issue may mask the second issue dependent on compiler behaviour.
5549 These problems could enable attacks where large amounts of untrusted data
5550 is passed to the BIO_*printf functions. If applications use these functions
5551 in this way then they could be vulnerable. OpenSSL itself uses these
5552 functions when printing out human-readable dumps of ASN.1 data. Therefore
5553 applications that print this data could be vulnerable if the data is from
5554 untrusted sources. OpenSSL command line applications could also be
5555 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5556 as command line arguments.
5558 Libssl is not considered directly vulnerable. Additionally certificates etc
5559 received via remote connections via libssl are also unlikely to be able to
5560 trigger these issues because of message size limits enforced within libssl.
5562 This issue was reported to OpenSSL Guido Vranken.
5567 * Side channel attack on modular exponentiation
5569 A side-channel attack was found which makes use of cache-bank conflicts on
5570 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5571 of RSA keys. The ability to exploit this issue is limited as it relies on
5572 an attacker who has control of code in a thread running on the same
5573 hyper-threaded core as the victim thread which is performing decryptions.
5575 This issue was reported to OpenSSL by Yuval Yarom, The University of
5576 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5577 Nadia Heninger, University of Pennsylvania with more information at
5578 http://cachebleed.info.
5583 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5584 if no keysize is specified with default_bits. This fixes an
5585 omission in an earlier change that changed all RSA/DSA key generation
5586 apps to use 2048 bits by default.
5590 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016] ###
5592 * Protection for DH small subgroup attacks
5594 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5595 switched on by default and cannot be disabled. This could have some
5600 * SSLv2 doesn't block disabled ciphers
5602 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5603 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5604 been disabled, provided that the SSLv2 protocol was not also disabled via
5607 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5608 and Sebastian Schinzel.
5613 * Reject DH handshakes with parameters shorter than 1024 bits.
5617 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015] ###
5619 * Certificate verify crash with missing PSS parameter
5621 The signature verification routines will crash with a NULL pointer
5622 dereference if presented with an ASN.1 signature using the RSA PSS
5623 algorithm and absent mask generation function parameter. Since these
5624 routines are used to verify certificate signature algorithms this can be
5625 used to crash any certificate verification operation and exploited in a
5626 DoS attack. Any application which performs certificate verification is
5627 vulnerable including OpenSSL clients and servers which enable client
5630 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5635 * X509_ATTRIBUTE memory leak
5637 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5638 memory. This structure is used by the PKCS#7 and CMS routines so any
5639 application which reads PKCS#7 or CMS data from untrusted sources is
5640 affected. SSL/TLS is not affected.
5642 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5648 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5649 This changes the decoding behaviour for some invalid messages,
5650 though the change is mostly in the more lenient direction, and
5651 legacy behaviour is preserved as much as possible.
5655 * In DSA_generate_parameters_ex, if the provided seed is too short,
5656 use a random seed, as already documented.
5658 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5660 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015] ###
5662 * Alternate chains certificate forgery
5664 During certificate verfification, OpenSSL will attempt to find an
5665 alternative certificate chain if the first attempt to build such a chain
5666 fails. An error in the implementation of this logic can mean that an
5667 attacker could cause certain checks on untrusted certificates to be
5668 bypassed, such as the CA flag, enabling them to use a valid leaf
5669 certificate to act as a CA and "issue" an invalid certificate.
5671 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5677 * Race condition handling PSK identify hint
5679 If PSK identity hints are received by a multi-threaded client then
5680 the values are wrongly updated in the parent SSL_CTX structure. This can
5681 result in a race condition potentially leading to a double free of the
5687 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015] ###
5688 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5689 incompatibility in the handling of HMAC. The previous ABI has now been
5692 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015] ###
5694 * Malformed ECParameters causes infinite loop
5696 When processing an ECParameters structure OpenSSL enters an infinite loop
5697 if the curve specified is over a specially malformed binary polynomial
5700 This can be used to perform denial of service against any
5701 system which processes public keys, certificate requests or
5702 certificates. This includes TLS clients and TLS servers with
5703 client authentication enabled.
5705 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5710 * Exploitable out-of-bounds read in X509_cmp_time
5712 X509_cmp_time does not properly check the length of the ASN1_TIME
5713 string and can read a few bytes out of bounds. In addition,
5714 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5717 An attacker can use this to craft malformed certificates and CRLs of
5718 various sizes and potentially cause a segmentation fault, resulting in
5719 a DoS on applications that verify certificates or CRLs. TLS clients
5720 that verify CRLs are affected. TLS clients and servers with client
5721 authentication enabled may be affected if they use custom verification
5724 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5725 independently by Hanno Böck.
5730 * PKCS7 crash with missing EnvelopedContent
5732 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5733 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5734 with missing content and trigger a NULL pointer dereference on parsing.
5736 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5737 structures from untrusted sources are affected. OpenSSL clients and
5738 servers are not affected.
5740 This issue was reported to OpenSSL by Michal Zalewski (Google).
5745 * CMS verify infinite loop with unknown hash function
5747 When verifying a signedData message the CMS code can enter an infinite loop
5748 if presented with an unknown hash function OID. This can be used to perform
5749 denial of service against any system which verifies signedData messages using
5751 This issue was reported to OpenSSL by Johannes Bauer.
5756 * Race condition handling NewSessionTicket
5758 If a NewSessionTicket is received by a multi-threaded client when attempting to
5759 reuse a previous ticket then a race condition can occur potentially leading to
5760 a double free of the ticket data.
5765 * Reject DH handshakes with parameters shorter than 768 bits.
5767 *Kurt Roeckx and Emilia Kasper*
5769 * dhparam: generate 2048-bit parameters by default.
5771 *Kurt Roeckx and Emilia Kasper*
5773 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015] ###
5775 * Segmentation fault in ASN1_TYPE_cmp fix
5777 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5778 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5779 certificate signature algorithm consistency this can be used to crash any
5780 certificate verification operation and exploited in a DoS attack. Any
5781 application which performs certificate verification is vulnerable including
5782 OpenSSL clients and servers which enable client authentication.
5787 * ASN.1 structure reuse memory corruption fix
5789 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5790 memory corruption via an invalid write. Such reuse is and has been
5791 strongly discouraged and is believed to be rare.
5793 Applications that parse structures containing CHOICE or ANY DEFINED BY
5794 components may be affected. Certificate parsing (d2i_X509 and related
5795 functions) are however not affected. OpenSSL clients and servers are
5801 * PKCS7 NULL pointer dereferences fix
5803 The PKCS#7 parsing code does not handle missing outer ContentInfo
5804 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5805 missing content and trigger a NULL pointer dereference on parsing.
5807 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5808 otherwise parse PKCS#7 structures from untrusted sources are
5809 affected. OpenSSL clients and servers are not affected.
5811 This issue was reported to OpenSSL by Michal Zalewski (Google).
5816 * DoS via reachable assert in SSLv2 servers fix
5818 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5819 servers that both support SSLv2 and enable export cipher suites by sending
5820 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5822 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5823 (OpenSSL development team).
5828 * Use After Free following d2i_ECPrivatekey error fix
5830 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5831 could cause a use after free condition. This, in turn, could cause a double
5832 free in several private key parsing functions (such as d2i_PrivateKey
5833 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5834 for applications that receive EC private keys from untrusted
5835 sources. This scenario is considered rare.
5837 This issue was discovered by the BoringSSL project and fixed in their
5843 * X509_to_X509_REQ NULL pointer deref fix
5845 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5846 the certificate key is invalid. This function is rarely used in practice.
5848 This issue was discovered by Brian Carpenter.
5853 * Removed the export ciphers from the DEFAULT ciphers
5857 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015] ###
5859 * Build fixes for the Windows and OpenVMS platforms
5861 *Matt Caswell and Richard Levitte*
5863 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015] ###
5865 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5866 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5867 dereference. This could lead to a Denial Of Service attack. Thanks to
5868 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5873 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5874 dtls1_buffer_record function under certain conditions. In particular this
5875 could occur if an attacker sent repeated DTLS records with the same
5876 sequence number but for the next epoch. The memory leak could be exploited
5877 by an attacker in a Denial of Service attack through memory exhaustion.
5878 Thanks to Chris Mueller for reporting this issue.
5883 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5884 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5885 method would be set to NULL which could later result in a NULL pointer
5886 dereference. Thanks to Frank Schmirler for reporting this issue.
5891 * Abort handshake if server key exchange message is omitted for ephemeral
5894 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5895 reporting this issue.
5900 * Remove non-export ephemeral RSA code on client and server. This code
5901 violated the TLS standard by allowing the use of temporary RSA keys in
5902 non-export ciphersuites and could be used by a server to effectively
5903 downgrade the RSA key length used to a value smaller than the server
5904 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
5905 INRIA or reporting this issue.
5910 * Fixed issue where DH client certificates are accepted without verification.
5911 An OpenSSL server will accept a DH certificate for client authentication
5912 without the certificate verify message. This effectively allows a client to
5913 authenticate without the use of a private key. This only affects servers
5914 which trust a client certificate authority which issues certificates
5915 containing DH keys: these are extremely rare and hardly ever encountered.
5916 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
5922 * Ensure that the session ID context of an SSL is updated when its
5923 SSL_CTX is updated via SSL_set_SSL_CTX.
5925 The session ID context is typically set from the parent SSL_CTX,
5926 and can vary with the CTX.
5930 * Fix various certificate fingerprint issues.
5932 By using non-DER or invalid encodings outside the signed portion of a
5933 certificate the fingerprint can be changed without breaking the signature.
5934 Although no details of the signed portion of the certificate can be changed
5935 this can cause problems with some applications: e.g. those using the
5936 certificate fingerprint for blacklists.
5938 1. Reject signatures with non zero unused bits.
5940 If the BIT STRING containing the signature has non zero unused bits reject
5941 the signature. All current signature algorithms require zero unused bits.
5943 2. Check certificate algorithm consistency.
5945 Check the AlgorithmIdentifier inside TBS matches the one in the
5946 certificate signature. NB: this will result in signature failure
5947 errors for some broken certificates.
5949 Thanks to Konrad Kraszewski from Google for reporting this issue.
5951 3. Check DSA/ECDSA signatures use DER.
5953 Re-encode DSA/ECDSA signatures and compare with the original received
5954 signature. Return an error if there is a mismatch.
5956 This will reject various cases including garbage after signature
5957 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
5958 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
5959 (negative or with leading zeroes).
5961 Further analysis was conducted and fixes were developed by Stephen Henson
5962 of the OpenSSL core team.
5968 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
5969 results on some platforms, including x86_64. This bug occurs at random
5970 with a very low probability, and is not known to be exploitable in any
5971 way, though its exact impact is difficult to determine. Thanks to Pieter
5972 Wuille (Blockstream) who reported this issue and also suggested an initial
5973 fix. Further analysis was conducted by the OpenSSL development team and
5974 Adam Langley of Google. The final fix was developed by Andy Polyakov of
5975 the OpenSSL core team.
5980 *) Do not resume sessions on the server if the negotiated protocol
5981 version does not match the session's version. Resuming with a different
5982 version, while not strictly forbidden by the RFC, is of questionable
5983 sanity and breaks all known clients.
5985 *David Benjamin, Emilia Käsper*
5987 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
5988 early CCS messages during renegotiation. (Note that because
5989 renegotiation is encrypted, this early CCS was not exploitable.)
5993 *) Tighten client-side session ticket handling during renegotiation:
5994 ensure that the client only accepts a session ticket if the server sends
5995 the extension anew in the ServerHello. Previously, a TLS client would
5996 reuse the old extension state and thus accept a session ticket if one was
5997 announced in the initial ServerHello.
5999 Similarly, ensure that the client requires a session ticket if one
6000 was advertised in the ServerHello. Previously, a TLS client would
6001 ignore a missing NewSessionTicket message.
6005 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014] ###
6009 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6010 sends a carefully crafted handshake message, to cause OpenSSL to fail
6011 to free up to 64k of memory causing a memory leak. This could be
6012 exploited in a Denial Of Service attack. This issue affects OpenSSL
6013 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6014 whether SRTP is used or configured. Implementations of OpenSSL that
6015 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6017 The fix was developed by the OpenSSL team.
6022 * Session Ticket Memory Leak.
6024 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6025 integrity of that ticket is first verified. In the event of a session
6026 ticket integrity check failing, OpenSSL will fail to free memory
6027 causing a memory leak. By sending a large number of invalid session
6028 tickets an attacker could exploit this issue in a Denial Of Service
6034 * Build option no-ssl3 is incomplete.
6036 When OpenSSL is configured with "no-ssl3" as a build option, servers
6037 could accept and complete a SSL 3.0 handshake, and clients could be
6038 configured to send them.
6041 *Akamai and the OpenSSL team*
6043 * Add support for TLS_FALLBACK_SCSV.
6044 Client applications doing fallback retries should call
6045 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6048 *Adam Langley, Bodo Moeller*
6050 * Add additional DigestInfo checks.
6052 Re-encode DigestInto in DER and check against the original when
6053 verifying RSA signature: this will reject any improperly encoded
6054 DigestInfo structures.
6056 Note: this is a precautionary measure and no attacks are currently known.
6061 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014] ###
6063 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6064 SRP code can be overrun an internal buffer. Add sanity check that
6065 g, A, B < N to SRP code.
6067 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6068 Group for discovering this issue.
6073 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6074 TLS 1.0 instead of higher protocol versions when the ClientHello message
6075 is badly fragmented. This allows a man-in-the-middle attacker to force a
6076 downgrade to TLS 1.0 even if both the server and the client support a
6077 higher protocol version, by modifying the client's TLS records.
6079 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6080 researching this issue.
6085 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6086 to a denial of service attack. A malicious server can crash the client
6087 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6088 ciphersuite and sending carefully crafted handshake messages.
6090 Thanks to Felix Gröbert (Google) for discovering and researching this
6096 * By sending carefully crafted DTLS packets an attacker could cause openssl
6097 to leak memory. This can be exploited through a Denial of Service attack.
6098 Thanks to Adam Langley for discovering and researching this issue.
6103 * An attacker can force openssl to consume large amounts of memory whilst
6104 processing DTLS handshake messages. This can be exploited through a
6105 Denial of Service attack.
6106 Thanks to Adam Langley for discovering and researching this issue.
6111 * An attacker can force an error condition which causes openssl to crash
6112 whilst processing DTLS packets due to memory being freed twice. This
6113 can be exploited through a Denial of Service attack.
6114 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6120 * If a multithreaded client connects to a malicious server using a resumed
6121 session and the server sends an ec point format extension it could write
6122 up to 255 bytes to freed memory.
6124 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6130 * A malicious server can crash an OpenSSL client with a null pointer
6131 dereference (read) by specifying an SRP ciphersuite even though it was not
6132 properly negotiated with the client. This can be exploited through a
6133 Denial of Service attack.
6135 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6136 discovering and researching this issue.
6141 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6142 X509_name_oneline, X509_name_print_ex et al. to leak some information
6143 from the stack. Applications may be affected if they echo pretty printing
6144 output to the attacker.
6146 Thanks to Ivan Fratric (Google) for discovering this issue.
6149 *Emilia Käsper, and Steve Henson*
6151 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6152 for corner cases. (Certain input points at infinity could lead to
6153 bogus results, with non-infinity inputs mapped to infinity too.)
6157 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014] ###
6159 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6160 handshake can force the use of weak keying material in OpenSSL
6161 SSL/TLS clients and servers.
6163 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6164 researching this issue. [CVE-2014-0224][]
6166 *KIKUCHI Masashi, Steve Henson*
6168 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6169 OpenSSL DTLS client the code can be made to recurse eventually crashing
6172 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6175 *Imre Rad, Steve Henson*
6177 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6178 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6179 client or server. This is potentially exploitable to run arbitrary
6180 code on a vulnerable client or server.
6182 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6184 *Jüri Aedla, Steve Henson*
6186 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6187 are subject to a denial of service attack.
6189 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6190 this issue. [CVE-2014-3470][]
6192 *Felix Gröbert, Ivan Fratric, Steve Henson*
6194 * Harmonize version and its documentation. -f flag is used to display
6197 *mancha <mancha1@zoho.com>*
6199 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6200 in i2d_ECPrivateKey.
6202 *mancha <mancha1@zoho.com>*
6204 * Fix some double frees. These are not thought to be exploitable.
6206 *mancha <mancha1@zoho.com>*
6208 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014] ###
6210 * A missing bounds check in the handling of the TLS heartbeat extension
6211 can be used to reveal up to 64k of memory to a connected client or
6214 Thanks for Neel Mehta of Google Security for discovering this bug and to
6215 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6216 preparing the fix [CVE-2014-0160][]
6218 *Adam Langley, Bodo Moeller*
6220 * Fix for the attack described in the paper "Recovering OpenSSL
6221 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6222 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6223 http://eprint.iacr.org/2014/140
6225 Thanks to Yuval Yarom and Naomi Benger for discovering this
6226 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6228 *Yuval Yarom and Naomi Benger*
6230 * TLS pad extension: draft-agl-tls-padding-03
6232 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6233 TLS client Hello record length value would otherwise be > 255 and
6234 less that 512 pad with a dummy extension containing zeroes so it
6235 is at least 512 bytes long.
6238 *Adam Langley, Steve Henson*
6240 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014] ###
6242 * Fix for TLS record tampering bug. A carefully crafted invalid
6243 handshake could crash OpenSSL with a NULL pointer exception.
6244 Thanks to Anton Johansson for reporting this issues.
6247 * Keep original DTLS digest and encryption contexts in retransmission
6248 structures so we can use the previous session parameters if they need
6249 to be resent. [CVE-2013-6450][]
6253 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6254 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6255 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6256 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6257 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6258 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6260 *Rob Stradling, Adam Langley*
6262 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013] ###
6264 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6265 supporting platforms or when small records were transferred.
6267 *Andy Polyakov, Steve Henson*
6269 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013] ###
6271 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6273 This addresses the flaw in CBC record processing discovered by
6274 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6275 at: http://www.isg.rhul.ac.uk/tls/
6277 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6278 Security Group at Royal Holloway, University of London
6279 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6280 Emilia Käsper for the initial patch.
6283 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6285 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6286 ciphersuites which can be exploited in a denial of service attack.
6287 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6288 and detecting this bug and to Wolfgang Ettlinger
6289 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6294 * Return an error when checking OCSP signatures when key is NULL.
6295 This fixes a DoS attack. [CVE-2013-0166][]
6299 * Make openssl verify return errors.
6301 *Chris Palmer <palmer@google.com> and Ben Laurie*
6303 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6304 the right response is stapled. Also change SSL_get_certificate()
6305 so it returns the certificate actually sent.
6306 See http://rt.openssl.org/Ticket/Display.html?id=2836.
6308 *Rob Stradling <rob.stradling@comodo.com>*
6310 * Fix possible deadlock when decoding public keys.
6314 * Don't use TLS 1.0 record version number in initial client hello
6319 ### Changes between 1.0.1b and 1.0.1c [10 May 2012] ###
6321 * Sanity check record length before skipping explicit IV in TLS
6322 1.2, 1.1 and DTLS to fix DoS attack.
6324 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6325 fuzzing as a service testing platform.
6330 * Initialise tkeylen properly when encrypting CMS messages.
6331 Thanks to Solar Designer of Openwall for reporting this issue.
6335 * In FIPS mode don't try to use composite ciphers as they are not
6340 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012] ###
6342 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6343 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6344 mean any application compiled against OpenSSL 1.0.0 headers setting
6345 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6346 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6347 0x10000000L Any application which was previously compiled against
6348 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6349 will need to be recompiled as a result. Letting be results in
6350 inability to disable specifically TLS 1.1 and in client context,
6351 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6355 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6356 disable just protocol X, but all protocols above X *if* there are
6357 protocols *below* X still enabled. In more practical terms it means
6358 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6359 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6360 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6365 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012] ###
6367 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6368 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6369 in CRYPTO_realloc_clean.
6371 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6372 issue and to Adam Langley <agl@chromium.org> for fixing it.
6375 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6377 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6381 * Workarounds for some broken servers that "hang" if a client hello
6382 record length exceeds 255 bytes.
6384 1. Do not use record version number > TLS 1.0 in initial client
6385 hello: some (but not all) hanging servers will now work.
6386 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6387 the number of ciphers sent in the client hello. This should be
6388 set to an even number, such as 50, for example by passing:
6389 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6390 Most broken servers should now work.
6391 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6392 TLS 1.2 client support entirely.
6396 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6400 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012] ###
6402 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6403 STRING form instead of a DigestInfo.
6407 * The format used for MDC2 RSA signatures is inconsistent between EVP
6408 and the RSA_sign/RSA_verify functions. This was made more apparent when
6409 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6410 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6411 the correct format in RSA_verify so both forms transparently work.
6415 * Some servers which support TLS 1.0 can choke if we initially indicate
6416 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6417 encrypted premaster secret. As a workaround use the maximum permitted
6418 client version in client hello, this should keep such servers happy
6419 and still work with previous versions of OpenSSL.
6423 * Add support for TLS/DTLS heartbeats.
6425 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6427 * Add support for SCTP.
6429 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6431 * Improved PRNG seeding for VOS.
6433 *Paul Green <Paul.Green@stratus.com>*
6435 * Extensive assembler packs updates, most notably:
6437 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6438 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6439 - x86_64: bit-sliced AES implementation;
6440 - ARM: NEON support, contemporary platforms optimizations;
6441 - s390x: z196 support;
6442 - *: GHASH and GF(2^m) multiplication implementations;
6446 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6447 (removal of unnecessary code)
6449 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6451 * Add TLS key material exporter from RFC 5705.
6455 * Add DTLS-SRTP negotiation from RFC 5764.
6459 * Add Next Protocol Negotiation,
6460 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
6461 disabled with a no-npn flag to config or Configure. Code donated
6464 *Adam Langley <agl@google.com> and Ben Laurie*
6466 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6467 NIST-P256, NIST-P521, with constant-time single point multiplication on
6468 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6469 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6470 Code made available under Apache License version 2.0.
6472 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6473 line to include this in your build of OpenSSL, and run "make depend" (or
6474 "make update"). This enables the following EC_METHODs:
6476 EC_GFp_nistp224_method()
6477 EC_GFp_nistp256_method()
6478 EC_GFp_nistp521_method()
6480 EC_GROUP_new_by_curve_name() will automatically use these (while
6481 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6484 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6486 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6487 all platforms. Move ssize_t definition from e_os.h to the public
6488 header file e_os2.h as it now appears in public header file cms.h
6492 * New -sigopt option to the ca, req and x509 utilities. Additional
6493 signature parameters can be passed using this option and in
6498 * Add RSA PSS signing function. This will generate and set the
6499 appropriate AlgorithmIdentifiers for PSS based on those in the
6500 corresponding EVP_MD_CTX structure. No application support yet.
6504 * Support for companion algorithm specific ASN1 signing routines.
6505 New function ASN1_item_sign_ctx() signs a pre-initialised
6506 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6507 the appropriate parameters.
6511 * Add new algorithm specific ASN1 verification initialisation function
6512 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6513 handling will be the same no matter what EVP_PKEY_METHOD is used.
6514 Add a PSS handler to support verification of PSS signatures: checked
6515 against a number of sample certificates.
6519 * Add signature printing for PSS. Add PSS OIDs.
6521 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6523 * Add algorithm specific signature printing. An individual ASN1 method
6524 can now print out signatures instead of the standard hex dump.
6526 More complex signatures (e.g. PSS) can print out more meaningful
6527 information. Include DSA version that prints out the signature
6532 * Password based recipient info support for CMS library: implementing
6537 * Split password based encryption into PBES2 and PBKDF2 functions. This
6538 neatly separates the code into cipher and PBE sections and is required
6539 for some algorithms that split PBES2 into separate pieces (such as
6540 password based CMS).
6544 * Session-handling fixes:
6545 - Fix handling of connections that are resuming with a session ID,
6546 but also support Session Tickets.
6547 - Fix a bug that suppressed issuing of a new ticket if the client
6548 presented a ticket with an expired session.
6549 - Try to set the ticket lifetime hint to something reasonable.
6550 - Make tickets shorter by excluding irrelevant information.
6551 - On the client side, don't ignore renewed tickets.
6553 *Adam Langley, Bodo Moeller (Google)*
6555 * Fix PSK session representation.
6559 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6561 This work was sponsored by Intel.
6565 * Add GCM support to TLS library. Some custom code is needed to split
6566 the IV between the fixed (from PRF) and explicit (from TLS record)
6567 portions. This adds all GCM ciphersuites supported by RFC5288 and
6568 RFC5289. Generalise some AES* cipherstrings to include GCM and
6569 add a special AESGCM string for GCM only.
6573 * Expand range of ctrls for AES GCM. Permit setting invocation
6574 field on decrypt and retrieval of invocation field only on encrypt.
6578 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6579 As required by RFC5289 these ciphersuites cannot be used if for
6580 versions of TLS earlier than 1.2.
6584 * For FIPS capable OpenSSL interpret a NULL default public key method
6585 as unset and return the appropriate default but do *not* set the default.
6586 This means we can return the appropriate method in applications that
6587 switch between FIPS and non-FIPS modes.
6591 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6592 ENGINE is used then we cannot handle that in the FIPS module so we
6593 keep original code iff non-FIPS operations are allowed.
6597 * Add -attime option to openssl utilities.
6599 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6601 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6605 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6606 FIPS EC methods unconditionally for now.
6610 * New build option no-ec2m to disable characteristic 2 code.
6614 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6615 all cases can be covered as some introduce binary incompatibilities.
6619 * Redirect RSA operations to FIPS module including keygen,
6620 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6624 * Add similar low level API blocking to ciphers.
6628 * Low level digest APIs are not approved in FIPS mode: any attempt
6629 to use these will cause a fatal error. Applications that *really* want
6630 to use them can use the private_* version instead.
6634 * Redirect cipher operations to FIPS module for FIPS builds.
6638 * Redirect digest operations to FIPS module for FIPS builds.
6642 * Update build system to add "fips" flag which will link in fipscanister.o
6643 for static and shared library builds embedding a signature if needed.
6647 * Output TLS supported curves in preference order instead of numerical
6648 order. This is currently hardcoded for the highest order curves first.
6649 This should be configurable so applications can judge speed vs strength.
6653 * Add TLS v1.2 server support for client authentication.
6657 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6662 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6663 FIPS modules versions.
6667 * Add TLS v1.2 client side support for client authentication. Keep cache
6668 of handshake records longer as we don't know the hash algorithm to use
6669 until after the certificate request message is received.
6673 * Initial TLS v1.2 client support. Add a default signature algorithms
6674 extension including all the algorithms we support. Parse new signature
6675 format in client key exchange. Relax some ECC signing restrictions for
6676 TLS v1.2 as indicated in RFC5246.
6680 * Add server support for TLS v1.2 signature algorithms extension. Switch
6681 to new signature format when needed using client digest preference.
6682 All server ciphersuites should now work correctly in TLS v1.2. No client
6683 support yet and no support for client certificates.
6687 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6688 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6689 ciphersuites. At present only RSA key exchange ciphersuites work with
6690 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6691 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6692 and version checking.
6696 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6697 with this defined it will not be affected by any changes to ssl internal
6698 structures. Add several utility functions to allow openssl application
6699 to work with OPENSSL_NO_SSL_INTERN defined.
6703 * A long standing patch to add support for SRP from EdelWeb (Peter
6704 Sylvester and Christophe Renou) was integrated.
6705 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6706 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6709 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6713 * Permit abbreviated handshakes when renegotiating using the function
6714 SSL_renegotiate_abbreviated().
6716 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6718 * Add call to ENGINE_register_all_complete() to
6719 ENGINE_load_builtin_engines(), so some implementations get used
6720 automatically instead of needing explicit application support.
6724 * Add support for TLS key exporter as described in RFC5705.
6726 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6728 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6729 a few changes are required:
6731 Add SSL_OP_NO_TLSv1_1 flag.
6732 Add TLSv1_1 methods.
6733 Update version checking logic to handle version 1.1.
6734 Add explicit IV handling (ported from DTLS code).
6735 Add command line options to s_client/s_server.
6742 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015] ###
6744 * X509_ATTRIBUTE memory leak
6746 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6747 memory. This structure is used by the PKCS#7 and CMS routines so any
6748 application which reads PKCS#7 or CMS data from untrusted sources is
6749 affected. SSL/TLS is not affected.
6751 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6757 * Race condition handling PSK identify hint
6759 If PSK identity hints are received by a multi-threaded client then
6760 the values are wrongly updated in the parent SSL_CTX structure. This can
6761 result in a race condition potentially leading to a double free of the
6767 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015] ###
6769 * Malformed ECParameters causes infinite loop
6771 When processing an ECParameters structure OpenSSL enters an infinite loop
6772 if the curve specified is over a specially malformed binary polynomial
6775 This can be used to perform denial of service against any
6776 system which processes public keys, certificate requests or
6777 certificates. This includes TLS clients and TLS servers with
6778 client authentication enabled.
6780 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6785 * Exploitable out-of-bounds read in X509_cmp_time
6787 X509_cmp_time does not properly check the length of the ASN1_TIME
6788 string and can read a few bytes out of bounds. In addition,
6789 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6792 An attacker can use this to craft malformed certificates and CRLs of
6793 various sizes and potentially cause a segmentation fault, resulting in
6794 a DoS on applications that verify certificates or CRLs. TLS clients
6795 that verify CRLs are affected. TLS clients and servers with client
6796 authentication enabled may be affected if they use custom verification
6799 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6800 independently by Hanno Böck.
6805 * PKCS7 crash with missing EnvelopedContent
6807 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6808 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6809 with missing content and trigger a NULL pointer dereference on parsing.
6811 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6812 structures from untrusted sources are affected. OpenSSL clients and
6813 servers are not affected.
6815 This issue was reported to OpenSSL by Michal Zalewski (Google).
6820 * CMS verify infinite loop with unknown hash function
6822 When verifying a signedData message the CMS code can enter an infinite loop
6823 if presented with an unknown hash function OID. This can be used to perform
6824 denial of service against any system which verifies signedData messages using
6826 This issue was reported to OpenSSL by Johannes Bauer.
6831 * Race condition handling NewSessionTicket
6833 If a NewSessionTicket is received by a multi-threaded client when attempting to
6834 reuse a previous ticket then a race condition can occur potentially leading to
6835 a double free of the ticket data.
6840 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015] ###
6842 * Segmentation fault in ASN1_TYPE_cmp fix
6844 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6845 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6846 certificate signature algorithm consistency this can be used to crash any
6847 certificate verification operation and exploited in a DoS attack. Any
6848 application which performs certificate verification is vulnerable including
6849 OpenSSL clients and servers which enable client authentication.
6854 * ASN.1 structure reuse memory corruption fix
6856 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6857 memory corruption via an invalid write. Such reuse is and has been
6858 strongly discouraged and is believed to be rare.
6860 Applications that parse structures containing CHOICE or ANY DEFINED BY
6861 components may be affected. Certificate parsing (d2i_X509 and related
6862 functions) are however not affected. OpenSSL clients and servers are
6868 * PKCS7 NULL pointer dereferences fix
6870 The PKCS#7 parsing code does not handle missing outer ContentInfo
6871 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6872 missing content and trigger a NULL pointer dereference on parsing.
6874 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6875 otherwise parse PKCS#7 structures from untrusted sources are
6876 affected. OpenSSL clients and servers are not affected.
6878 This issue was reported to OpenSSL by Michal Zalewski (Google).
6883 * DoS via reachable assert in SSLv2 servers fix
6885 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6886 servers that both support SSLv2 and enable export cipher suites by sending
6887 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6889 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6890 (OpenSSL development team).
6895 * Use After Free following d2i_ECPrivatekey error fix
6897 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6898 could cause a use after free condition. This, in turn, could cause a double
6899 free in several private key parsing functions (such as d2i_PrivateKey
6900 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6901 for applications that receive EC private keys from untrusted
6902 sources. This scenario is considered rare.
6904 This issue was discovered by the BoringSSL project and fixed in their
6910 * X509_to_X509_REQ NULL pointer deref fix
6912 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6913 the certificate key is invalid. This function is rarely used in practice.
6915 This issue was discovered by Brian Carpenter.
6920 * Removed the export ciphers from the DEFAULT ciphers
6924 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015] ###
6926 * Build fixes for the Windows and OpenVMS platforms
6928 *Matt Caswell and Richard Levitte*
6930 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015] ###
6932 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6933 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6934 dereference. This could lead to a Denial Of Service attack. Thanks to
6935 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6940 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6941 dtls1_buffer_record function under certain conditions. In particular this
6942 could occur if an attacker sent repeated DTLS records with the same
6943 sequence number but for the next epoch. The memory leak could be exploited
6944 by an attacker in a Denial of Service attack through memory exhaustion.
6945 Thanks to Chris Mueller for reporting this issue.
6950 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6951 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6952 method would be set to NULL which could later result in a NULL pointer
6953 dereference. Thanks to Frank Schmirler for reporting this issue.
6958 * Abort handshake if server key exchange message is omitted for ephemeral
6961 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6962 reporting this issue.
6967 * Remove non-export ephemeral RSA code on client and server. This code
6968 violated the TLS standard by allowing the use of temporary RSA keys in
6969 non-export ciphersuites and could be used by a server to effectively
6970 downgrade the RSA key length used to a value smaller than the server
6971 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6972 INRIA or reporting this issue.
6977 * Fixed issue where DH client certificates are accepted without verification.
6978 An OpenSSL server will accept a DH certificate for client authentication
6979 without the certificate verify message. This effectively allows a client to
6980 authenticate without the use of a private key. This only affects servers
6981 which trust a client certificate authority which issues certificates
6982 containing DH keys: these are extremely rare and hardly ever encountered.
6983 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6989 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6990 results on some platforms, including x86_64. This bug occurs at random
6991 with a very low probability, and is not known to be exploitable in any
6992 way, though its exact impact is difficult to determine. Thanks to Pieter
6993 Wuille (Blockstream) who reported this issue and also suggested an initial
6994 fix. Further analysis was conducted by the OpenSSL development team and
6995 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6996 the OpenSSL core team.
7001 * Fix various certificate fingerprint issues.
7003 By using non-DER or invalid encodings outside the signed portion of a
7004 certificate the fingerprint can be changed without breaking the signature.
7005 Although no details of the signed portion of the certificate can be changed
7006 this can cause problems with some applications: e.g. those using the
7007 certificate fingerprint for blacklists.
7009 1. Reject signatures with non zero unused bits.
7011 If the BIT STRING containing the signature has non zero unused bits reject
7012 the signature. All current signature algorithms require zero unused bits.
7014 2. Check certificate algorithm consistency.
7016 Check the AlgorithmIdentifier inside TBS matches the one in the
7017 certificate signature. NB: this will result in signature failure
7018 errors for some broken certificates.
7020 Thanks to Konrad Kraszewski from Google for reporting this issue.
7022 3. Check DSA/ECDSA signatures use DER.
7024 Reencode DSA/ECDSA signatures and compare with the original received
7025 signature. Return an error if there is a mismatch.
7027 This will reject various cases including garbage after signature
7028 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7029 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7030 (negative or with leading zeroes).
7032 Further analysis was conducted and fixes were developed by Stephen Henson
7033 of the OpenSSL core team.
7039 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014] ###
7041 * Session Ticket Memory Leak.
7043 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7044 integrity of that ticket is first verified. In the event of a session
7045 ticket integrity check failing, OpenSSL will fail to free memory
7046 causing a memory leak. By sending a large number of invalid session
7047 tickets an attacker could exploit this issue in a Denial Of Service
7053 * Build option no-ssl3 is incomplete.
7055 When OpenSSL is configured with "no-ssl3" as a build option, servers
7056 could accept and complete a SSL 3.0 handshake, and clients could be
7057 configured to send them.
7060 *Akamai and the OpenSSL team*
7062 * Add support for TLS_FALLBACK_SCSV.
7063 Client applications doing fallback retries should call
7064 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7067 *Adam Langley, Bodo Moeller*
7069 * Add additional DigestInfo checks.
7071 Reencode DigestInto in DER and check against the original when
7072 verifying RSA signature: this will reject any improperly encoded
7073 DigestInfo structures.
7075 Note: this is a precautionary measure and no attacks are currently known.
7080 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014] ###
7082 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7083 to a denial of service attack. A malicious server can crash the client
7084 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7085 ciphersuite and sending carefully crafted handshake messages.
7087 Thanks to Felix Gröbert (Google) for discovering and researching this
7093 * By sending carefully crafted DTLS packets an attacker could cause openssl
7094 to leak memory. This can be exploited through a Denial of Service attack.
7095 Thanks to Adam Langley for discovering and researching this issue.
7100 * An attacker can force openssl to consume large amounts of memory whilst
7101 processing DTLS handshake messages. This can be exploited through a
7102 Denial of Service attack.
7103 Thanks to Adam Langley for discovering and researching this issue.
7108 * An attacker can force an error condition which causes openssl to crash
7109 whilst processing DTLS packets due to memory being freed twice. This
7110 can be exploited through a Denial of Service attack.
7111 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7117 * If a multithreaded client connects to a malicious server using a resumed
7118 session and the server sends an ec point format extension it could write
7119 up to 255 bytes to freed memory.
7121 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7127 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7128 X509_name_oneline, X509_name_print_ex et al. to leak some information
7129 from the stack. Applications may be affected if they echo pretty printing
7130 output to the attacker.
7132 Thanks to Ivan Fratric (Google) for discovering this issue.
7135 *Emilia Käsper, and Steve Henson*
7137 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7138 for corner cases. (Certain input points at infinity could lead to
7139 bogus results, with non-infinity inputs mapped to infinity too.)
7143 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014] ###
7145 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7146 handshake can force the use of weak keying material in OpenSSL
7147 SSL/TLS clients and servers.
7149 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7150 researching this issue. [CVE-2014-0224][]
7152 *KIKUCHI Masashi, Steve Henson*
7154 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7155 OpenSSL DTLS client the code can be made to recurse eventually crashing
7158 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7161 *Imre Rad, Steve Henson*
7163 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7164 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7165 client or server. This is potentially exploitable to run arbitrary
7166 code on a vulnerable client or server.
7168 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7170 *Jüri Aedla, Steve Henson*
7172 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7173 are subject to a denial of service attack.
7175 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7176 this issue. [CVE-2014-3470][]
7178 *Felix Gröbert, Ivan Fratric, Steve Henson*
7180 * Harmonize version and its documentation. -f flag is used to display
7183 *mancha <mancha1@zoho.com>*
7185 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7186 in i2d_ECPrivateKey.
7188 *mancha <mancha1@zoho.com>*
7190 * Fix some double frees. These are not thought to be exploitable.
7192 *mancha <mancha1@zoho.com>*
7194 * Fix for the attack described in the paper "Recovering OpenSSL
7195 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7196 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7197 http://eprint.iacr.org/2014/140
7199 Thanks to Yuval Yarom and Naomi Benger for discovering this
7200 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7202 *Yuval Yarom and Naomi Benger*
7204 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014] ###
7206 * Keep original DTLS digest and encryption contexts in retransmission
7207 structures so we can use the previous session parameters if they need
7208 to be resent. [CVE-2013-6450][]
7212 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7213 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7214 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7215 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7216 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7217 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7219 *Rob Stradling, Adam Langley*
7221 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013] ###
7223 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7225 This addresses the flaw in CBC record processing discovered by
7226 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7227 at: http://www.isg.rhul.ac.uk/tls/
7229 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7230 Security Group at Royal Holloway, University of London
7231 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7232 Emilia Käsper for the initial patch.
7235 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7237 * Return an error when checking OCSP signatures when key is NULL.
7238 This fixes a DoS attack. [CVE-2013-0166][]
7242 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7243 the right response is stapled. Also change SSL_get_certificate()
7244 so it returns the certificate actually sent.
7245 See http://rt.openssl.org/Ticket/Display.html?id=2836.
7246 (This is a backport)
7248 *Rob Stradling <rob.stradling@comodo.com>*
7250 * Fix possible deadlock when decoding public keys.
7254 ### Changes between 1.0.0i and 1.0.0j [10 May 2012] ###
7256 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7259 * Sanity check record length before skipping explicit IV in DTLS
7262 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7263 fuzzing as a service testing platform.
7268 * Initialise tkeylen properly when encrypting CMS messages.
7269 Thanks to Solar Designer of Openwall for reporting this issue.
7273 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012] ###
7275 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7276 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7277 in CRYPTO_realloc_clean.
7279 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7280 issue and to Adam Langley <agl@chromium.org> for fixing it.
7283 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7285 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012] ###
7287 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7288 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7289 content decryption and always return the same error. Note: this attack
7290 needs on average 2^20 messages so it only affects automated senders. The
7291 old behaviour can be re-enabled in the CMS code by setting the
7292 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7293 an MMA defence is not necessary.
7294 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7295 this issue. [CVE-2012-0884][]
7299 * Fix CVE-2011-4619: make sure we really are receiving a
7300 client hello before rejecting multiple SGC restarts. Thanks to
7301 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7305 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012] ###
7307 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7308 Thanks to Antonio Martin, Enterprise Secure Access Research and
7309 Development, Cisco Systems, Inc. for discovering this bug and
7310 preparing a fix. [CVE-2012-0050][]
7314 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012] ###
7316 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7317 of the Vaudenay padding oracle attack on CBC mode encryption
7318 which enables an efficient plaintext recovery attack against
7319 the OpenSSL implementation of DTLS. Their attack exploits timing
7320 differences arising during decryption processing. A research
7321 paper describing this attack can be found at:
7322 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
7323 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7324 Security Group at Royal Holloway, University of London
7325 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7326 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7327 for preparing the fix. [CVE-2011-4108][]
7329 *Robin Seggelmann, Michael Tuexen*
7331 * Clear bytes used for block padding of SSL 3.0 records.
7334 *Adam Langley (Google)*
7336 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7337 Kadianakis <desnacked@gmail.com> for discovering this issue and
7338 Adam Langley for preparing the fix. [CVE-2011-4619][]
7340 *Adam Langley (Google)*
7342 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7344 *Andrey Kulikov <amdeich@gmail.com>*
7346 * Prevent malformed RFC3779 data triggering an assertion failure.
7347 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7348 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7350 *Rob Austein <sra@hactrn.net>*
7352 * Improved PRNG seeding for VOS.
7354 *Paul Green <Paul.Green@stratus.com>*
7356 * Fix ssl_ciph.c set-up race.
7358 *Adam Langley (Google)*
7360 * Fix spurious failures in ecdsatest.c.
7362 *Emilia Käsper (Google)*
7364 * Fix the BIO_f_buffer() implementation (which was mixing different
7365 interpretations of the '..._len' fields).
7367 *Adam Langley (Google)*
7369 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7370 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7371 threads won't reuse the same blinding coefficients.
7373 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7374 lock to call BN_BLINDING_invert_ex, and avoids one use of
7375 BN_BLINDING_update for each BN_BLINDING structure (previously,
7376 the last update always remained unused).
7378 *Emilia Käsper (Google)*
7380 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7382 *Bob Buckholz (Google)*
7384 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011] ###
7386 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7387 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7389 *Kaspar Brand <ossl@velox.ch>*
7391 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7392 for multi-threaded use of ECDH. [CVE-2011-3210][]
7394 *Adam Langley (Google)*
7396 * Fix x509_name_ex_d2i memory leak on bad inputs.
7400 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7401 signature public key algorithm by using OID xref utilities instead.
7402 Before this you could only use some ECC ciphersuites with SHA1 only.
7406 * Add protection against ECDSA timing attacks as mentioned in the paper
7407 by Billy Bob Brumley and Nicola Tuveri, see:
7409 http://eprint.iacr.org/2011/232.pdf
7412 *Billy Bob Brumley and Nicola Tuveri*
7414 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011] ###
7416 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7418 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7420 * Fix bug in string printing code: if *any* escaping is enabled we must
7421 escape the escape character (backslash) or the resulting string is
7426 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010] ###
7428 * Disable code workaround for ancient and obsolete Netscape browsers
7429 and servers: an attacker can use it in a ciphersuite downgrade attack.
7430 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7434 * Fixed J-PAKE implementation error, originally discovered by
7435 Sebastien Martini, further info and confirmation from Stefan
7436 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7440 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010] ###
7442 * Fix extension code to avoid race conditions which can result in a buffer
7443 overrun vulnerability: resumed sessions must not be modified as they can
7444 be shared by multiple threads. CVE-2010-3864
7448 * Fix WIN32 build system to correctly link an ENGINE directory into
7453 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010] ###
7455 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7458 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7460 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010] ###
7462 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7463 context. The operation can be customised via the ctrl mechanism in
7464 case ENGINEs want to include additional functionality.
7468 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7472 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7473 output hashes compatible with older versions of OpenSSL.
7475 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7477 * Fix compression algorithm handling: if resuming a session use the
7478 compression algorithm of the resumed session instead of determining
7479 it from client hello again. Don't allow server to change algorithm.
7483 * Add load_crls() function to apps tidying load_certs() too. Add option
7484 to verify utility to allow additional CRLs to be included.
7488 * Update OCSP request code to permit adding custom headers to the request:
7489 some responders need this.
7493 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7496 *Julia Lawall <julia@diku.dk>*
7498 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7499 needlessly dereferenced structures, used obsolete functions and
7500 didn't handle all updated verify codes correctly.
7504 * Disable MD2 in the default configuration.
7508 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7509 indicate the initial BIO being pushed or popped. This makes it possible
7510 to determine whether the BIO is the one explicitly called or as a result
7511 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7512 it handles reference counts correctly and doesn't zero out the I/O bio
7513 when it is not being explicitly popped. WARNING: applications which
7514 included workarounds for the old buggy behaviour will need to be modified
7515 or they could free up already freed BIOs.
7519 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7520 renaming to all platforms (within the 0.9.8 branch, this was
7521 done conditionally on Netware platforms to avoid a name clash).
7523 *Guenter <lists@gknw.net>*
7525 * Add ECDHE and PSK support to DTLS.
7527 *Michael Tuexen <tuexen@fh-muenster.de>*
7529 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7534 * Add "missing" function EVP_MD_flags() (without this the only way to
7535 retrieve a digest flags is by accessing the structure directly. Update
7536 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
7537 or cipher is registered as in the "from" argument. Print out all
7538 registered digests in the dgst usage message instead of manually
7539 attempting to work them out.
7543 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7544 this allows the use of compression and extensions. Change default cipher
7545 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7546 by default unless an application cipher string requests it.
7550 * Alter match criteria in PKCS12_parse(). It used to try to use local
7551 key ids to find matching certificates and keys but some PKCS#12 files
7552 don't follow the (somewhat unwritten) rules and this strategy fails.
7553 Now just gather all certificates together and the first private key
7554 then look for the first certificate that matches the key.
7558 * Support use of registered digest and cipher names for dgst and cipher
7559 commands instead of having to add each one as a special case. So now
7566 openssl dgst -sha256 foo
7568 and this works for ENGINE based algorithms too.
7573 * Update Gost ENGINE to support parameter files.
7575 *Victor B. Wagner <vitus@cryptocom.ru>*
7577 * Support GeneralizedTime in ca utility.
7579 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7581 * Enhance the hash format used for certificate directory links. The new
7582 form uses the canonical encoding (meaning equivalent names will work
7583 even if they aren't identical) and uses SHA1 instead of MD5. This form
7584 is incompatible with the older format and as a result c_rehash should
7585 be used to rebuild symbolic links.
7589 * Make PKCS#8 the default write format for private keys, replacing the
7590 traditional format. This form is standardised, more secure and doesn't
7591 include an implicit MD5 dependency.
7595 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7596 committed to OpenSSL should pass this lot as a minimum.
7600 * Add session ticket override functionality for use by EAP-FAST.
7602 *Jouni Malinen <j@w1.fi>*
7604 * Modify HMAC functions to return a value. Since these can be implemented
7605 in an ENGINE errors can occur.
7609 * Type-checked OBJ_bsearch_ex.
7613 * Type-checked OBJ_bsearch. Also some constification necessitated
7614 by type-checking. Still to come: TXT_DB, bsearch(?),
7615 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7620 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7621 seconds to a tm structure directly, instead of going through OS
7622 specific date routines. This avoids any issues with OS routines such
7623 as the year 2038 bug. New *_adj() functions for ASN1 time structures
7624 and X509_time_adj_ex() to cover the extended range. The existing
7625 X509_time_adj() is still usable and will no longer have any date issues.
7629 * Delta CRL support. New use deltas option which will attempt to locate
7630 and search any appropriate delta CRLs available.
7632 This work was sponsored by Google.
7636 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7637 code and add additional score elements. Validate alternate CRL paths
7638 as part of the CRL checking and indicate a new error "CRL path validation
7639 error" in this case. Applications wanting additional details can use
7640 the verify callback and check the new "parent" field. If this is not
7641 NULL CRL path validation is taking place. Existing applications won't
7642 see this because it requires extended CRL support which is off by
7645 This work was sponsored by Google.
7649 * Support for freshest CRL extension.
7651 This work was sponsored by Google.
7655 * Initial indirect CRL support. Currently only supported in the CRLs
7656 passed directly and not via lookup. Process certificate issuer
7657 CRL entry extension and lookup CRL entries by bother issuer name
7658 and serial number. Check and process CRL issuer entry in IDP extension.
7660 This work was sponsored by Google.
7664 * Add support for distinct certificate and CRL paths. The CRL issuer
7665 certificate is validated separately in this case. Only enabled if
7666 an extended CRL support flag is set: this flag will enable additional
7667 CRL functionality in future.
7669 This work was sponsored by Google.
7673 * Add support for policy mappings extension.
7675 This work was sponsored by Google.
7679 * Fixes to pathlength constraint, self issued certificate handling,
7680 policy processing to align with RFC3280 and PKITS tests.
7682 This work was sponsored by Google.
7686 * Support for name constraints certificate extension. DN, email, DNS
7687 and URI types are currently supported.
7689 This work was sponsored by Google.
7693 * To cater for systems that provide a pointer-based thread ID rather
7694 than numeric, deprecate the current numeric thread ID mechanism and
7695 replace it with a structure and associated callback type. This
7696 mechanism allows a numeric "hash" to be extracted from a thread ID in
7697 either case, and on platforms where pointers are larger than 'long',
7698 mixing is done to help ensure the numeric 'hash' is usable even if it
7699 can't be guaranteed unique. The default mechanism is to use "&errno"
7700 as a pointer-based thread ID to distinguish between threads.
7702 Applications that want to provide their own thread IDs should now use
7703 CRYPTO_THREADID_set_callback() to register a callback that will call
7704 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7706 Note that ERR_remove_state() is now deprecated, because it is tied
7707 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7708 to free the current thread's error state should be replaced by
7709 ERR_remove_thread_state(NULL).
7711 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7712 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7713 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7714 application was previously providing a numeric thread callback that
7715 was inappropriate for distinguishing threads, then uniqueness might
7716 have been obtained with &errno that happened immediately in the
7717 intermediate development versions of OpenSSL; this is no longer the
7718 case, the numeric thread callback will now override the automatic use
7721 *Geoff Thorpe, with help from Bodo Moeller*
7723 * Initial support for different CRL issuing certificates. This covers a
7724 simple case where the self issued certificates in the chain exist and
7725 the real CRL issuer is higher in the existing chain.
7727 This work was sponsored by Google.
7731 * Removed effectively defunct crypto/store from the build.
7735 * Revamp of STACK to provide stronger type-checking. Still to come:
7736 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7737 ASN1_STRING, CONF_VALUE.
7741 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7742 RAM on SSL connections. This option can save about 34k per idle SSL.
7746 * Revamp of LHASH to provide stronger type-checking. Still to come:
7747 STACK, TXT_DB, bsearch, qsort.
7751 * Initial support for Cryptographic Message Syntax (aka CMS) based
7752 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7753 support for data, signedData, compressedData, digestedData and
7754 encryptedData, envelopedData types included. Scripts to check against
7755 RFC4134 examples draft and interop and consistency checks of many
7756 content types and variants.
7760 * Add options to enc utility to support use of zlib compression BIO.
7764 * Extend mk1mf to support importing of options and assembly language
7765 files from Configure script, currently only included in VC-WIN32.
7766 The assembly language rules can now optionally generate the source
7767 files from the associated perl scripts.
7771 * Implement remaining functionality needed to support GOST ciphersuites.
7772 Interop testing has been performed using CryptoPro implementations.
7774 *Victor B. Wagner <vitus@cryptocom.ru>*
7776 * s390x assembler pack.
7780 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7785 * Implement Opaque PRF Input TLS extension as specified in
7786 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7787 official specification yet and no extension type assignment by
7788 IANA exists, this extension (for now) will have to be explicitly
7789 enabled when building OpenSSL by providing the extension number
7790 to use. For example, specify an option
7792 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7794 to the "config" or "Configure" script to enable the extension,
7795 assuming extension number 0x9527 (which is a completely arbitrary
7796 and unofficial assignment based on the MD5 hash of the Internet
7797 Draft). Note that by doing so, you potentially lose
7798 interoperability with other TLS implementations since these might
7799 be using the same extension number for other purposes.
7801 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7802 opaque PRF input value to use in the handshake. This will create
7803 an internal copy of the length-'len' string at 'src', and will
7804 return non-zero for success.
7806 To get more control and flexibility, provide a callback function
7809 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7810 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7814 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7817 Callback function 'cb' will be called in handshakes, and is
7818 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7819 Argument 'arg' is for application purposes (the value as given to
7820 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7821 be provided to the callback function). The callback function
7822 has to return non-zero to report success: usually 1 to use opaque
7823 PRF input just if possible, or 2 to enforce use of the opaque PRF
7824 input. In the latter case, the library will abort the handshake
7825 if opaque PRF input is not successfully negotiated.
7827 Arguments 'peerinput' and 'len' given to the callback function
7828 will always be NULL and 0 in the case of a client. A server will
7829 see the client's opaque PRF input through these variables if
7830 available (NULL and 0 otherwise). Note that if the server
7831 provides an opaque PRF input, the length must be the same as the
7832 length of the client's opaque PRF input.
7834 Note that the callback function will only be called when creating
7835 a new session (session resumption can resume whatever was
7836 previously negotiated), and will not be called in SSL 2.0
7837 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7838 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7839 for applications that need to enforce opaque PRF input.
7844 * Update ssl code to support digests other than SHA1+MD5 for handshake
7848 *Victor B. Wagner <vitus@cryptocom.ru>*
7850 * Add RFC4507 support to OpenSSL. This includes the corrections in
7851 RFC4507bis. The encrypted ticket format is an encrypted encoded
7852 SSL_SESSION structure, that way new session features are automatically
7855 If a client application caches session in an SSL_SESSION structure
7856 support is transparent because tickets are now stored in the encoded
7859 The SSL_CTX structure automatically generates keys for ticket
7860 protection in servers so again support should be possible
7861 with no application modification.
7863 If a client or server wishes to disable RFC4507 support then the option
7864 SSL_OP_NO_TICKET can be set.
7866 Add a TLS extension debugging callback to allow the contents of any client
7867 or server extensions to be examined.
7869 This work was sponsored by Google.
7873 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7874 OpenSSL should now compile cleanly on gcc 4.2
7876 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7878 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7879 support including streaming MAC support: this is required for GOST
7880 ciphersuite support.
7882 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7884 * Add option -stream to use PKCS#7 streaming in smime utility. New
7885 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7886 to output in BER and PEM format.
7890 * Experimental support for use of HMAC via EVP_PKEY interface. This
7891 allows HMAC to be handled via the EVP_DigestSign*() interface. The
7892 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7893 ENGINE support for HMAC keys which are unextractable. New -mac and
7894 -macopt options to dgst utility.
7898 * New option -sigopt to dgst utility. Update dgst to use
7899 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
7900 alternative signing parameters such as X9.31 or PSS in the dgst
7905 * Change ssl_cipher_apply_rule(), the internal function that does
7906 the work each time a ciphersuite string requests enabling
7907 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7908 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7909 the order of disabled ciphersuites such that those ciphersuites
7910 that most recently went from enabled to disabled not only stay
7911 in order with respect to each other, but also have higher priority
7912 than other disabled ciphersuites the next time ciphersuites are
7915 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
7916 the same ciphersuites as with "HIGH" alone, but in a specific
7917 order where the PSK ciphersuites come first (since they are the
7918 most recently disabled ciphersuites when "HIGH" is parsed).
7920 Also, change ssl_create_cipher_list() (using this new
7921 functionality) such that between otherwise identical
7922 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
7927 * Change ssl_create_cipher_list() so that it automatically
7928 arranges the ciphersuites in reasonable order before starting
7929 to process the rule string. Thus, the definition for "DEFAULT"
7930 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
7931 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
7932 This makes it much easier to arrive at a reasonable default order
7933 in applications for which anonymous ciphers are OK (meaning
7934 that you can't actually use DEFAULT).
7936 *Bodo Moeller; suggested by Victor Duchovni*
7938 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
7939 processing) into multiple integers instead of setting
7940 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
7941 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
7942 (These masks as well as the individual bit definitions are hidden
7943 away into the non-exported interface ssl/ssl_locl.h, so this
7944 change to the definition of the SSL_CIPHER structure shouldn't
7945 affect applications.) This give us more bits for each of these
7946 categories, so there is no longer a need to coagulate AES128 and
7947 AES256 into a single algorithm bit, and to coagulate Camellia128
7948 and Camellia256 into a single algorithm bit, which has led to all
7951 Thus, among other things, the kludge introduced in 0.9.7m and
7952 0.9.8e for masking out AES256 independently of AES128 or masking
7953 out Camellia256 independently of AES256 is not needed here in 0.9.9.
7955 With the change, we also introduce new ciphersuite aliases that
7956 so far were missing: "AES128", "AES256", "CAMELLIA128", and
7961 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
7962 Use the leftmost N bytes of the signature input if the input is
7963 larger than the prime q (with N being the size in bytes of q).
7967 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
7968 it yet and it is largely untested.
7972 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
7976 * Initial incomplete changes to avoid need for function casts in OpenSSL
7977 some compilers (gcc 4.2 and later) reject their use. Safestack is
7978 reimplemented. Update ASN1 to avoid use of legacy functions.
7982 * Win32/64 targets are linked with Winsock2.
7986 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
7987 to external functions. This can be used to increase CRL handling
7988 efficiency especially when CRLs are very large by (for example) storing
7989 the CRL revoked certificates in a database.
7993 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
7994 new CRLs added to a directory can be used. New command line option
7995 -verify_return_error to s_client and s_server. This causes real errors
7996 to be returned by the verify callback instead of carrying on no matter
7997 what. This reflects the way a "real world" verify callback would behave.
8001 * GOST engine, supporting several GOST algorithms and public key formats.
8002 Kindly donated by Cryptocom.
8006 * Partial support for Issuing Distribution Point CRL extension. CRLs
8007 partitioned by DP are handled but no indirect CRL or reason partitioning
8008 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8009 selected via a scoring technique which handles IDP and AKID in CRLs.
8013 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8014 will ultimately be used for all verify operations: this will remove the
8015 X509_STORE dependency on certificate verification and allow alternative
8016 lookup methods. X509_STORE based implementations of these two callbacks.
8020 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8021 Modify get_crl() to find a valid (unexpired) CRL if possible.
8025 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8026 this would be called X509_CRL_cmp() but that name is already used by
8027 a function that just compares CRL issuer names. Cache several CRL
8028 extensions in X509_CRL structure and cache CRLDP in X509.
8032 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8033 this maps equivalent X509_NAME structures into a consistent structure.
8034 Name comparison can then be performed rapidly using memcmp().
8038 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8043 * Allow digests to supply their own micalg string for S/MIME type using
8044 the ctrl EVP_MD_CTRL_MICALG.
8048 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8049 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8050 ctrl. It can then customise the structure before and/or after signing
8055 * New function OBJ_add_sigid() to allow application defined signature OIDs
8056 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8057 to free up any added signature OIDs.
8061 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8062 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8063 digest and cipher tables. New options added to openssl utility:
8064 list-message-digest-algorithms and list-cipher-algorithms.
8068 * Change the array representation of binary polynomials: the list
8069 of degrees of non-zero coefficients is now terminated with -1.
8070 Previously it was terminated with 0, which was also part of the
8071 value; thus, the array representation was not applicable to
8072 polynomials where t^0 has coefficient zero. This change makes
8073 the array representation useful in a more general context.
8077 * Various modifications and fixes to SSL/TLS cipher string
8078 handling. For ECC, the code now distinguishes between fixed ECDH
8079 with RSA certificates on the one hand and with ECDSA certificates
8080 on the other hand, since these are separate ciphersuites. The
8081 unused code for Fortezza ciphersuites has been removed.
8083 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8084 (not "ECDHE"). For consistency with the code for DH
8085 certificates, use of ECDH certificates is now considered ECDH
8086 authentication, not RSA or ECDSA authentication (the latter is
8087 merely the CA's signing algorithm and not actively used in the
8090 The temporary ciphersuite alias "ECCdraft" is no longer
8091 available, and ECC ciphersuites are no longer excluded from "ALL"
8092 and "DEFAULT". The following aliases now exist for RFC 4492
8093 ciphersuites, most of these by analogy with the DH case:
8095 kECDHr - ECDH cert, signed with RSA
8096 kECDHe - ECDH cert, signed with ECDSA
8097 kECDH - ECDH cert (signed with either RSA or ECDSA)
8098 kEECDH - ephemeral ECDH
8099 ECDH - ECDH cert or ephemeral ECDH
8105 AECDH - anonymous ECDH
8106 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8111 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8112 Use correct micalg parameters depending on digest(s) in signed message.
8116 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8117 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8121 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8122 an engine to register a method. Add ENGINE lookups for methods and
8123 functional reference processing.
8127 * New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
8128 EVP_{Sign,Verify}* which allow an application to customise the signature
8133 * New -resign option to smime utility. This adds one or more signers
8134 to an existing PKCS#7 signedData structure. Also -md option to use an
8135 alternative message digest algorithm for signing.
8139 * Tidy up PKCS#7 routines and add new functions to make it easier to
8140 create PKCS7 structures containing multiple signers. Update smime
8141 application to support multiple signers.
8145 * New -macalg option to pkcs12 utility to allow setting of an alternative
8150 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8151 Reorganize PBE internals to lookup from a static table using NIDs,
8152 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8153 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8154 PRF which will be automatically used with PBES2.
8158 * Replace the algorithm specific calls to generate keys in "req" with the
8163 * Update PKCS#7 enveloped data routines to use new API. This is now
8164 supported by any public key method supporting the encrypt operation. A
8165 ctrl is added to allow the public key algorithm to examine or modify
8166 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8171 * Add a ctrl to asn1 method to allow a public key algorithm to express
8172 a default digest type to use. In most cases this will be SHA1 but some
8173 algorithms (such as GOST) need to specify an alternative digest. The
8174 return value indicates how strong the preference is 1 means optional and
8175 2 is mandatory (that is it is the only supported type). Modify
8176 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8177 use the default md. Update openssl utilities to use the default digest
8178 type for signing if it is not explicitly indicated.
8182 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8183 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8184 signing method from the key type. This effectively removes the link
8185 between digests and public key types.
8189 * Add an OID cross reference table and utility functions. Its purpose is to
8190 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8191 rsaEncryption. This will allow some of the algorithm specific hackery
8192 needed to use the correct OID to be removed.
8196 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8197 structures for PKCS7_sign(). They are now set up by the relevant public
8202 * Add provisional EC pkey method with support for ECDSA and ECDH.
8206 * Add support for key derivation (agreement) in the API, DH method and
8211 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8212 public and private key formats. As a side effect these add additional
8213 command line functionality not previously available: DSA signatures can be
8214 generated and verified using pkeyutl and DH key support and generation in
8221 *Oliver Tappe <zooey@hirschkaefer.de>*
8223 * New make target "install_html_docs" installs HTML renditions of the
8226 *Oliver Tappe <zooey@hirschkaefer.de>*
8228 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8229 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8230 support key and parameter generation and add initial key generation
8231 functionality for RSA.
8235 * Add functions for main EVP_PKEY_method operations. The undocumented
8236 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
8237 EVP_PKEY_{encrypt,decrypt}_old.
8241 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8242 key API, doesn't do much yet.
8246 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8247 public key algorithms. New option to openssl utility:
8248 "list-public-key-algorithms" to print out info.
8252 * Implement the Supported Elliptic Curves Extension for
8253 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8257 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8258 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8262 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8263 utilities such as rsa, dsa, dsaparam etc except they process any key
8268 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8269 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8270 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8275 * Initial support for pluggable public key ASN1.
8276 De-spaghettify the public key ASN1 handling. Move public and private
8277 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8278 algorithm specific handling to a single module within the relevant
8279 algorithm directory. Add functions to allow (near) opaque processing
8280 of public and private key structures.
8284 * Implement the Supported Point Formats Extension for
8285 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8289 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8290 for the psk identity [hint] and the psk callback functions to the
8291 SSL_SESSION, SSL and SSL_CTX structure.
8294 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8298 SSL_CTX_use_psk_identity_hint
8299 SSL_get_psk_identity_hint
8300 SSL_get_psk_identity
8301 SSL_use_psk_identity_hint
8304 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8306 * Add RFC 3161 compliant time stamp request creation, response generation
8307 and response verification functionality.
8309 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8311 * Add initial support for TLS extensions, specifically for the server_name
8312 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8313 have new members for a host name. The SSL data structure has an
8314 additional member SSL_CTX *initial_ctx so that new sessions can be
8315 stored in that context to allow for session resumption, even after the
8316 SSL has been switched to a new SSL_CTX in reaction to a client's
8317 server_name extension.
8319 New functions (subject to change):
8321 SSL_get_servername()
8322 SSL_get_servername_type()
8325 New CTRL codes and macros (subject to change):
8327 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8328 - SSL_CTX_set_tlsext_servername_callback()
8329 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8330 - SSL_CTX_set_tlsext_servername_arg()
8331 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8333 openssl s_client has a new '-servername ...' option.
8335 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8336 '-key2 ...', '-servername_fatal' (subject to change). This allows
8337 testing the HostName extension for a specific single host name ('-cert'
8338 and '-key' remain fallbacks for handshakes without HostName
8339 negotiation). If the unrecognized_name alert has to be sent, this by
8340 default is a warning; it becomes fatal with the '-servername_fatal'
8344 *Peter Sylvester, Remy Allais, Christophe Renou*
8346 * Whirlpool hash implementation is added.
8350 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8351 bn(64,32). Because of instruction set limitations it doesn't have
8352 any negative impact on performance. This was done mostly in order
8353 to make it possible to share assembler modules, such as bn_mul_mont
8354 implementations, between 32- and 64-bit builds without hassle.
8358 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8359 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8364 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8365 dedicated Montgomery multiplication procedure, is introduced.
8366 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8367 "64-bit" performance on certain 32-bit targets.
8371 * New option SSL_OP_NO_COMP to disable use of compression selectively
8372 in SSL structures. New SSL ctrl to set maximum send fragment size.
8373 Save memory by setting the I/O buffer sizes dynamically instead of
8374 using the maximum available value.
8378 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8379 in addition to the text details.
8383 * Very, very preliminary EXPERIMENTAL support for printing of general
8384 ASN1 structures. This currently produces rather ugly output and doesn't
8385 handle several customised structures at all.
8389 * Integrated support for PVK file format and some related formats such
8390 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8391 these in the 'rsa' and 'dsa' utilities.
8395 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8399 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8400 place for the (very old) "NETSCAPE" format certificates which are now
8401 handled using new ASN1 code equivalents.
8405 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8406 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8407 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8411 * Modify CRL distribution points extension code to print out previously
8412 unsupported fields. Enhance extension setting code to allow setting of
8417 * Add print and set support for Issuing Distribution Point CRL extension.
8421 * Change 'Configure' script to enable Camellia by default.
8428 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010] ###
8430 * When rejecting SSL/TLS records due to an incorrect version number, never
8431 update s->server with a new major version number. As of
8432 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8433 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8434 the previous behavior could result in a read attempt at NULL when
8435 receiving specific incorrect SSL/TLS records once record payload
8436 protection is active. [CVE-2010-0740][]
8438 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8440 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8441 could be crashed if the relevant tables were not present (e.g. chrooted).
8443 *Tomas Hoger <thoger@redhat.com>*
8445 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010] ###
8447 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8449 *Martin Olsson, Neel Mehta*
8451 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8452 accommodate for stack sorting, always a write lock!).
8456 * On some versions of WIN32 Heap32Next is very slow. This can cause
8457 excessive delays in the RAND_poll(): over a minute. As a workaround
8458 include a time check in the inner Heap32Next loop too.
8462 * The code that handled flushing of data in SSL/TLS originally used the
8463 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8464 the problem outlined in PR#1949. The fix suggested there however can
8465 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8466 of Apache). So instead simplify the code to flush unconditionally.
8467 This should be fine since flushing with no data to flush is a no op.
8471 * Handle TLS versions 2.0 and later properly and correctly use the
8472 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8473 off ancient servers have a habit of sticking around for a while...
8477 * Modify compression code so it frees up structures without using the
8478 ex_data callbacks. This works around a problem where some applications
8479 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8480 restarting) then use compression (e.g. SSL with compression) later.
8481 This results in significant per-connection memory leaks and
8482 has caused some security issues including CVE-2008-1678 and
8487 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8488 change when encrypting or decrypting.
8492 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8493 connect and renegotiate with servers which do not support RI.
8494 Until RI is more widely deployed this option is enabled by default.
8498 * Add "missing" ssl ctrls to clear options and mode.
8502 * If client attempts to renegotiate and doesn't support RI respond with
8503 a no_renegotiation alert as required by RFC5746. Some renegotiating
8504 TLS clients will continue a connection gracefully when they receive
8505 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8506 waiting for a server hello which it will never receive. Now we treat a
8507 received no_renegotiation alert as a fatal error. This is because
8508 applications requesting a renegotiation might well expect it to succeed
8509 and would have no code in place to handle the server denying it so the
8510 only safe thing to do is to terminate the connection.
8514 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8515 peer supports secure renegotiation and 0 otherwise. Print out peer
8516 renegotiation support in s_client/s_server.
8520 * Replace the highly broken and deprecated SPKAC certification method with
8521 the updated NID creation version. This should correctly handle UTF8.
8525 * Implement RFC5746. Re-enable renegotiation but require the extension
8526 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8527 turns out to be a bad idea. It has been replaced by
8528 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8529 SSL_CTX_set_options(). This is really not recommended unless you
8530 know what you are doing.
8532 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8534 * Fixes to stateless session resumption handling. Use initial_ctx when
8535 issuing and attempting to decrypt tickets in case it has changed during
8536 servername handling. Use a non-zero length session ID when attempting
8537 stateless session resumption: this makes it possible to determine if
8538 a resumption has occurred immediately after receiving server hello
8539 (several places in OpenSSL subtly assume this) instead of later in
8544 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8545 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8546 fixes for a few places where the return code is not checked
8549 *Julia Lawall <julia@diku.dk>*
8551 * Add --strict-warnings option to Configure script to include devteam
8552 warnings in other configurations.
8556 * Add support for --libdir option and LIBDIR variable in makefiles. This
8557 makes it possible to install openssl libraries in locations which
8558 have names other than "lib", for example "/usr/lib64" which some
8561 *Steve Henson, based on patch from Jeremy Utley*
8563 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8564 X690 8.9.12 and can produce some misleading textual output of OIDs.
8566 *Steve Henson, reported by Dan Kaminsky*
8568 * Delete MD2 from algorithm tables. This follows the recommendation in
8569 several standards that it is not used in new applications due to
8570 several cryptographic weaknesses. For binary compatibility reasons
8571 the MD2 API is still compiled in by default.
8575 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8580 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8581 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8584 *Guenter <lists@gknw.net>*
8586 * Fix the server certificate chain building code to use X509_verify_cert(),
8587 it used to have an ad-hoc builder which was unable to cope with anything
8588 other than a simple chain.
8590 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8592 * Don't check self signed certificate signatures in X509_verify_cert()
8593 by default (a flag can override this): it just wastes time without
8594 adding any security. As a useful side effect self signed root CAs
8595 with non-FIPS digests are now usable in FIPS mode.
8599 * In dtls1_process_out_of_seq_message() the check if the current message
8600 is already buffered was missing. For every new message was memory
8601 allocated, allowing an attacker to perform an denial of service attack
8602 with sending out of seq handshake messages until there is no memory
8603 left. Additionally every future message was buffered, even if the
8604 sequence number made no sense and would be part of another handshake.
8605 So only messages with sequence numbers less than 10 in advance will be
8606 buffered. [CVE-2009-1378][]
8608 *Robin Seggelmann, discovered by Daniel Mentz*
8610 * Records are buffered if they arrive with a future epoch to be
8611 processed after finishing the corresponding handshake. There is
8612 currently no limitation to this buffer allowing an attacker to perform
8613 a DOS attack with sending records with future epochs until there is no
8614 memory left. This patch adds the pqueue_size() function to determine
8615 the size of a buffer and limits the record buffer to 100 entries.
8618 *Robin Seggelmann, discovered by Daniel Mentz*
8620 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8621 parent structure is freed. [CVE-2009-1379][]
8625 * Handle non-blocking I/O properly in SSL_shutdown() call.
8627 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8631 *Ilya O. <vrghost@gmail.com>*
8633 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009] ###
8635 * Disable renegotiation completely - this fixes a severe security
8636 problem [CVE-2009-3555][] at the cost of breaking all
8637 renegotiation. Renegotiation can be re-enabled by setting
8638 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8639 run-time. This is really not recommended unless you know what
8644 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009] ###
8646 * Don't set val to NULL when freeing up structures, it is freed up by
8647 underlying code. If sizeof(void *) > sizeof(long) this can result in
8648 zeroing past the valid field. [CVE-2009-0789][]
8650 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8652 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8653 checked correctly. This would allow some invalid signed attributes to
8654 appear to verify correctly. [CVE-2009-0591][]
8656 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8658 * Reject UniversalString and BMPString types with invalid lengths. This
8659 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8660 a legal length. [CVE-2009-0590][]
8664 * Set S/MIME signing as the default purpose rather than setting it
8665 unconditionally. This allows applications to override it at the store
8670 * Permit restricted recursion of ASN1 strings. This is needed in practice
8671 to handle some structures.
8675 * Improve efficiency of mem_gets: don't search whole buffer each time
8678 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8680 * New -hex option for openssl rand.
8684 * Print out UTF8String and NumericString when parsing ASN1.
8688 * Support NumericString type for name components.
8692 * Allow CC in the environment to override the automatically chosen
8693 compiler. Note that nothing is done to ensure flags work with the
8698 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009] ###
8700 * Properly check EVP_VerifyFinal() and similar return values
8703 *Ben Laurie, Bodo Moeller, Google Security Team*
8705 * Enable TLS extensions by default.
8709 * Allow the CHIL engine to be loaded, whether the application is
8710 multithreaded or not. (This does not release the developer from the
8711 obligation to set up the dynamic locking callbacks.)
8713 *Sander Temme <sander@temme.net>*
8715 * Use correct exit code if there is an error in dgst command.
8717 *Steve Henson; problem pointed out by Roland Dirlewanger*
8719 * Tweak Configure so that you need to say "experimental-jpake" to enable
8720 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8724 * Add experimental JPAKE support, including demo authentication in
8725 s_client and s_server.
8729 * Set the comparison function in v3_addr_canonize().
8731 *Rob Austein <sra@hactrn.net>*
8733 * Add support for XMPP STARTTLS in s_client.
8735 *Philip Paeps <philip@freebsd.org>*
8737 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8738 to ensure that even with this option, only ciphersuites in the
8739 server's preference list will be accepted. (Note that the option
8740 applies only when resuming a session, so the earlier behavior was
8741 just about the algorithm choice for symmetric cryptography.)
8745 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008] ###
8747 * Fix NULL pointer dereference if a DTLS server received
8748 ChangeCipherSpec as first record [CVE-2009-1386][].
8752 * Fix a state transition in s3_srvr.c and d1_srvr.c
8753 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
8757 * The fix in 0.9.8c that supposedly got rid of unsafe
8758 double-checked locking was incomplete for RSA blinding,
8759 addressing just one layer of what turns out to have been
8760 doubly unsafe triple-checked locking.
8762 So now fix this for real by retiring the MONT_HELPER macro
8763 in crypto/rsa/rsa_eay.c.
8766 *Bodo Moeller; problem pointed out by Marius Schilder*
8768 * Various precautionary measures:
8770 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8772 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8773 (NB: This would require knowledge of the secret session ticket key
8774 to exploit, in which case you'd be SOL either way.)
8776 - Change bn_nist.c so that it will properly handle input BIGNUMs
8777 outside the expected range.
8779 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8783 *Neel Mehta, Bodo Moeller*
8785 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8786 the load fails. Useful for distros.
8788 *Ben Laurie and the FreeBSD team*
8790 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8794 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8798 * Expand ENGINE to support engine supplied SSL client certificate functions.
8800 This work was sponsored by Logica.
8804 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8805 keystores. Support for SSL/TLS client authentication too.
8806 Not compiled unless enable-capieng specified to Configure.
8808 This work was sponsored by Logica.
8812 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8813 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8814 attribute creation routines such as certificate requests and PKCS#12
8819 ### Changes between 0.9.8g and 0.9.8h [28 May 2008] ###
8821 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8822 handshake which could lead to a client crash as found using the
8823 Codenomicon TLS test suite [CVE-2008-1672][]
8825 *Steve Henson, Mark Cox*
8827 * Fix double free in TLS server name extensions which could lead to
8828 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8832 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8834 Clear the error queue to ensure that error entries left from
8835 older function calls do not interfere with the correct operation.
8837 *Lutz Jaenicke, Erik de Castro Lopo*
8839 * Remove root CA certificates of commercial CAs:
8841 The OpenSSL project does not recommend any specific CA and does not
8842 have any policy with respect to including or excluding any CA.
8843 Therefore it does not make any sense to ship an arbitrary selection
8844 of root CA certificates with the OpenSSL software.
8848 * RSA OAEP patches to fix two separate invalid memory reads.
8849 The first one involves inputs when 'lzero' is greater than
8850 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8851 before the beginning of from). The second one involves inputs where
8852 the 'db' section contains nothing but zeroes (there is a one-byte
8853 invalid read after the end of 'db').
8855 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8857 * Partial backport from 0.9.9-dev:
8859 Introduce bn_mul_mont (dedicated Montgomery multiplication
8860 procedure) as a candidate for BIGNUM assembler implementation.
8861 While 0.9.9-dev uses assembler for various architectures, only
8862 x86_64 is available by default here in the 0.9.8 branch, and
8863 32-bit x86 is available through a compile-time setting.
8865 To try the 32-bit x86 assembler implementation, use Configure
8866 option "enable-montasm" (which exists only for this backport).
8868 As "enable-montasm" for 32-bit x86 disclaims code stability
8869 anyway, in this constellation we activate additional code
8870 backported from 0.9.9-dev for further performance improvements,
8871 namely BN_from_montgomery_word. (To enable this otherwise,
8872 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8875 *Andy Polyakov (backport partially by Bodo Moeller)*
8877 * Add TLS session ticket callback. This allows an application to set
8878 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8879 values. This is useful for key rollover for example where several key
8880 sets may exist with different names.
8884 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8885 This was broken until now in 0.9.8 releases, such that the only way
8886 a registered ENGINE could be used (assuming it initialises
8887 successfully on the host) was to explicitly set it as the default
8888 for the relevant algorithms. This is in contradiction with 0.9.7
8889 behaviour and the documentation. With this fix, when an ENGINE is
8890 registered into a given algorithm's table of implementations, the
8891 'uptodate' flag is reset so that auto-discovery will be used next
8892 time a new context for that algorithm attempts to select an
8895 *Ian Lister (tweaked by Geoff Thorpe)*
8897 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8898 implementation in the following ways:
8900 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8903 Lack of BER streaming support means one pass streaming processing is
8904 only supported if data is detached: setting the streaming flag is
8905 ignored for embedded content.
8907 CMS support is disabled by default and must be explicitly enabled
8908 with the enable-cms configuration option.
8912 * Update the GMP engine glue to do direct copies between BIGNUM and
8913 mpz_t when openssl and GMP use the same limb size. Otherwise the
8914 existing "conversion via a text string export" trick is still used.
8916 *Paul Sheer <paulsheer@gmail.com>*
8918 * Zlib compression BIO. This is a filter BIO which compressed and
8919 uncompresses any data passed through it.
8923 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
8924 RFC3394 compatible AES key wrapping.
8928 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
8929 sets string data without copying. X509_ALGOR_set0() and
8930 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
8931 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
8932 from an X509_ATTRIBUTE structure optionally checking it occurs only
8933 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
8938 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
8939 to get the expected BN_FLG_CONSTTIME behavior.
8941 *Bodo Moeller (Google)*
8945 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
8946 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
8947 - added some more tests to do_tests.pl
8948 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
8949 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
8950 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
8951 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
8952 - various changes to netware.pl to enable gcc-cross builds on Win32
8954 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
8955 - various changes to fix missing prototype warnings
8956 - fixed x86nasm.pl to create correct asm files for NASM COFF output
8957 - added AES, WHIRLPOOL and CPUID assembler code to build files
8958 - added missing AES assembler make rules to mk1mf.pl
8959 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
8961 *Guenter Knauf <eflash@gmx.net>*
8963 * Implement certificate status request TLS extension defined in RFC3546.
8964 A client can set the appropriate parameters and receive the encoded
8965 OCSP response via a callback. A server can query the supplied parameters
8966 and set the encoded OCSP response in the callback. Add simplified examples
8967 to s_client and s_server.
8971 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007] ###
8974 + Binary incompatibility of ssl_ctx_st structure
8975 + DTLS interoperation with non-compliant servers
8976 + Don't call get_session_cb() without proposed session
8977 + Fix ia64 assembler code
8979 *Andy Polyakov, Steve Henson*
8981 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007] ###
8983 * DTLS Handshake overhaul. There were longstanding issues with
8984 OpenSSL DTLS implementation, which were making it impossible for
8985 RFC 4347 compliant client to communicate with OpenSSL server.
8986 Unfortunately just fixing these incompatibilities would "cut off"
8987 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
8988 server keeps tolerating non RFC compliant syntax. The opposite is
8989 not true, 0.9.8f client can not communicate with earlier server.
8990 This update even addresses CVE-2007-4995.
8994 * Changes to avoid need for function casts in OpenSSL: some compilers
8995 (gcc 4.2 and later) reject their use.
8996 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
8999 * Add RFC4507 support to OpenSSL. This includes the corrections in
9000 RFC4507bis. The encrypted ticket format is an encrypted encoded
9001 SSL_SESSION structure, that way new session features are automatically
9004 If a client application caches session in an SSL_SESSION structure
9005 support is transparent because tickets are now stored in the encoded
9008 The SSL_CTX structure automatically generates keys for ticket
9009 protection in servers so again support should be possible
9010 with no application modification.
9012 If a client or server wishes to disable RFC4507 support then the option
9013 SSL_OP_NO_TICKET can be set.
9015 Add a TLS extension debugging callback to allow the contents of any client
9016 or server extensions to be examined.
9018 This work was sponsored by Google.
9022 * Add initial support for TLS extensions, specifically for the server_name
9023 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9024 have new members for a host name. The SSL data structure has an
9025 additional member SSL_CTX *initial_ctx so that new sessions can be
9026 stored in that context to allow for session resumption, even after the
9027 SSL has been switched to a new SSL_CTX in reaction to a client's
9028 server_name extension.
9030 New functions (subject to change):
9032 SSL_get_servername()
9033 SSL_get_servername_type()
9036 New CTRL codes and macros (subject to change):
9038 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9039 - SSL_CTX_set_tlsext_servername_callback()
9040 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9041 - SSL_CTX_set_tlsext_servername_arg()
9042 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9044 openssl s_client has a new '-servername ...' option.
9046 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9047 '-key2 ...', '-servername_fatal' (subject to change). This allows
9048 testing the HostName extension for a specific single host name ('-cert'
9049 and '-key' remain fallbacks for handshakes without HostName
9050 negotiation). If the unrecognized_name alert has to be sent, this by
9051 default is a warning; it becomes fatal with the '-servername_fatal'
9055 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9057 * Add AES and SSE2 assembly language support to VC++ build.
9061 * Mitigate attack on final subtraction in Montgomery reduction.
9065 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9066 (which previously caused an internal error).
9070 * Squeeze another 10% out of IGE mode when in != out.
9074 * AES IGE mode speedup.
9076 *Dean Gaudet (Google)*
9078 * Add the Korean symmetric 128-bit cipher SEED (see
9079 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
9080 add SEED ciphersuites from RFC 4162:
9082 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9083 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9084 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9085 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9087 To minimize changes between patchlevels in the OpenSSL 0.9.8
9088 series, SEED remains excluded from compilation unless OpenSSL
9089 is configured with 'enable-seed'.
9091 *KISA, Bodo Moeller*
9093 * Mitigate branch prediction attacks, which can be practical if a
9094 single processor is shared, allowing a spy process to extract
9095 information. For detailed background information, see
9096 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
9097 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9098 and Necessary Software Countermeasures"). The core of the change
9099 are new versions BN_div_no_branch() and
9100 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9101 respectively, which are slower, but avoid the security-relevant
9102 conditional branches. These are automatically called by BN_div()
9103 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9104 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9105 remove a conditional branch.
9107 BN_FLG_CONSTTIME is the new name for the previous
9108 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9109 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9110 in the exponent causes BN_mod_exp_mont() to use the alternative
9111 implementation in BN_mod_exp_mont_consttime().) The old name
9112 remains as a deprecated alias.
9114 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9115 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9116 constant-time implementations for more than just exponentiation.
9117 Here too the old name is kept as a deprecated alias.
9119 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9120 the BN_BLINDING structure gets an independent copy of the
9121 modulus. This means that the previous "BIGNUM *m" argument to
9122 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9123 essentially becomes "const BIGNUM *m", although we can't actually
9124 change this in the header file before 0.9.9. It allows
9125 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9126 enable BN_FLG_CONSTTIME.
9129 *Matthew D Wood (Intel Corp)*
9131 * In the SSL/TLS server implementation, be strict about session ID
9132 context matching (which matters if an application uses a single
9133 external cache for different purposes). Previously,
9134 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9135 set. This did ensure strict client verification, but meant that,
9136 with applications using a single external cache for quite
9137 different requirements, clients could circumvent ciphersuite
9138 restrictions for a given session ID context by starting a session
9139 in a different context.
9143 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9144 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9145 authentication-only ciphersuites.
9149 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9150 not complete and could lead to a possible single byte overflow
9151 [CVE-2007-5135][] [Ben Laurie]
9153 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007] ###
9155 * Since AES128 and AES256 (and similarly Camellia128 and
9156 Camellia256) share a single mask bit in the logic of
9157 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9158 kludge to work properly if AES128 is available and AES256 isn't
9159 (or if Camellia128 is available and Camellia256 isn't).
9163 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9164 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9165 When a point or a seed is encoded in a BIT STRING, we need to
9166 prevent the removal of trailing zero bits to get the proper DER
9167 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9168 of a NamedBitList, for which trailing 0 bits need to be removed.)
9172 * Have SSL/TLS server implementation tolerate "mismatched" record
9173 protocol version while receiving ClientHello even if the
9174 ClientHello is fragmented. (The server can't insist on the
9175 particular protocol version it has chosen before the ServerHello
9176 message has informed the client about his choice.)
9180 * Add RFC 3779 support.
9182 *Rob Austein for ARIN, Ben Laurie*
9184 * Load error codes if they are not already present instead of using a
9185 static variable. This allows them to be cleanly unloaded and reloaded.
9186 Improve header file function name parsing.
9190 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9191 or CAPABILITY handshake as required by RFCs.
9195 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006] ###
9197 * Introduce limits to prevent malicious keys being able to
9198 cause a denial of service. [CVE-2006-2940][]
9200 *Steve Henson, Bodo Moeller*
9202 * Fix ASN.1 parsing of certain invalid structures that can result
9203 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9205 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9206 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9208 * Fix SSL client code which could crash if connecting to a
9209 malicious SSLv2 server. [CVE-2006-4343][]
9211 *Tavis Ormandy and Will Drewry, Google Security Team*
9213 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9214 match only those. Before that, "AES256-SHA" would be interpreted
9215 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9216 the same strength classification in 0.9.7h) as we currently only
9217 have a single AES bit in the ciphersuite description bitmap.
9218 That change, however, also applied to ciphersuite strings such as
9219 "RC4-MD5" that intentionally matched multiple ciphersuites --
9220 namely, SSL 2.0 ciphersuites in addition to the more common ones
9221 from SSL 3.0/TLS 1.0.
9223 So we change the selection algorithm again: Naming an explicit
9224 ciphersuite selects this one ciphersuite, and any other similar
9225 ciphersuite (same bitmap) from *other* protocol versions.
9226 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9227 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9229 Since SSL 2.0 does not have any ciphersuites for which the
9230 128/256 bit distinction would be relevant, this works for now.
9231 The proper fix will be to use different bits for AES128 and
9232 AES256, which would have avoided the problems from the beginning;
9233 however, bits are scarce, so we can only do this in a new release
9234 (not just a patchlevel) when we can change the SSL_CIPHER
9235 definition to split the single 'unsigned long mask' bitmap into
9236 multiple values to extend the available space.
9241 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006] ###
9243 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9244 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9246 * Add AES IGE and biIGE modes.
9250 * Change the Unix randomness entropy gathering to use poll() when
9251 possible instead of select(), since the latter has some
9252 undesirable limitations.
9254 *Darryl Miles via Richard Levitte and Bodo Moeller*
9256 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9257 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9258 cannot be implicitly activated as part of, e.g., the "AES" alias.
9259 However, please upgrade to OpenSSL 0.9.9[-dev] for
9260 non-experimental use of the ECC ciphersuites to get TLS extension
9261 support, which is required for curve and point format negotiation
9262 to avoid potential handshake problems.
9266 * Disable rogue ciphersuites:
9268 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9269 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9270 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9272 The latter two were purportedly from
9273 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9276 Also deactivate the remaining ciphersuites from
9277 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9278 unofficial, and the ID has long expired.
9282 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9283 dual-core machines) and other potential thread-safety issues.
9287 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9288 versions), which is now available for royalty-free use
9289 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
9290 Also, add Camellia TLS ciphersuites from RFC 4132.
9292 To minimize changes between patchlevels in the OpenSSL 0.9.8
9293 series, Camellia remains excluded from compilation unless OpenSSL
9294 is configured with 'enable-camellia'.
9298 * Disable the padding bug check when compression is in use. The padding
9299 bug check assumes the first packet is of even length, this is not
9300 necessarily true if compression is enabled and can result in false
9301 positives causing handshake failure. The actual bug test is ancient
9302 code so it is hoped that implementations will either have fixed it by
9303 now or any which still have the bug do not support compression.
9307 ### Changes between 0.9.8a and 0.9.8b [04 May 2006] ###
9309 * When applying a cipher rule check to see if string match is an explicit
9310 cipher suite and only match that one cipher suite if it is.
9314 * Link in manifests for VC++ if needed.
9316 *Austin Ziegler <halostatue@gmail.com>*
9318 * Update support for ECC-based TLS ciphersuites according to
9319 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9320 TLS extensions, which are supported starting with the 0.9.9
9321 branch, not in the OpenSSL 0.9.8 branch).
9325 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9326 opaque EVP_CIPHER_CTX handling.
9330 * Fixes and enhancements to zlib compression code. We now only use
9331 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9332 to conform with the standards mentioned here:
9333 http://www.zlib.net/DLL_FAQ.txt
9334 Static zlib linking now works on Windows and the new --with-zlib-include
9335 --with-zlib-lib options to Configure can be used to supply the location
9336 of the headers and library. Gracefully handle case where zlib library
9341 * Several fixes and enhancements to the OID generation code. The old code
9342 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9343 handle numbers larger than ULONG_MAX, truncated printing and had a
9344 non standard OBJ_obj2txt() behaviour.
9348 * Add support for building of engines under engine/ as shared libraries
9349 under VC++ build system.
9353 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9354 Hopefully, we will not see any false combination of paths any more.
9358 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005] ###
9360 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9361 (part of SSL_OP_ALL). This option used to disable the
9362 countermeasure against man-in-the-middle protocol-version
9363 rollback in the SSL 2.0 server implementation, which is a bad
9364 idea. [CVE-2005-2969][]
9366 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9367 for Information Security, National Institute of Advanced Industrial
9368 Science and Technology [AIST*, Japan)]
9370 * Add two function to clear and return the verify parameter flags.
9374 * Keep cipherlists sorted in the source instead of sorting them at
9375 runtime, thus removing the need for a lock.
9379 * Avoid some small subgroup attacks in Diffie-Hellman.
9381 *Nick Mathewson and Ben Laurie*
9383 * Add functions for well-known primes.
9387 * Extended Windows CE support.
9389 *Satoshi Nakamura and Andy Polyakov*
9391 * Initialize SSL_METHOD structures at compile time instead of during
9392 runtime, thus removing the need for a lock.
9396 * Make PKCS7_decrypt() work even if no certificate is supplied by
9397 attempting to decrypt each encrypted key in turn. Add support to
9402 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005] ###
9404 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9407 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9411 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9412 key into the same file any more.
9416 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9420 * Add -utf8 command line and config file option to 'ca'.
9422 *Stefan <stf@udoma.org*
9424 * Removed the macro des_crypt(), as it seems to conflict with some
9425 libraries. Use DES_crypt().
9429 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9430 involves renaming the source and generated shared-libs for
9431 both. The engines will accept the corrected or legacy ids
9432 ('ncipher' and '4758_cca' respectively) when binding. NB,
9433 this only applies when building 'shared'.
9435 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9437 * Add attribute functions to EVP_PKEY structure. Modify
9438 PKCS12_create() to recognize a CSP name attribute and
9439 use it. Make -CSP option work again in pkcs12 utility.
9443 * Add new functionality to the bn blinding code:
9444 - automatic re-creation of the BN_BLINDING parameters after
9445 a fixed number of uses (currently 32)
9446 - add new function for parameter creation
9447 - introduce flags to control the update behaviour of the
9448 BN_BLINDING parameters
9449 - hide BN_BLINDING structure
9450 Add a second BN_BLINDING slot to the RSA structure to improve
9451 performance when a single RSA object is shared among several
9456 * Add support for DTLS.
9458 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9460 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9461 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9465 * Remove buggy and incomplete DH cert support from
9466 ssl/ssl_rsa.c and ssl/s3_both.c
9470 * Use SHA-1 instead of MD5 as the default digest algorithm for
9471 the apps/openssl applications.
9475 * Compile clean with "-Wall -Wmissing-prototypes
9476 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9477 DEBUG_SAFESTACK must also be set.
9481 * Change ./Configure so that certain algorithms can be disabled by default.
9482 The new counterpiece to "no-xxx" is "enable-xxx".
9484 The patented RC5 and MDC2 algorithms will now be disabled unless
9485 "enable-rc5" and "enable-mdc2", respectively, are specified.
9487 (IDEA remains enabled despite being patented. This is because IDEA
9488 is frequently required for interoperability, and there is no license
9489 fee for non-commercial use. As before, "no-idea" can be used to
9490 avoid this algorithm.)
9495 * Add processing of proxy certificates (see RFC 3820). This work was
9496 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9497 EGEE (Enabling Grids for E-science in Europe).
9501 * RC4 performance overhaul on modern architectures/implementations, such
9502 as Intel P4, IA-64 and AMD64.
9506 * New utility extract-section.pl. This can be used specify an alternative
9507 section number in a pod file instead of having to treat each file as
9508 a separate case in Makefile. This can be done by adding two lines to the
9511 =for comment openssl_section:XXX
9513 The blank line is mandatory.
9518 * New arguments -certform, -keyform and -pass for s_client and s_server
9519 to allow alternative format key and certificate files and passphrase
9524 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9525 update associated structures and add various utility functions.
9527 Add new policy related verify parameters, include policy checking in
9528 standard verify code. Enhance 'smime' application with extra parameters
9529 to support policy checking and print out.
9533 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9534 Nehemiah processors. These extensions support AES encryption in hardware
9535 as well as RNG (though RNG support is currently disabled).
9537 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9539 * Deprecate BN_[get|set]_params() functions (they were ignored internally).
9543 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9545 *Andy Polyakov and a number of other people*
9547 * Improved PowerPC platform support. Most notably BIGNUM assembler
9548 implementation contributed by IBM.
9550 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9552 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9553 exponent rather than 'unsigned long'. There is a corresponding change to
9554 the new 'rsa_keygen' element of the RSA_METHOD structure.
9556 *Jelte Jansen, Geoff Thorpe*
9558 * Functionality for creating the initial serial number file is now
9559 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9561 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9562 number file to 1, which is bound to cause problems. To avoid
9563 the problems while respecting compatibility between different 0.9.7
9564 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9565 CA.pl for serial number initialization. With the new release 0.9.8,
9566 we can fix the problem directly in the 'ca' utility.)
9570 * Reduced header interdependencies by declaring more opaque objects in
9571 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9572 give fewer recursive includes, which could break lazy source code - so
9573 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9574 developers should define this symbol when building and using openssl to
9575 ensure they track the recommended behaviour, interfaces, [etc], but
9576 backwards-compatible behaviour prevails when this isn't defined.
9580 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9584 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9585 This will generate a random key of the appropriate length based on the
9586 cipher context. The EVP_CIPHER can provide its own random key generation
9587 routine to support keys of a specific form. This is used in the des and
9588 3des routines to generate a key of the correct parity. Update S/MIME
9589 code to use new functions and hence generate correct parity DES keys.
9590 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9591 valid (weak or incorrect parity).
9595 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9596 as looking them up. This is useful when the verified structure may contain
9597 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9598 present unless the new PKCS7_NO_CRL flag is asserted.
9602 * Extend ASN1 oid configuration module. It now additionally accepts the
9605 shortName = some long name, 1.2.3.4
9609 * Reimplemented the BN_CTX implementation. There is now no more static
9610 limitation on the number of variables it can handle nor the depth of the
9611 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9612 information can now expand as required, and rather than having a single
9613 static array of bignums, BN_CTX now uses a linked-list of such arrays
9614 allowing it to expand on demand whilst maintaining the usefulness of
9615 BN_CTX's "bundling".
9619 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9620 to allow all RSA operations to function using a single BN_CTX.
9624 * Preliminary support for certificate policy evaluation and checking. This
9625 is initially intended to pass the tests outlined in "Conformance Testing
9626 of Relying Party Client Certificate Path Processing Logic" v1.07.
9630 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9631 remained unused and not that useful. A variety of other little bignum
9632 tweaks and fixes have also been made continuing on from the audit (see
9637 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9638 associated ASN1, EVP and SSL functions and old ASN1 macros.
9642 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9643 and this should never fail. So the return value from the use of
9644 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9645 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9649 * BN_CTX_get() should return zero-valued bignums, providing the same
9650 initialised value as BN_new().
9652 *Geoff Thorpe, suggested by Ulf Möller*
9654 * Support for inhibitAnyPolicy certificate extension.
9658 * An audit of the BIGNUM code is underway, for which debugging code is
9659 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9660 is considered valid when processing BIGNUMs, and causes execution to
9661 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9662 further steps are taken to deliberately pollute unused data in BIGNUM
9663 structures to try and expose faulty code further on. For now, openssl will
9664 (in its default mode of operation) continue to tolerate the inconsistent
9665 forms that it has tolerated in the past, but authors and packagers should
9666 consider trying openssl and their own applications when compiled with
9667 these debugging symbols defined. It will help highlight potential bugs in
9668 their own code, and will improve the test coverage for OpenSSL itself. At
9669 some point, these tighter rules will become openssl's default to improve
9670 maintainability, though the assert()s and other overheads will remain only
9671 in debugging configurations. See bn.h for more details.
9673 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9675 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9676 that can only be obtained through BN_CTX_new() (which implicitly
9677 initialises it). The presence of this function only made it possible
9678 to overwrite an existing structure (and cause memory leaks).
9682 * Because of the callback-based approach for implementing LHASH as a
9683 template type, lh_insert() adds opaque objects to hash-tables and
9684 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9685 to clean up those corresponding objects before destroying the hash table
9686 (and losing the object pointers). So some over-zealous constifications in
9687 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9688 objects as "const" and the lh_doall[_arg] callback wrappers are not
9689 prototyped to have "const" restrictions on the object pointers they are
9690 given (and so aren't required to cast them away any more).
9694 * The tmdiff.h API was so ugly and minimal that our own timing utility
9695 (speed) prefers to use its own implementation. The two implementations
9696 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9697 its object type properly exposed (MS_TM) instead of casting to/from "char
9698 *". This may still change yet if someone realises MS_TM and "ms_time_***"
9699 aren't necessarily the greatest nomenclatures - but this is what was used
9700 internally to the implementation so I've used that for now.
9704 * Ensure that deprecated functions do not get compiled when
9705 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9706 the self-tests were still using deprecated key-generation functions so
9707 these have been updated also.
9711 * Reorganise PKCS#7 code to separate the digest location functionality
9712 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9713 New function PKCS7_set_digest() to set the digest type for PKCS#7
9714 digestedData type. Add additional code to correctly generate the
9715 digestedData type and add support for this type in PKCS7 initialization
9720 * New function PKCS7_set0_type_other() this initializes a PKCS7
9721 structure of type "other".
9725 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9726 sure the loop does correctly stop and breaking ("division by zero")
9727 modulus operations are not performed. The (pre-generated) prime
9728 table crypto/bn/bn_prime.h was already correct, but it could not be
9729 re-generated on some platforms because of the "division by zero"
9730 situation in the script.
9732 *Ralf S. Engelschall*
9734 * Update support for ECC-based TLS ciphersuites according to
9735 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9736 SHA-1 now is only used for "small" curves (where the
9737 representation of a field element takes up to 24 bytes); for
9738 larger curves, the field element resulting from ECDH is directly
9739 used as premaster secret.
9741 *Douglas Stebila (Sun Microsystems Laboratories)*
9743 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9744 curve secp160r1 to the tests.
9746 *Douglas Stebila (Sun Microsystems Laboratories)*
9748 * Add the possibility to load symbols globally with DSO.
9750 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9752 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9753 control of the error stack.
9757 * Add support for STORE in ENGINE.
9761 * Add the STORE type. The intention is to provide a common interface
9762 to certificate and key stores, be they simple file-based stores, or
9763 HSM-type store, or LDAP stores, or...
9764 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9768 * Add a generic structure called OPENSSL_ITEM. This can be used to
9769 pass a list of arguments to any function as well as provide a way
9770 for a function to pass data back to the caller.
9774 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9775 works like BUF_strdup() but can be used to duplicate a portion of
9776 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9781 * Add the function sk_find_ex() which works like sk_find(), but will
9782 return an index to an element even if an exact match couldn't be
9783 found. The index is guaranteed to point at the element where the
9784 searched-for key would be inserted to preserve sorting order.
9788 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9789 takes an extra flags argument for optional functionality. Currently,
9790 the following flags are defined:
9792 OBJ_BSEARCH_VALUE_ON_NOMATCH
9793 This one gets OBJ_bsearch_ex() to return a pointer to the first
9794 element where the comparing function returns a negative or zero
9797 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9798 This one gets OBJ_bsearch_ex() to return a pointer to the first
9799 element where the comparing function returns zero. This is useful
9800 if there are more than one element where the comparing function
9805 * Make it possible to create self-signed certificates with 'openssl ca'
9806 in such a way that the self-signed certificate becomes part of the
9807 CA database and uses the same mechanisms for serial number generation
9808 as all other certificate signing. The new flag '-selfsign' enables
9809 this functionality. Adapt CA.sh and CA.pl.in.
9813 * Add functionality to check the public key of a certificate request
9814 against a given private. This is useful to check that a certificate
9815 request can be signed by that key (self-signing).
9819 * Make it possible to have multiple active certificates with the same
9820 subject in the CA index file. This is done only if the keyword
9821 'unique_subject' is set to 'no' in the main CA section (default
9822 if 'CA_default') of the configuration file. The value is saved
9823 with the database itself in a separate index attribute file,
9824 named like the index file with '.attr' appended to the name.
9828 * Generate multi-valued AVAs using '+' notation in config files for
9833 * Support for nameConstraints certificate extension.
9837 * Support for policyConstraints certificate extension.
9841 * Support for policyMappings certificate extension.
9845 * Make sure the default DSA_METHOD implementation only uses its
9846 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9847 and change its own handlers to be NULL so as to remove unnecessary
9848 indirection. This lets alternative implementations fallback to the
9849 default implementation more easily.
9853 * Support for directoryName in GeneralName related extensions
9858 * Make it possible to link applications using Makefile.shared.
9859 Make that possible even when linking against static libraries!
9863 * Support for single pass processing for S/MIME signing. This now
9864 means that S/MIME signing can be done from a pipe, in addition
9865 cleartext signing (multipart/signed type) is effectively streaming
9866 and the signed data does not need to be all held in memory.
9868 This is done with a new flag PKCS7_STREAM. When this flag is set
9869 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9870 is done after the data is output (and digests calculated) in
9871 SMIME_write_PKCS7().
9875 * Add full support for -rpath/-R, both in shared libraries and
9876 applications, at least on the platforms where it's known how
9881 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9882 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9883 will now compute a table of multiples of the generator that
9884 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9885 faster (notably in the case of a single point multiplication,
9886 scalar * generator).
9888 *Nils Larsch, Bodo Moeller*
9890 * IPv6 support for certificate extensions. The various extensions
9891 which use the IP:a.b.c.d can now take IPv6 addresses using the
9892 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9897 * Added an ENGINE that implements RSA by performing private key
9898 exponentiations with the GMP library. The conversions to and from
9899 GMP's mpz_t format aren't optimised nor are any montgomery forms
9900 cached, and on x86 it appears OpenSSL's own performance has caught up.
9901 However there are likely to be other architectures where GMP could
9902 provide a boost. This ENGINE is not built in by default, but it can be
9903 specified at Configure time and should be accompanied by the necessary
9904 linker additions, eg;
9905 ./config -DOPENSSL_USE_GMP -lgmp
9909 * "openssl engine" will not display ENGINE/DSO load failure errors when
9910 testing availability of engines with "-t" - the old behaviour is
9911 produced by increasing the feature's verbosity with "-tt".
9915 * ECDSA routines: under certain error conditions uninitialized BN objects
9916 could be freed. Solution: make sure initialization is performed early
9917 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
9922 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
9923 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9924 software implementations. For DSA and DH, parameter generation can
9925 also be overridden by providing the appropriate method callbacks.
9929 * Change the "progress" mechanism used in key-generation and
9930 primality testing to functions that take a new BN_GENCB pointer in
9931 place of callback/argument pairs. The new API functions have "_ex"
9932 postfixes and the older functions are reimplemented as wrappers for
9933 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
9934 declarations of the old functions to help (graceful) attempts to
9935 migrate to the new functions. Also, the new key-generation API
9936 functions operate on a caller-supplied key-structure and return
9937 success/failure rather than returning a key or NULL - this is to
9938 help make "keygen" another member function of RSA_METHOD etc.
9940 Example for using the new callback interface:
9942 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
9946 BN_GENCB_set(&my_cb, my_callback, my_arg);
9948 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
9949 /* For the meaning of a, b in calls to my_callback(), see the
9950 * documentation of the function that calls the callback.
9951 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
9952 * my_callback should return 1 if it wants BN_is_prime_ex()
9953 * to continue, or 0 to stop.
9958 * Change the ZLIB compression method to be stateful, and make it
9959 available to TLS with the number defined in
9960 draft-ietf-tls-compression-04.txt.
9964 * Add the ASN.1 structures and functions for CertificatePair, which
9965 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
9967 CertificatePair ::= SEQUENCE {
9968 forward [0] Certificate OPTIONAL,
9969 reverse [1] Certificate OPTIONAL,
9970 -- at least one of the pair shall be present -- }
9972 Also implement the PEM functions to read and write certificate
9973 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
9975 This needed to be defined, mostly for the sake of the LDAP
9976 attribute crossCertificatePair, but may prove useful elsewhere as
9981 * Make it possible to inhibit symlinking of shared libraries in
9982 Makefile.shared, for Cygwin's sake.
9986 * Extend the BIGNUM API by creating a function
9987 void BN_set_negative(BIGNUM *a, int neg);
9988 and a macro that behave like
9989 int BN_is_negative(const BIGNUM *a);
9991 to avoid the need to access 'a->neg' directly in applications.
9995 * Implement fast modular reduction for pseudo-Mersenne primes
9996 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
9997 EC_GROUP_new_curve_GFp() will now automatically use this
10000 *Nils Larsch <nla@trustcenter.de>*
10002 * Add new lock type (CRYPTO_LOCK_BN).
10006 * Change the ENGINE framework to automatically load engines
10007 dynamically from specific directories unless they could be
10008 found to already be built in or loaded. Move all the
10009 current engines except for the cryptodev one to a new
10010 directory engines/.
10011 The engines in engines/ are built as shared libraries if
10012 the "shared" options was given to ./Configure or ./config.
10013 Otherwise, they are inserted in libcrypto.a.
10014 /usr/local/ssl/engines is the default directory for dynamic
10015 engines, but that can be overridden at configure time through
10016 the usual use of --prefix and/or --openssldir, and at run
10017 time with the environment variable OPENSSL_ENGINES.
10019 *Geoff Thorpe and Richard Levitte*
10021 * Add Makefile.shared, a helper makefile to build shared
10022 libraries. Adapt Makefile.org.
10026 * Add version info to Win32 DLLs.
10028 *Peter 'Luna' Runestig" <peter@runestig.com>*
10030 * Add new 'medium level' PKCS#12 API. Certificates and keys
10031 can be added using this API to created arbitrary PKCS#12
10032 files while avoiding the low level API.
10034 New options to PKCS12_create(), key or cert can be NULL and
10035 will then be omitted from the output file. The encryption
10036 algorithm NIDs can be set to -1 for no encryption, the mac
10037 iteration count can be set to 0 to omit the mac.
10039 Enhance pkcs12 utility by making the -nokeys and -nocerts
10040 options work when creating a PKCS#12 file. New option -nomac
10041 to omit the mac, NONE can be set for an encryption algorithm.
10042 New code is modified to use the enhanced PKCS12_create()
10043 instead of the low level API.
10047 * Extend ASN1 encoder to support indefinite length constructed
10048 encoding. This can output sequences tags and octet strings in
10049 this form. Modify pk7_asn1.c to support indefinite length
10050 encoding. This is experimental and needs additional code to
10051 be useful, such as an ASN1 bio and some enhanced streaming
10054 Extend template encode functionality so that tagging is passed
10055 down to the template encoder.
10059 * Let 'openssl req' fail if an argument to '-newkey' is not
10060 recognized instead of using RSA as a default.
10064 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10065 As these are not official, they are not included in "ALL";
10066 the "ECCdraft" ciphersuite group alias can be used to select them.
10068 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10070 * Add ECDH engine support.
10072 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10074 * Add ECDH in new directory crypto/ecdh/.
10076 *Douglas Stebila (Sun Microsystems Laboratories)*
10078 * Let BN_rand_range() abort with an error after 100 iterations
10079 without success (which indicates a broken PRNG).
10083 * Change BN_mod_sqrt() so that it verifies that the input value
10084 is really the square of the return value. (Previously,
10085 BN_mod_sqrt would show GIGO behaviour.)
10089 * Add named elliptic curves over binary fields from X9.62, SECG,
10090 and WAP/WTLS; add OIDs that were still missing.
10092 *Sheueling Chang Shantz and Douglas Stebila
10093 (Sun Microsystems Laboratories)*
10095 * Extend the EC library for elliptic curves over binary fields
10096 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10099 EC_GF2m_simple_method
10103 EC_GROUP_new_curve_GF2m
10104 EC_GROUP_set_curve_GF2m
10105 EC_GROUP_get_curve_GF2m
10106 EC_POINT_set_affine_coordinates_GF2m
10107 EC_POINT_get_affine_coordinates_GF2m
10108 EC_POINT_set_compressed_coordinates_GF2m
10110 Point compression for binary fields is disabled by default for
10111 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10114 As binary polynomials are represented as BIGNUMs, various members
10115 of the EC_GROUP and EC_POINT data structures can be shared
10116 between the implementations for prime fields and binary fields;
10117 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
10118 are essentially identical to their ..._GFp counterparts.
10119 (For simplicity, the '..._GFp' prefix has been dropped from
10120 various internal method names.)
10122 An internal 'field_div' method (similar to 'field_mul' and
10123 'field_sqr') has been added; this is used only for binary fields.
10125 *Sheueling Chang Shantz and Douglas Stebila
10126 (Sun Microsystems Laboratories)*
10128 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10129 through methods ('mul', 'precompute_mult').
10131 The generic implementations (now internally called 'ec_wNAF_mul'
10132 and 'ec_wNAF_precomputed_mult') remain the default if these
10133 methods are undefined.
10135 *Sheueling Chang Shantz and Douglas Stebila
10136 (Sun Microsystems Laboratories)*
10138 * New function EC_GROUP_get_degree, which is defined through
10139 EC_METHOD. For curves over prime fields, this returns the bit
10140 length of the modulus.
10142 *Sheueling Chang Shantz and Douglas Stebila
10143 (Sun Microsystems Laboratories)*
10145 * New functions EC_GROUP_dup, EC_POINT_dup.
10146 (These simply call ..._new and ..._copy).
10148 *Sheueling Chang Shantz and Douglas Stebila
10149 (Sun Microsystems Laboratories)*
10151 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10152 Polynomials are represented as BIGNUMs (where the sign bit is not
10153 used) in the following functions [macros]:
10156 BN_GF2m_sub [= BN_GF2m_add]
10157 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10158 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10159 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10161 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10162 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10163 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10164 BN_GF2m_cmp [= BN_ucmp]
10166 (Note that only the 'mod' functions are actually for fields GF(2^m).
10167 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10169 For some functions, an the irreducible polynomial defining a
10170 field can be given as an 'unsigned int[]' with strictly
10171 decreasing elements giving the indices of those bits that are set;
10172 i.e., p[] represents the polynomial
10173 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10175 p[0] > p[1] > ... > p[k] = 0.
10176 This applies to the following functions:
10179 BN_GF2m_mod_mul_arr
10180 BN_GF2m_mod_sqr_arr
10181 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10182 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10183 BN_GF2m_mod_exp_arr
10184 BN_GF2m_mod_sqrt_arr
10185 BN_GF2m_mod_solve_quad_arr
10189 Conversion can be performed by the following functions:
10194 bntest.c has additional tests for binary polynomial arithmetic.
10196 Two implementations for BN_GF2m_mod_div() are available.
10197 The default algorithm simply uses BN_GF2m_mod_inv() and
10198 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10199 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10200 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10202 *Sheueling Chang Shantz and Douglas Stebila
10203 (Sun Microsystems Laboratories)*
10205 * Add new error code 'ERR_R_DISABLED' that can be used when some
10206 functionality is disabled at compile-time.
10208 *Douglas Stebila <douglas.stebila@sun.com>*
10210 * Change default behaviour of 'openssl asn1parse' so that more
10211 information is visible when viewing, e.g., a certificate:
10213 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10214 mode the content of non-printable OCTET STRINGs is output in a
10215 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10216 avoid the appearance of a printable string.
10218 *Nils Larsch <nla@trustcenter.de>*
10220 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10222 EC_GROUP_set_asn1_flag()
10223 EC_GROUP_get_asn1_flag()
10224 EC_GROUP_set_point_conversion_form()
10225 EC_GROUP_get_point_conversion_form()
10226 These control ASN1 encoding details:
10227 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10228 has been set to OPENSSL_EC_NAMED_CURVE.
10229 - Points are encoded in uncompressed form by default; options for
10230 asn1_for are as for point2oct, namely
10231 POINT_CONVERSION_COMPRESSED
10232 POINT_CONVERSION_UNCOMPRESSED
10233 POINT_CONVERSION_HYBRID
10235 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10237 EC_GROUP_set_seed()
10238 EC_GROUP_get0_seed()
10239 EC_GROUP_get_seed_len()
10240 This is used only for ASN1 purposes (so far).
10242 *Nils Larsch <nla@trustcenter.de>*
10244 * Add 'field_type' member to EC_METHOD, which holds the NID
10245 of the appropriate field type OID. The new function
10246 EC_METHOD_get_field_type() returns this value.
10248 *Nils Larsch <nla@trustcenter.de>*
10251 EC_POINT_point2bn()
10252 EC_POINT_bn2point()
10253 EC_POINT_point2hex()
10254 EC_POINT_hex2point()
10255 providing useful interfaces to EC_POINT_point2oct() and
10256 EC_POINT_oct2point().
10258 *Nils Larsch <nla@trustcenter.de>*
10260 * Change internals of the EC library so that the functions
10261 EC_GROUP_set_generator()
10262 EC_GROUP_get_generator()
10263 EC_GROUP_get_order()
10264 EC_GROUP_get_cofactor()
10265 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10266 to methods, which would lead to unnecessary code duplication when
10267 adding different types of curves.
10269 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10271 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10272 arithmetic, and such that modified wNAFs are generated
10273 (which avoid length expansion in many cases).
10277 * Add a function EC_GROUP_check_discriminant() (defined via
10278 EC_METHOD) that verifies that the curve discriminant is non-zero.
10280 Add a function EC_GROUP_check() that makes some sanity tests
10281 on a EC_GROUP, its generator and order. This includes
10282 EC_GROUP_check_discriminant().
10284 *Nils Larsch <nla@trustcenter.de>*
10286 * Add ECDSA in new directory crypto/ecdsa/.
10288 Add applications 'openssl ecparam' and 'openssl ecdsa'
10289 (these are based on 'openssl dsaparam' and 'openssl dsa').
10291 ECDSA support is also included in various other files across the
10292 library. Most notably,
10293 - 'openssl req' now has a '-newkey ecdsa:file' option;
10294 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10295 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10296 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10297 them suitable for ECDSA where domain parameters must be
10298 extracted before the specific public key;
10299 - ECDSA engine support has been added.
10301 *Nils Larsch <nla@trustcenter.de>*
10303 * Include some named elliptic curves, and add OIDs from X9.62,
10304 SECG, and WAP/WTLS. Each curve can be obtained from the new
10306 EC_GROUP_new_by_curve_name(),
10307 and the list of available named curves can be obtained with
10308 EC_get_builtin_curves().
10309 Also add a 'curve_name' member to EC_GROUP objects, which can be
10311 EC_GROUP_set_curve_name()
10312 EC_GROUP_get_curve_name()
10314 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10316 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10317 was actually never needed) and in BN_mul(). The removal in BN_mul()
10318 required a small change in bn_mul_part_recursive() and the addition
10319 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10320 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10321 bn_sub_words() and bn_add_words() except they take arrays with
10326 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007] ###
10328 * Cleanse PEM buffers before freeing them since they may contain
10331 *Benjamin Bennett <ben@psc.edu>*
10333 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10334 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10335 authentication-only ciphersuites.
10339 * Since AES128 and AES256 share a single mask bit in the logic of
10340 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10341 kludge to work properly if AES128 is available and AES256 isn't.
10345 * Expand security boundary to match 1.1.1 module.
10349 * Remove redundant features: hash file source, editing of test vectors
10350 modify fipsld to use external fips_premain.c signature.
10354 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10355 run algorithm test programs.
10359 * Make algorithm test programs more tolerant of whitespace.
10363 * Have SSL/TLS server implementation tolerate "mismatched" record
10364 protocol version while receiving ClientHello even if the
10365 ClientHello is fragmented. (The server can't insist on the
10366 particular protocol version it has chosen before the ServerHello
10367 message has informed the client about his choice.)
10371 * Load error codes if they are not already present instead of using a
10372 static variable. This allows them to be cleanly unloaded and reloaded.
10376 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006] ###
10378 * Introduce limits to prevent malicious keys being able to
10379 cause a denial of service. [CVE-2006-2940][]
10381 *Steve Henson, Bodo Moeller*
10383 * Fix ASN.1 parsing of certain invalid structures that can result
10384 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10386 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10387 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10389 * Fix SSL client code which could crash if connecting to a
10390 malicious SSLv2 server. [CVE-2006-4343][]
10392 *Tavis Ormandy and Will Drewry, Google Security Team*
10394 * Change ciphersuite string processing so that an explicit
10395 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10396 will no longer include "AES128-SHA"), and any other similar
10397 ciphersuite (same bitmap) from *other* protocol versions (so that
10398 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10399 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10400 changes from 0.9.8b and 0.9.8d.
10404 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006] ###
10406 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10407 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10409 * Change the Unix randomness entropy gathering to use poll() when
10410 possible instead of select(), since the latter has some
10411 undesirable limitations.
10413 *Darryl Miles via Richard Levitte and Bodo Moeller*
10415 * Disable rogue ciphersuites:
10417 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10418 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10419 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10421 The latter two were purportedly from
10422 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10425 Also deactivate the remaining ciphersuites from
10426 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10427 unofficial, and the ID has long expired.
10431 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10432 dual-core machines) and other potential thread-safety issues.
10436 ### Changes between 0.9.7i and 0.9.7j [04 May 2006] ###
10438 * Adapt fipsld and the build system to link against the validated FIPS
10439 module in FIPS mode.
10443 * Fixes for VC++ 2005 build under Windows.
10447 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10448 from a Windows bash shell such as MSYS. It is autodetected from the
10449 "config" script when run from a VC++ environment. Modify standard VC++
10450 build to use fipscanister.o from the GNU make build.
10454 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005] ###
10456 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10457 The value now differs depending on if you build for FIPS or not.
10458 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10459 safely run with a non-FIPSed libcrypto, as it may crash because of
10460 the difference induced by this change.
10464 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005] ###
10466 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10467 (part of SSL_OP_ALL). This option used to disable the
10468 countermeasure against man-in-the-middle protocol-version
10469 rollback in the SSL 2.0 server implementation, which is a bad
10470 idea. [CVE-2005-2969][]
10472 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10473 for Information Security, National Institute of Advanced Industrial
10474 Science and Technology [AIST*, Japan)]
10476 * Minimal support for X9.31 signatures and PSS padding modes. This is
10477 mainly for FIPS compliance and not fully integrated at this stage.
10481 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10482 the exponentiation using a fixed-length exponent. (Otherwise,
10483 the information leaked through timing could expose the secret key
10484 after many signatures; cf. Bleichenbacher's attack on DSA with
10489 * Make a new fixed-window mod_exp implementation the default for
10490 RSA, DSA, and DH private-key operations so that the sequence of
10491 squares and multiplies and the memory access pattern are
10492 independent of the particular secret key. This will mitigate
10493 cache-timing and potential related attacks.
10495 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10496 and this is automatically used by BN_mod_exp_mont() if the new flag
10497 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10498 will use this BN flag for private exponents unless the flag
10499 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10500 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10503 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10505 * Change the client implementation for SSLv23_method() and
10506 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10507 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10508 (Previously, the SSL 2.0 backwards compatible Client Hello
10509 message format would be used even with SSL_OP_NO_SSLv2.)
10513 * Add support for smime-type MIME parameter in S/MIME messages which some
10518 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10519 a threadsafe manner. Modify rsa code to use new function and add calls
10520 to dsa and dh code (which had race conditions before).
10524 * Include the fixed error library code in the C error file definitions
10525 instead of fixing them up at runtime. This keeps the error code
10526 structures constant.
10530 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005] ###
10532 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10535 * Fixes for newer kerberos headers. NB: the casts are needed because
10536 the 'length' field is signed on one version and unsigned on another
10537 with no (?) obvious way to tell the difference, without these VC++
10538 complains. Also the "definition" of FAR (blank) is no longer included
10539 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10540 some needed definitions.
10544 * Undo Cygwin change.
10548 * Added support for proxy certificates according to RFC 3820.
10549 Because they may be a security thread to unaware applications,
10550 they must be explicitly allowed in run-time. See
10551 docs/HOWTO/proxy_certificates.txt for further information.
10555 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005] ###
10557 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10558 server and client random values. Previously
10559 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10560 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10562 This change has negligible security impact because:
10564 1. Server and client random values still have 24 bytes of pseudo random
10567 2. Server and client random values are sent in the clear in the initial
10570 3. The master secret is derived using the premaster secret (48 bytes in
10571 size for static RSA ciphersuites) as well as client server and random
10574 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10577 *Stephen Henson, reported by UK NISCC*
10579 * Use Windows randomness collection on Cygwin.
10583 * Fix hang in EGD/PRNGD query when communication socket is closed
10584 prematurely by EGD/PRNGD.
10586 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10588 * Prompt for pass phrases when appropriate for PKCS12 input format.
10592 * Back-port of selected performance improvements from development
10593 branch, as well as improved support for PowerPC platforms.
10597 * Add lots of checks for memory allocation failure, error codes to indicate
10598 failure and freeing up memory if a failure occurs.
10600 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10602 * Add new -passin argument to dgst.
10606 * Perform some character comparisons of different types in X509_NAME_cmp:
10607 this is needed for some certificates that re-encode DNs into UTF8Strings
10608 (in violation of RFC3280) and can't or won't issue name rollover
10613 * Make an explicit check during certificate validation to see that
10614 the CA setting in each certificate on the chain is correct. As a
10615 side effect always do the following basic checks on extensions,
10616 not just when there's an associated purpose to the check:
10618 - if there is an unhandled critical extension (unless the user
10619 has chosen to ignore this fault)
10620 - if the path length has been exceeded (if one is set at all)
10621 - that certain extensions fit the associated purpose (if one has
10626 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004] ###
10628 * Avoid a race condition when CRLs are checked in a multi threaded
10629 environment. This would happen due to the reordering of the revoked
10630 entries during signature checking and serial number lookup. Now the
10631 encoding is cached and the serial number sort performed under a lock.
10632 Add new STACK function sk_is_sorted().
10636 * Add Delta CRL to the extension code.
10640 * Various fixes to s3_pkt.c so alerts are sent properly.
10642 *David Holmes <d.holmes@f5.com>*
10644 * Reduce the chances of duplicate issuer name and serial numbers (in
10645 violation of RFC3280) using the OpenSSL certificate creation utilities.
10646 This is done by creating a random 64 bit value for the initial serial
10647 number when a serial number file is created or when a self signed
10648 certificate is created using 'openssl req -x509'. The initial serial
10649 number file is created using 'openssl x509 -next_serial' in CA.pl
10650 rather than being initialized to 1.
10654 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004] ###
10656 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10657 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10659 *Joe Orton, Steve Henson*
10661 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10664 *Joe Orton, Steve Henson*
10666 * Make it possible to have multiple active certificates with the same
10667 subject in the CA index file. This is done only if the keyword
10668 'unique_subject' is set to 'no' in the main CA section (default
10669 if 'CA_default') of the configuration file. The value is saved
10670 with the database itself in a separate index attribute file,
10671 named like the index file with '.attr' appended to the name.
10675 * X509 verify fixes. Disable broken certificate workarounds when
10676 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10677 keyUsage extension present. Don't accept CRLs with unhandled critical
10678 extensions: since verify currently doesn't process CRL extensions this
10679 rejects a CRL with *any* critical extensions. Add new verify error codes
10684 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10685 A clarification of RFC2560 will require the use of OCTET STRINGs and
10686 some implementations cannot handle the current raw format. Since OpenSSL
10687 copies and compares OCSP nonces as opaque blobs without any attempt at
10688 parsing them this should not create any compatibility issues.
10692 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10693 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10694 this HMAC (and other) operations are several times slower than OpenSSL
10699 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10701 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10703 * Use the correct content when signing type "other".
10707 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003] ###
10709 * Fix various bugs revealed by running the NISCC test suite:
10711 Stop out of bounds reads in the ASN1 code when presented with
10712 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10714 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10716 If verify callback ignores invalid public key errors don't try to check
10717 certificate signature with the NULL public key.
10722 * New -ignore_err option in ocsp application to stop the server
10723 exiting on the first error in a request.
10727 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10728 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10733 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10734 extra data after the compression methods not only for TLS 1.0
10735 but also for SSL 3.0 (as required by the specification).
10737 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10739 * Change X509_certificate_type() to mark the key as exported/exportable
10740 when it's 512 *bits* long, not 512 bytes.
10744 * Change AES_cbc_encrypt() so it outputs exact multiple of
10745 blocks during encryption.
10749 * Various fixes to base64 BIO and non blocking I/O. On write
10750 flushes were not handled properly if the BIO retried. On read
10751 data was not being buffered properly and had various logic bugs.
10752 This also affects blocking I/O when the data being decoded is a
10757 * Various S/MIME bugfixes and compatibility changes:
10758 output correct application/pkcs7 MIME type if
10759 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10760 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10761 of files as .eml work). Correctly handle very long lines in MIME
10766 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003] ###
10768 * Countermeasure against the Klima-Pokorny-Rosa extension of
10769 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10770 a protocol version number mismatch like a decryption error
10771 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10775 * Turn on RSA blinding by default in the default implementation
10776 to avoid a timing attack. Applications that don't want it can call
10777 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10778 They would be ill-advised to do so in most cases.
10780 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10782 * Change RSA blinding code so that it works when the PRNG is not
10783 seeded (in this case, the secret RSA exponent is abused as
10784 an unpredictable seed -- if it is not unpredictable, there
10785 is no point in blinding anyway). Make RSA blinding thread-safe
10786 by remembering the creator's thread ID in rsa->blinding and
10787 having all other threads use local one-time blinding factors
10788 (this requires more computation than sharing rsa->blinding, but
10789 avoids excessive locking; and if an RSA object is not shared
10790 between threads, blinding will still be very fast).
10794 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10795 ENGINE as defaults for all supported algorithms irrespective of
10796 the 'flags' parameter. 'flags' is now honoured, so applications
10797 should make sure they are passing it correctly.
10801 * Target "mingw" now allows native Windows code to be generated in
10802 the Cygwin environment as well as with the MinGW compiler.
10806 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003] ###
10808 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10809 via timing by performing a MAC computation even if incorrect
10810 block cipher padding has been found. This is a countermeasure
10811 against active attacks where the attacker has to distinguish
10812 between bad padding and a MAC verification error. [CVE-2003-0078][]
10814 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10815 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10816 Martin Vuagnoux (EPFL, Ilion)*
10818 * Make the no-err option work as intended. The intention with no-err
10819 is not to have the whole error stack handling routines removed from
10820 libcrypto, it's only intended to remove all the function name and
10821 reason texts, thereby removing some of the footprint that may not
10822 be interesting if those errors aren't displayed anyway.
10824 NOTE: it's still possible for any application or module to have its
10825 own set of error texts inserted. The routines are there, just not
10826 used by default when no-err is given.
10830 * Add support for FreeBSD on IA64.
10832 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10834 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10835 Kerberos function mit_des_cbc_cksum(). Before this change,
10836 the value returned by DES_cbc_cksum() was like the one from
10837 mit_des_cbc_cksum(), except the bytes were swapped.
10839 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10841 * Allow an application to disable the automatic SSL chain building.
10842 Before this a rather primitive chain build was always performed in
10843 ssl3_output_cert_chain(): an application had no way to send the
10844 correct chain if the automatic operation produced an incorrect result.
10846 Now the chain builder is disabled if either:
10848 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10850 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10852 The reasoning behind this is that an application would not want the
10853 auto chain building to take place if extra chain certificates are
10854 present and it might also want a means of sending no additional
10855 certificates (for example the chain has two certificates and the
10860 * Add the possibility to build without the ENGINE framework.
10862 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10864 * Under Win32 gmtime() can return NULL: check return value in
10865 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10869 * DSA routines: under certain error conditions uninitialized BN objects
10870 could be freed. Solution: make sure initialization is performed early
10871 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10872 Nils Larsch <nla@trustcenter.de> via PR#459)
10876 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10877 checked on reconnect on the client side, therefore session resumption
10878 could still fail with a "ssl session id is different" error. This
10879 behaviour is masked when SSL_OP_ALL is used due to
10880 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10881 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10882 followup to PR #377.
10886 * IA-32 assembler support enhancements: unified ELF targets, support
10887 for SCO/Caldera platforms, fix for Cygwin shared build.
10891 * Add support for FreeBSD on sparc64. As a consequence, support for
10892 FreeBSD on non-x86 processors is separate from x86 processors on
10893 the config script, much like the NetBSD support.
10895 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10897 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002] ###
10899 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10902 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10903 code (06) was taken as the first octet of the session ID and the last
10904 octet was ignored consequently. As a result SSLv2 client side session
10905 caching could not have worked due to the session ID mismatch between
10907 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10912 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10913 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10918 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10919 seems that in spite of existing for more than a year, many application
10920 author have done nothing to provide the necessary callbacks, which
10921 means that this particular engine will not work properly anywhere.
10922 This is a very unfortunate situation which forces us, in the name
10923 of usability, to give the hw_ncipher.c a static lock, which is part
10925 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10926 appear in 0.9.8 or later. We EXPECT application authors to have
10927 dealt properly with this when 0.9.8 is released (unless we actually
10928 make such changes in the libcrypto locking code that changes will
10929 have to be made anyway).
10933 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
10934 octets have been read, EOF or an error occurs. Without this change
10935 some truncated ASN1 structures will not produce an error.
10939 * Disable Heimdal support, since it hasn't been fully implemented.
10940 Still give the possibility to force the use of Heimdal, but with
10941 warnings and a request that patches get sent to openssl-dev.
10945 * Add the VC-CE target, introduce the WINCE sysname, and add
10946 INSTALL.WCE and appropriate conditionals to make it build.
10948 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10950 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
10951 cygssl-x.y.z.dll, where x, y and z are the major, minor and
10952 edit numbers of the version.
10954 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10956 * Introduce safe string copy and catenation functions
10957 (BUF_strlcpy() and BUF_strlcat()).
10959 *Ben Laurie (CHATS) and Richard Levitte*
10961 * Avoid using fixed-size buffers for one-line DNs.
10963 *Ben Laurie (CHATS)*
10965 * Add BUF_MEM_grow_clean() to avoid information leakage when
10966 resizing buffers containing secrets, and use where appropriate.
10968 *Ben Laurie (CHATS)*
10970 * Avoid using fixed size buffers for configuration file location.
10972 *Ben Laurie (CHATS)*
10974 * Avoid filename truncation for various CA files.
10976 *Ben Laurie (CHATS)*
10978 * Use sizeof in preference to magic numbers.
10980 *Ben Laurie (CHATS)*
10982 * Avoid filename truncation in cert requests.
10984 *Ben Laurie (CHATS)*
10986 * Add assertions to check for (supposedly impossible) buffer
10989 *Ben Laurie (CHATS)*
10991 * Don't cache truncated DNS entries in the local cache (this could
10992 potentially lead to a spoofing attack).
10994 *Ben Laurie (CHATS)*
10996 * Fix various buffers to be large enough for hex/decimal
10997 representations in a platform independent manner.
10999 *Ben Laurie (CHATS)*
11001 * Add CRYPTO_realloc_clean() to avoid information leakage when
11002 resizing buffers containing secrets, and use where appropriate.
11004 *Ben Laurie (CHATS)*
11006 * Add BIO_indent() to avoid much slightly worrying code to do
11009 *Ben Laurie (CHATS)*
11011 * Convert sprintf()/BIO_puts() to BIO_printf().
11013 *Ben Laurie (CHATS)*
11015 * buffer_gets() could terminate with the buffer only half
11018 *Ben Laurie (CHATS)*
11020 * Add assertions to prevent user-supplied crypto functions from
11021 overflowing internal buffers by having large block sizes, etc.
11023 *Ben Laurie (CHATS)*
11025 * New OPENSSL_assert() macro (similar to assert(), but enabled
11028 *Ben Laurie (CHATS)*
11030 * Eliminate unused copy of key in RC4.
11032 *Ben Laurie (CHATS)*
11034 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11036 *Ben Laurie (CHATS)*
11038 * Fix off-by-one error in EGD path.
11040 *Ben Laurie (CHATS)*
11042 * If RANDFILE path is too long, ignore instead of truncating.
11044 *Ben Laurie (CHATS)*
11046 * Eliminate unused and incorrectly sized X.509 structure
11049 *Ben Laurie (CHATS)*
11051 * Eliminate unused and dangerous function knumber().
11053 *Ben Laurie (CHATS)*
11055 * Eliminate unused and dangerous structure, KSSL_ERR.
11057 *Ben Laurie (CHATS)*
11059 * Protect against overlong session ID context length in an encoded
11060 session object. Since these are local, this does not appear to be
11063 *Ben Laurie (CHATS)*
11065 * Change from security patch (see 0.9.6e below) that did not affect
11066 the 0.9.6 release series:
11068 Remote buffer overflow in SSL3 protocol - an attacker could
11069 supply an oversized master key in Kerberos-enabled versions.
11072 *Ben Laurie (CHATS)*
11074 * Change the SSL kerb5 codes to match RFC 2712.
11078 * Make -nameopt work fully for req and add -reqopt switch.
11080 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11082 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11084 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11086 * Make sure tests can be performed even if the corresponding algorithms
11087 have been removed entirely. This was also the last step to make
11088 OpenSSL compilable with DJGPP under all reasonable conditions.
11090 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11092 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11093 to allow version independent disabling of normally unselected ciphers,
11094 which may be activated as a side-effect of selecting a single cipher.
11096 (E.g., cipher list string "RSA" enables ciphersuites that are left
11097 out of "ALL" because they do not provide symmetric encryption.
11098 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11100 *Lutz Jaenicke, Bodo Moeller*
11102 * Add appropriate support for separate platform-dependent build
11103 directories. The recommended way to make a platform-dependent
11104 build directory is the following (tested on Linux), maybe with
11107 # Place yourself outside of the OpenSSL source tree. In
11108 # this example, the environment variable OPENSSL_SOURCE
11109 # is assumed to contain the absolute OpenSSL source directory.
11110 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11111 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11112 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11113 mkdir -p `dirname $F`
11114 ln -s $OPENSSL_SOURCE/$F $F
11117 To be absolutely sure not to disturb the source tree, a "make clean"
11118 is a good thing. If it isn't successful, don't worry about it,
11119 it probably means the source directory is very clean.
11123 * Make sure any ENGINE control commands make local copies of string
11124 pointers passed to them whenever necessary. Otherwise it is possible
11125 the caller may have overwritten (or deallocated) the original string
11126 data when a later ENGINE operation tries to use the stored values.
11128 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11130 * Improve diagnostics in file reading and command-line digests.
11132 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11134 * Add AES modes CFB and OFB to the object database. Correct an
11135 error in AES-CFB decryption.
11139 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11140 allows existing EVP_CIPHER_CTX structures to be reused after
11141 calling EVP_*Final(). This behaviour is used by encryption
11142 BIOs and some applications. This has the side effect that
11143 applications must explicitly clean up cipher contexts with
11144 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11148 * Check the values of dna and dnb in bn_mul_recursive before calling
11149 bn_mul_comba (a non zero value means the a or b arrays do not contain
11150 n2 elements) and fallback to bn_mul_normal if either is not zero.
11154 * Fix escaping of non-ASCII characters when using the -subj option
11155 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11159 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11160 form for "surname", serialNumber has no short form.
11161 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11162 therefore remove "mail" short name for "internet 7".
11163 The OID for unique identifiers in X509 certificates is
11164 x500UniqueIdentifier, not uniqueIdentifier.
11165 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11169 * Add an "init" command to the ENGINE config module and auto initialize
11170 ENGINEs. Without any "init" command the ENGINE will be initialized
11171 after all ctrl commands have been executed on it. If init=1 the
11172 ENGINE is initialized at that point (ctrls before that point are run
11173 on the uninitialized ENGINE and after on the initialized one). If
11174 init=0 then the ENGINE will not be initialized at all.
11178 * Fix the 'app_verify_callback' interface so that the user-defined
11179 argument is actually passed to the callback: In the
11180 SSL_CTX_set_cert_verify_callback() prototype, the callback
11181 declaration has been changed from
11184 int (*cb)(X509_STORE_CTX *,void *);
11185 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11186 i=s->ctx->app_verify_callback(&ctx)
11187 has been changed into
11188 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11190 To update applications using SSL_CTX_set_cert_verify_callback(),
11191 a dummy argument can be added to their callback functions.
11193 *D. K. Smetters <smetters@parc.xerox.com>*
11195 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11197 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11199 * Add and OPENSSL_LOAD_CONF define which will cause
11200 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11201 This allows older applications to transparently support certain
11202 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11203 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11204 load the config file and OPENSSL_add_all_algorithms_conf() which will
11205 always load it have also been added.
11209 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11210 Adjust NIDs and EVP layer.
11212 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11214 * Config modules support in openssl utility.
11216 Most commands now load modules from the config file,
11217 though in a few (such as version) this isn't done
11218 because it couldn't be used for anything.
11220 In the case of ca and req the config file used is
11221 the same as the utility itself: that is the -config
11222 command line option can be used to specify an
11227 * Move default behaviour from OPENSSL_config(). If appname is NULL
11228 use "openssl_conf" if filename is NULL use default openssl config file.
11232 * Add an argument to OPENSSL_config() to allow the use of an alternative
11233 config section name. Add a new flag to tolerate a missing config file
11234 and move code to CONF_modules_load_file().
11238 * Support for crypto accelerator cards from Accelerated Encryption
11239 Processing, www.aep.ie. (Use engine 'aep')
11240 The support was copied from 0.9.6c [engine] and adapted/corrected
11241 to work with the new engine framework.
11243 *AEP Inc. and Richard Levitte*
11245 * Support for SureWare crypto accelerator cards from Baltimore
11246 Technologies. (Use engine 'sureware')
11247 The support was copied from 0.9.6c [engine] and adapted
11248 to work with the new engine framework.
11252 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11253 make the newer ENGINE framework commands for the CHIL engine work.
11255 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11257 * Make it possible to produce shared libraries on ReliantUNIX.
11259 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11261 * Add the configuration target debug-linux-ppro.
11262 Make 'openssl rsa' use the general key loading routines
11263 implemented in apps.c, and make those routines able to
11264 handle the key format FORMAT_NETSCAPE and the variant
11267 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11269 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11271 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11273 * Add -keyform to rsautl, and document -engine.
11275 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11277 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11278 BIO_R_NO_SUCH_FILE error code rather than the generic
11279 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11283 * Add new functions
11284 ERR_peek_last_error
11285 ERR_peek_last_error_line
11286 ERR_peek_last_error_line_data.
11287 These are similar to
11289 ERR_peek_error_line
11290 ERR_peek_error_line_data,
11291 but report on the latest error recorded rather than the first one
11292 still in the error queue.
11294 *Ben Laurie, Bodo Moeller*
11296 * default_algorithms option in ENGINE config module. This allows things
11298 default_algorithms = ALL
11299 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11303 * Preliminary ENGINE config module.
11307 * New experimental application configuration code.
11311 * Change the AES code to follow the same name structure as all other
11312 symmetric ciphers, and behave the same way. Move everything to
11313 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11315 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11317 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11319 *Ben Laurie and Theo de Raadt*
11321 * Add option to output public keys in req command.
11323 *Massimiliano Pala madwolf@openca.org*
11325 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11326 (up to about 10% better than before for P-192 and P-224).
11330 * New functions/macros
11332 SSL_CTX_set_msg_callback(ctx, cb)
11333 SSL_CTX_set_msg_callback_arg(ctx, arg)
11334 SSL_set_msg_callback(ssl, cb)
11335 SSL_set_msg_callback_arg(ssl, arg)
11337 to request calling a callback function
11339 void cb(int write_p, int version, int content_type,
11340 const void *buf, size_t len, SSL *ssl, void *arg)
11342 whenever a protocol message has been completely received
11343 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11344 protocol version according to which the SSL library interprets
11345 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11346 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11347 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11348 specification (change_cipher_spec(20), alert(21), handshake(22)).
11349 'buf' and 'len' point to the actual message, 'ssl' to the
11350 SSL object, and 'arg' is the application-defined value set by
11351 SSL[_CTX]_set_msg_callback_arg().
11353 'openssl s_client' and 'openssl s_server' have new '-msg' options
11354 to enable a callback that displays all protocol messages.
11358 * Change the shared library support so shared libraries are built as
11359 soon as the corresponding static library is finished, and thereby get
11360 openssl and the test programs linked against the shared library.
11361 This still only happens when the keyword "shard" has been given to
11362 the configuration scripts.
11364 NOTE: shared library support is still an experimental thing, and
11365 backward binary compatibility is still not guaranteed.
11367 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11369 * Add support for Subject Information Access extension.
11371 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11373 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11374 additional bytes when new memory had to be allocated, not just
11375 when reusing an existing buffer.
11379 * New command line and configuration option 'utf8' for the req command.
11380 This allows field values to be specified as UTF8 strings.
11384 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11385 runs for the former and machine-readable output for the latter.
11389 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11390 of the e-mail address in the DN (i.e., it will go into a certificate
11391 extension only). The new configuration file option 'email_in_dn = no'
11392 has the same effect.
11394 *Massimiliano Pala madwolf@openca.org*
11396 * Change all functions with names starting with des_ to be starting
11397 with DES_ instead. Add wrappers that are compatible with libdes,
11398 but are named _ossl_old_des_*. Finally, add macros that map the
11399 des_* symbols to the corresponding _ossl_old_des_* if libdes
11400 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11401 desired, the des_* symbols will be mapped to DES_*, with one
11404 Since we provide two compatibility mappings, the user needs to
11405 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11406 compatibility is desired. The default (i.e., when that macro
11407 isn't defined) is OpenSSL 0.9.6c compatibility.
11409 There are also macros that enable and disable the support of old
11410 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11411 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11412 are defined, the default will apply: to support the old des routines.
11414 In either case, one must include openssl/des.h to get the correct
11415 definitions. Do not try to just include openssl/des_old.h, that
11418 NOTE: This is a major break of an old API into a new one. Software
11419 authors are encouraged to switch to the DES_ style functions. Some
11420 time in the future, des_old.h and the libdes compatibility functions
11421 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11422 default), and then completely removed.
11426 * Test for certificates which contain unsupported critical extensions.
11427 If such a certificate is found during a verify operation it is
11428 rejected by default: this behaviour can be overridden by either
11429 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11430 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11431 X509_supported_extension() has also been added which returns 1 if a
11432 particular extension is supported.
11436 * Modify the behaviour of EVP cipher functions in similar way to digests
11437 to retain compatibility with existing code.
11441 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11442 compatibility with existing code. In particular the 'ctx' parameter does
11443 not have to be to be initialized before the call to EVP_DigestInit() and
11444 it is tidied up after a call to EVP_DigestFinal(). New function
11445 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11446 EVP_MD_CTX_copy() changed to not require the destination to be
11447 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11448 requires the destination to be valid.
11450 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11451 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11455 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11456 so that complete 'Handshake' protocol structures are kept in memory
11457 instead of overwriting 'msg_type' and 'length' with 'body' data.
11461 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11463 *Massimo Santin via Richard Levitte*
11465 * Major restructuring to the underlying ENGINE code. This includes
11466 reduction of linker bloat, separation of pure "ENGINE" manipulation
11467 (initialisation, etc) from functionality dealing with implementations
11468 of specific crypto interfaces. This change also introduces integrated
11469 support for symmetric ciphers and digest implementations - so ENGINEs
11470 can now accelerate these by providing EVP_CIPHER and EVP_MD
11471 implementations of their own. This is detailed in crypto/engine/README
11472 as it couldn't be adequately described here. However, there are a few
11473 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11474 were changed in the original introduction of ENGINE code have now
11475 reverted back - the hooking from this code to ENGINE is now a good
11476 deal more passive and at run-time, operations deal directly with
11477 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11478 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11479 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
11480 they were not being used by the framework as there is no concept of a
11481 BIGNUM_METHOD and they could not be generalised to the new
11482 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11483 ENGINE_cpy() has been removed as it cannot be consistently defined in
11488 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11492 * Change mkdef.pl to sort symbols that get the same entry number,
11493 and make sure the automatically generated functions ERR_load_*
11494 become part of libeay.num as well.
11498 * New function SSL_renegotiate_pending(). This returns true once
11499 renegotiation has been requested (either SSL_renegotiate() call
11500 or HelloRequest/ClientHello received from the peer) and becomes
11501 false once a handshake has been completed.
11502 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11503 sends a HelloRequest, but does not ensure that a handshake takes
11504 place. SSL_renegotiate_pending() is useful for checking if the
11505 client has followed the request.)
11509 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11510 By default, clients may request session resumption even during
11511 renegotiation (if session ID contexts permit); with this option,
11512 session resumption is possible only in the first handshake.
11514 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11515 more bits available for options that should not be part of
11516 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11520 * Add some demos for certificate and certificate request creation.
11524 * Make maximum certificate chain size accepted from the peer application
11525 settable (SSL*_get/set_max_cert_list()), as proposed by
11526 "Douglas E. Engert" <deengert@anl.gov>.
11530 * Add support for shared libraries for Unixware-7
11531 (Boyd Lynn Gerber <gerberb@zenez.com>).
11535 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11536 be done prior to destruction. Use this to unload error strings from
11537 ENGINEs that load their own error strings. NB: This adds two new API
11538 functions to "get" and "set" this destroy handler in an ENGINE.
11542 * Alter all existing ENGINE implementations (except "openssl" and
11543 "openbsd") to dynamically instantiate their own error strings. This
11544 makes them more flexible to be built both as statically-linked ENGINEs
11545 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11546 Also, add stub code to each that makes building them as self-contained
11547 shared-libraries easier (see README.ENGINE).
11551 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11552 implementations into applications that are completely implemented in
11553 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11554 commands that can be used to configure what shared-library to load and
11555 to control aspects of the way it is handled. Also, made an update to
11556 the README.ENGINE file that brings its information up-to-date and
11557 provides some information and instructions on the "dynamic" ENGINE
11558 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11562 * Make it possible to unload ranges of ERR strings with a new
11563 "ERR_unload_strings" function.
11567 * Add a copy() function to EVP_MD.
11571 * Make EVP_MD routines take a context pointer instead of just the
11572 md_data void pointer.
11576 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11577 that the digest can only process a single chunk of data
11578 (typically because it is provided by a piece of
11579 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11580 is only going to provide a single chunk of data, and hence the
11581 framework needn't accumulate the data for oneshot drivers.
11585 * As with "ERR", make it possible to replace the underlying "ex_data"
11586 functions. This change also alters the storage and management of global
11587 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11588 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11589 index counters. The API functions that use this state have been changed
11590 to take a "class_index" rather than pointers to the class's local STACK
11591 and counter, and there is now an API function to dynamically create new
11592 classes. This centralisation allows us to (a) plug a lot of the
11593 thread-safety problems that existed, and (b) makes it possible to clean
11594 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11595 such data would previously have always leaked in application code and
11596 workarounds were in place to make the memory debugging turn a blind eye
11597 to it. Application code that doesn't use this new function will still
11598 leak as before, but their memory debugging output will announce it now
11599 rather than letting it slide.
11601 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11602 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11603 has a return value to indicate success or failure.
11607 * Make it possible to replace the underlying "ERR" functions such that the
11608 global state (2 LHASH tables and 2 locks) is only used by the "default"
11609 implementation. This change also adds two functions to "get" and "set"
11610 the implementation prior to it being automatically set the first time
11611 any other ERR function takes place. Ie. an application can call "get",
11612 pass the return value to a module it has just loaded, and that module
11613 can call its own "set" function using that value. This means the
11614 module's "ERR" operations will use (and modify) the error state in the
11615 application and not in its own statically linked copy of OpenSSL code.
11619 * Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
11620 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11621 the operation, and provides a more encapsulated way for external code
11622 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11623 to use these functions rather than manually incrementing the counts.
11625 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11629 * Add EVP test program.
11633 * Add symmetric cipher support to ENGINE. Expect the API to change!
11637 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11638 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11639 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11640 These allow a CRL to be built without having to access X509_CRL fields
11641 directly. Modify 'ca' application to use new functions.
11645 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11646 bug workarounds. Rollback attack detection is a security feature.
11647 The problem will only arise on OpenSSL servers when TLSv1 is not
11648 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11649 Software authors not wanting to support TLSv1 will have special reasons
11650 for their choice and can explicitly enable this option.
11652 *Bodo Moeller, Lutz Jaenicke*
11654 * Rationalise EVP so it can be extended: don't include a union of
11655 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11656 (similar to those existing for EVP_CIPHER_CTX).
11661 EVP_MD_CTX_init(&md); /* new function call */
11662 EVP_DigestInit(&md, EVP_sha1());
11663 EVP_DigestUpdate(&md, in, len);
11664 EVP_DigestFinal(&md, out, NULL);
11665 EVP_MD_CTX_cleanup(&md); /* new function call */
11670 * Make DES key schedule conform to the usual scheme, as well as
11671 correcting its structure. This means that calls to DES functions
11672 now have to pass a pointer to a des_key_schedule instead of a
11673 plain des_key_schedule (which was actually always a pointer
11676 des_key_schedule ks;
11678 des_set_key_checked(..., &ks);
11679 des_ncbc_encrypt(..., &ks, ...);
11681 (Note that a later change renames 'des_...' into 'DES_...'.)
11685 * Initial reduction of linker bloat: the use of some functions, such as
11686 PEM causes large amounts of unused functions to be linked in due to
11687 poor organisation. For example pem_all.c contains every PEM function
11688 which has a knock on effect of linking in large amounts of (unused)
11689 ASN1 code. Grouping together similar functions and splitting unrelated
11690 functions prevents this.
11694 * Cleanup of EVP macros.
11698 * Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
11699 correct _ecb suffix.
11703 * Add initial OCSP responder support to ocsp application. The
11704 revocation information is handled using the text based index
11705 use by the ca application. The responder can either handle
11706 requests generated internally, supplied in files (for example
11707 via a CGI script) or using an internal minimal server.
11711 * Add configuration choices to get zlib compression for TLS.
11715 * Changes to Kerberos SSL for RFC 2712 compliance:
11716 1. Implemented real KerberosWrapper, instead of just using
11717 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11718 2. Implemented optional authenticator field of KerberosWrapper.
11720 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11721 and authenticator structs; see crypto/krb5/.
11723 Generalized Kerberos calls to support multiple Kerberos libraries.
11724 *Vern Staats <staatsvr@asc.hpc.mil>,
11725 Jeffrey Altman <jaltman@columbia.edu>
11726 via Richard Levitte*
11728 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11729 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11730 values for each of the key sizes rather than having just
11731 parameters (and 'speed' generating keys each time).
11735 * Speed up EVP routines.
11738 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11739 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11740 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11741 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11743 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11744 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11745 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11748 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11750 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11754 * Added the OS2-EMX target.
11756 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11758 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11759 to support NCONF routines in extension code. New function CONF_set_nconf()
11760 to allow functions which take an NCONF to also handle the old LHASH
11761 structure: this means that the old CONF compatible routines can be
11762 retained (in particular wrt extensions) without having to duplicate the
11763 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11767 * Enhance the general user interface with mechanisms for inner control
11768 and with possibilities to have yes/no kind of prompts.
11772 * Change all calls to low level digest routines in the library and
11773 applications to use EVP. Add missing calls to HMAC_cleanup() and
11774 don't assume HMAC_CTX can be copied using memcpy().
11776 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11778 * Add the possibility to control engines through control names but with
11779 arbitrary arguments instead of just a string.
11780 Change the key loaders to take a UI_METHOD instead of a callback
11781 function pointer. NOTE: this breaks binary compatibility with earlier
11782 versions of OpenSSL [engine].
11783 Adapt the nCipher code for these new conditions and add a card insertion
11788 * Enhance the general user interface with mechanisms to better support
11789 dialog box interfaces, application-defined prompts, the possibility
11790 to use defaults (for example default passwords from somewhere else)
11791 and interrupts/cancellations.
11795 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11796 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11800 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11801 tidy up some unnecessarily weird code in 'sk_new()').
11803 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11805 * Change the key loading routines for ENGINEs to use the same kind
11806 callback (pem_password_cb) as all other routines that need this
11811 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11812 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11813 than this minimum value is recommended.
11817 * New random seeder for OpenVMS, using the system process statistics
11818 that are easily reachable.
11822 * Windows apparently can't transparently handle global
11823 variables defined in DLLs. Initialisations such as:
11825 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11827 won't compile. This is used by the any applications that need to
11828 declare their own ASN1 modules. This was fixed by adding the option
11829 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11830 needed for static libraries under Win32.
11834 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11835 setting of purpose and trust fields. New X509_STORE trust and
11836 purpose functions and tidy up setting in other SSL functions.
11840 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11841 structure. These are inherited by X509_STORE_CTX when it is
11842 initialised. This allows various defaults to be set in the
11843 X509_STORE structure (such as flags for CRL checking and custom
11844 purpose or trust settings) for functions which only use X509_STORE_CTX
11845 internally such as S/MIME.
11847 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11848 trust settings if they are not set in X509_STORE. This allows X509_STORE
11849 purposes and trust (in S/MIME for example) to override any set by default.
11851 Add command line options for CRL checking to smime, s_client and s_server
11856 * Initial CRL based revocation checking. If the CRL checking flag(s)
11857 are set then the CRL is looked up in the X509_STORE structure and
11858 its validity and signature checked, then if the certificate is found
11859 in the CRL the verify fails with a revoked error.
11861 Various new CRL related callbacks added to X509_STORE_CTX structure.
11863 Command line options added to 'verify' application to support this.
11865 This needs some additional work, such as being able to handle multiple
11866 CRLs with different times, extension based lookup (rather than just
11867 by subject name) and ultimately more complete V2 CRL extension
11872 * Add a general user interface API (crypto/ui/). This is designed
11873 to replace things like des_read_password and friends (backward
11874 compatibility functions using this new API are provided).
11875 The purpose is to remove prompting functions from the DES code
11876 section as well as provide for prompting through dialog boxes in
11877 a window system and the like.
11881 * Add "ex_data" support to ENGINE so implementations can add state at a
11882 per-structure level rather than having to store it globally.
11886 * Make it possible for ENGINE structures to be copied when retrieved by
11887 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11888 This causes the "original" ENGINE structure to act like a template,
11889 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11890 operational state can be localised to each ENGINE structure, despite the
11891 fact they all share the same "methods". New ENGINE structures returned in
11892 this case have no functional references and the return value is the single
11893 structural reference. This matches the single structural reference returned
11894 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11899 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11900 needs to match any other type at all we need to manually clear the
11905 * Changes to the "openssl engine" utility to include;
11906 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11907 about an ENGINE's available control commands.
11908 - executing control commands from command line arguments using the
11909 '-pre' and '-post' switches. '-post' is only used if '-t' is
11910 specified and the ENGINE is successfully initialised. The syntax for
11911 the individual commands are colon-separated, for example;
11912 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11916 * New dynamic control command support for ENGINEs. ENGINEs can now
11917 declare their own commands (numbers), names (strings), descriptions,
11918 and input types for run-time discovery by calling applications. A
11919 subset of these commands are implicitly classed as "executable"
11920 depending on their input type, and only these can be invoked through
11921 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11922 can be based on user input, config files, etc). The distinction is
11923 that "executable" commands cannot return anything other than a boolean
11924 result and can only support numeric or string input, whereas some
11925 discoverable commands may only be for direct use through
11926 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11927 pointers, or other custom uses. The "executable" commands are to
11928 support parameterisations of ENGINE behaviour that can be
11929 unambiguously defined by ENGINEs and used consistently across any
11930 OpenSSL-based application. Commands have been added to all the
11931 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
11932 control over shared-library paths without source code alterations.
11936 * Changed all ENGINE implementations to dynamically allocate their
11937 ENGINEs rather than declaring them statically. Apart from this being
11938 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
11939 this also allows the implementations to compile without using the
11940 internal engine_int.h header.
11944 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
11945 'const' value. Any code that should be able to modify a RAND_METHOD
11946 should already have non-const pointers to it (ie. they should only
11947 modify their own ones).
11951 * Made a variety of little tweaks to the ENGINE code.
11952 - "atalla" and "ubsec" string definitions were moved from header files
11953 to C code. "nuron" string definitions were placed in variables
11954 rather than hard-coded - allowing parameterisation of these values
11955 later on via ctrl() commands.
11956 - Removed unused "#if 0"'d code.
11957 - Fixed engine list iteration code so it uses ENGINE_free() to release
11958 structural references.
11959 - Constified the RAND_METHOD element of ENGINE structures.
11960 - Constified various get/set functions as appropriate and added
11961 missing functions (including a catch-all ENGINE_cpy that duplicates
11962 all ENGINE values onto a new ENGINE except reference counts/state).
11963 - Removed NULL parameter checks in get/set functions. Setting a method
11964 or function to NULL is a way of cancelling out a previously set
11965 value. Passing a NULL ENGINE parameter is just plain stupid anyway
11966 and doesn't justify the extra error symbols and code.
11967 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
11968 flags from engine_int.h to engine.h.
11969 - Changed prototypes for ENGINE handler functions (init(), finish(),
11970 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
11974 * Implement binary inversion algorithm for BN_mod_inverse in addition
11975 to the algorithm using long division. The binary algorithm can be
11976 used only if the modulus is odd. On 32-bit systems, it is faster
11977 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
11978 roughly 5-15% for 256-bit moduli), so we use it only for moduli
11979 up to 450 bits. In 64-bit environments, the binary algorithm
11980 appears to be advantageous for much longer moduli; here we use it
11981 for moduli up to 2048 bits.
11985 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
11986 could not support the combine flag in choice fields.
11990 * Add a 'copy_extensions' option to the 'ca' utility. This copies
11991 extensions from a certificate request to the certificate.
11995 * Allow multiple 'certopt' and 'nameopt' options to be separated
11996 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
11997 file: this allows the display of the certificate about to be
11998 signed to be customised, to allow certain fields to be included
11999 or excluded and extension details. The old system didn't display
12000 multicharacter strings properly, omitted fields not in the policy
12001 and couldn't display additional details such as extensions.
12005 * Function EC_POINTs_mul for multiple scalar multiplication
12006 of an arbitrary number of elliptic curve points
12007 \sum scalars[i]*points[i],
12008 optionally including the generator defined for the EC_GROUP:
12009 scalar*generator + \sum scalars[i]*points[i].
12011 EC_POINT_mul is a simple wrapper function for the typical case
12012 that the point list has just one item (besides the optional
12017 * First EC_METHODs for curves over GF(p):
12019 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12020 operations and provides various method functions that can also
12021 operate with faster implementations of modular arithmetic.
12023 EC_GFp_mont_method() reuses most functions that are part of
12024 EC_GFp_simple_method, but uses Montgomery arithmetic.
12026 *Bodo Moeller; point addition and point doubling
12027 implementation directly derived from source code provided by
12028 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12030 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12031 crypto/ec/ec_lib.c):
12033 Curves are EC_GROUP objects (with an optional group generator)
12034 based on EC_METHODs that are built into the library.
12036 Points are EC_POINT objects based on EC_GROUP objects.
12038 Most of the framework would be able to handle curves over arbitrary
12039 finite fields, but as there are no obvious types for fields other
12040 than GF(p), some functions are limited to that for now.
12044 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12045 that the file contains a complete HTTP response.
12049 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12050 change the def and num file printf format specifier from "%-40sXXX"
12051 to "%-39s XXX". The latter will always guarantee a space after the
12052 field while the former will cause them to run together if the field
12053 is 40 of more characters long.
12057 * Constify the cipher and digest 'method' functions and structures
12058 and modify related functions to take constant EVP_MD and EVP_CIPHER
12063 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12064 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12068 * Modify EVP_Digest*() routines so they now return values. Although the
12069 internal software routines can never fail additional hardware versions
12074 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12076 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12077 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12080 ERR_R_NESTED_ASN1_ERROR
12082 ERR_R_MISSING_ASN1_EOS
12083 were 4 .. 9, conflicting with
12084 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12086 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12087 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12089 Add new error code 'ERR_R_INTERNAL_ERROR'.
12093 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12098 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12099 sets the subject name for a new request or supersedes the
12100 subject name in a given request. Formats that can be parsed are
12101 'CN=Some Name, OU=myOU, C=IT'
12103 'CN=Some Name/OU=myOU/C=IT'.
12105 Add options '-batch' and '-verbose' to 'openssl req'.
12107 *Massimiliano Pala <madwolf@hackmasters.net>*
12109 * Introduce the possibility to access global variables through
12110 functions on platform were that's the best way to handle exporting
12111 global variables in shared libraries. To enable this functionality,
12112 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12113 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12114 is normally done by Configure or something similar).
12116 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12117 in the source file (foo.c) like this:
12119 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12120 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12122 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12123 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12125 OPENSSL_DECLARE_GLOBAL(int,foo);
12126 #define foo OPENSSL_GLOBAL_REF(foo)
12127 OPENSSL_DECLARE_GLOBAL(double,bar);
12128 #define bar OPENSSL_GLOBAL_REF(bar)
12130 The #defines are very important, and therefore so is including the
12131 header file everywhere where the defined globals are used.
12133 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12134 of ASN.1 items, but that structure is a bit different.
12136 The largest change is in util/mkdef.pl which has been enhanced with
12137 better and easier to understand logic to choose which symbols should
12138 go into the Windows .def files as well as a number of fixes and code
12139 cleanup (among others, algorithm keywords are now sorted
12140 lexicographically to avoid constant rewrites).
12144 * In BN_div() keep a copy of the sign of 'num' before writing the
12145 result to 'rm' because if rm==num the value will be overwritten
12146 and produce the wrong result if 'num' is negative: this caused
12147 problems with BN_mod() and BN_nnmod().
12151 * Function OCSP_request_verify(). This checks the signature on an
12152 OCSP request and verifies the signer certificate. The signer
12153 certificate is just checked for a generic purpose and OCSP request
12158 * Add OCSP_check_validity() function to check the validity of OCSP
12159 responses. OCSP responses are prepared in real time and may only
12160 be a few seconds old. Simply checking that the current time lies
12161 between thisUpdate and nextUpdate max reject otherwise valid responses
12162 caused by either OCSP responder or client clock inaccuracy. Instead
12163 we allow thisUpdate and nextUpdate to fall within a certain period of
12164 the current time. The age of the response can also optionally be
12165 checked. Two new options -validity_period and -status_age added to
12170 * If signature or public key algorithm is unrecognized print out its
12171 OID rather that just UNKNOWN.
12175 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12176 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12177 ID to be generated from the issuer certificate alone which can then be
12178 passed to OCSP_id_issuer_cmp().
12182 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12183 ASN1 modules to export functions returning ASN1_ITEM pointers
12184 instead of the ASN1_ITEM structures themselves. This adds several
12185 new macros which allow the underlying ASN1 function/structure to
12186 be accessed transparently. As a result code should not use ASN1_ITEM
12187 references directly (such as &X509_it) but instead use the relevant
12188 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12189 use of the new ASN1 code on platforms where exporting structures
12190 is problematical (for example in shared libraries) but exporting
12191 functions returning pointers to structures is not.
12195 * Add support for overriding the generation of SSL/TLS session IDs.
12196 These callbacks can be registered either in an SSL_CTX or per SSL.
12197 The purpose of this is to allow applications to control, if they wish,
12198 the arbitrary values chosen for use as session IDs, particularly as it
12199 can be useful for session caching in multiple-server environments. A
12200 command-line switch for testing this (and any client code that wishes
12201 to use such a feature) has been added to "s_server".
12203 *Geoff Thorpe, Lutz Jaenicke*
12205 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12206 of the form '#if defined(...) || defined(...) || ...' and
12207 '#if !defined(...) && !defined(...) && ...'. This also avoids
12208 the growing number of special cases it was previously handling.
12212 * Make all configuration macros available for application by making
12213 sure they are available in opensslconf.h, by giving them names starting
12214 with "OPENSSL_" to avoid conflicts with other packages and by making
12215 sure e_os2.h will cover all platform-specific cases together with
12217 Additionally, it is now possible to define configuration/platform-
12218 specific names (called "system identities"). In the C code, these
12219 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
12220 macro with the name beginning with "OPENSSL_SYS_", which is determined
12221 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
12226 * New option -set_serial to 'req' and 'x509' this allows the serial
12227 number to use to be specified on the command line. Previously self
12228 signed certificates were hard coded with serial number 0 and the
12229 CA options of 'x509' had to use a serial number in a file which was
12234 * New options to 'ca' utility to support V2 CRL entry extensions.
12235 Currently CRL reason, invalidity date and hold instruction are
12236 supported. Add new CRL extensions to V3 code and some new objects.
12240 * New function EVP_CIPHER_CTX_set_padding() this is used to
12241 disable standard block padding (aka PKCS#5 padding) in the EVP
12242 API, which was previously mandatory. This means that the data is
12243 not padded in any way and so the total length much be a multiple
12244 of the block size, otherwise an error occurs.
12248 * Initial (incomplete) OCSP SSL support.
12252 * New function OCSP_parse_url(). This splits up a URL into its host,
12253 port and path components: primarily to parse OCSP URLs. New -url
12254 option to ocsp utility.
12258 * New nonce behavior. The return value of OCSP_check_nonce() now
12259 reflects the various checks performed. Applications can decide
12260 whether to tolerate certain situations such as an absent nonce
12261 in a response when one was present in a request: the ocsp application
12262 just prints out a warning. New function OCSP_add1_basic_nonce()
12263 this is to allow responders to include a nonce in a response even if
12264 the request is nonce-less.
12268 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12269 skipped when using openssl x509 multiple times on a single input file,
12270 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
12274 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12275 set string type: to handle setting ASN1_TIME structures. Fix ca
12276 utility to correctly initialize revocation date of CRLs.
12280 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12281 the clients preferred ciphersuites and rather use its own preferences.
12282 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12283 Internet Explorer by ensuring unchanged hash method during stepup.
12284 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12288 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12289 to aes and add a new 'exist' option to print out symbols that don't
12294 * Additional options to ocsp utility to allow flags to be set and
12295 additional certificates supplied.
12299 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12300 OCSP client a number of certificate to only verify the response
12305 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12306 handle the new API. Currently only ECB, CBC modes supported. Add new
12309 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12310 Encryption Standard (AES) Ciphersuites for Transport Layer
12311 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12312 not enabled by default and were not part of the "ALL" ciphersuite
12313 alias because they were not yet official; they could be
12314 explicitly requested by specifying the "AESdraft" ciphersuite
12315 group alias. In the final release of OpenSSL 0.9.7, the group
12316 alias is called "AES" and is part of "ALL".)
12318 *Ben Laurie, Steve Henson, Bodo Moeller*
12320 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12321 request to response.
12325 * Functions for OCSP responders. OCSP_request_onereq_count(),
12326 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12327 extract information from a certificate request. OCSP_response_create()
12328 creates a response and optionally adds a basic response structure.
12329 OCSP_basic_add1_status() adds a complete single response to a basic
12330 response and returns the OCSP_SINGLERESP structure just added (to allow
12331 extensions to be included for example). OCSP_basic_add1_cert() adds a
12332 certificate to a basic response and OCSP_basic_sign() signs a basic
12333 response with various flags. New helper functions ASN1_TIME_check()
12334 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12335 (converts ASN1_TIME to GeneralizedTime).
12339 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12340 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12341 structure from a certificate. X509_pubkey_digest() digests the public_key
12342 contents: this is used in various key identifiers.
12346 * Make sk_sort() tolerate a NULL argument.
12348 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12350 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12351 passed by the function are trusted implicitly. If any of them signed the
12352 response then it is assumed to be valid and is not verified.
12356 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12357 to data. This was previously part of the PKCS7 ASN1 code. This
12358 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12359 *Steve Henson, reported by Kenneth R. Robinette
12360 <support@securenetterm.com>*
12362 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12363 routines: without these tracing memory leaks is very painful.
12364 Fix leaks in PKCS12 and PKCS7 routines.
12368 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12369 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12370 effectively meant GeneralizedTime would never be used. Now it
12371 is initialised to -1 but X509_time_adj() now has to check the value
12372 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12373 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12374 *Steve Henson, reported by Kenneth R. Robinette
12375 <support@securenetterm.com>*
12377 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12378 result in a zero length in the ASN1_INTEGER structure which was
12379 not consistent with the structure when d2i_ASN1_INTEGER() was used
12380 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12381 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12382 where it did not print out a minus for negative ASN1_INTEGER.
12386 * Add summary printout to ocsp utility. The various functions which
12387 convert status values to strings have been renamed to:
12388 OCSP_response_status_str(), OCSP_cert_status_str() and
12389 OCSP_crl_reason_str() and are no longer static. New options
12390 to verify nonce values and to disable verification. OCSP response
12391 printout format cleaned up.
12395 * Add additional OCSP certificate checks. These are those specified
12396 in RFC2560. This consists of two separate checks: the CA of the
12397 certificate being checked must either be the OCSP signer certificate
12398 or the issuer of the OCSP signer certificate. In the latter case the
12399 OCSP signer certificate must contain the OCSP signing extended key
12400 usage. This check is performed by attempting to match the OCSP
12401 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12402 in the OCSP_CERTID structures of the response.
12406 * Initial OCSP certificate verification added to OCSP_basic_verify()
12407 and related routines. This uses the standard OpenSSL certificate
12408 verify routines to perform initial checks (just CA validity) and
12409 to obtain the certificate chain. Then additional checks will be
12410 performed on the chain. Currently the root CA is checked to see
12411 if it is explicitly trusted for OCSP signing. This is used to set
12412 a root CA as a global signing root: that is any certificate that
12413 chains to that CA is an acceptable OCSP signing certificate.
12417 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12418 extensions from a separate configuration file.
12419 As when reading extensions from the main configuration file,
12420 the '-extensions ...' option may be used for specifying the
12423 *Massimiliano Pala <madwolf@comune.modena.it>*
12425 * New OCSP utility. Allows OCSP requests to be generated or
12426 read. The request can be sent to a responder and the output
12427 parsed, outputted or printed in text form. Not complete yet:
12428 still needs to check the OCSP response validity.
12432 * New subcommands for 'openssl ca':
12433 'openssl ca -status <serial>' prints the status of the cert with
12434 the given serial number (according to the index file).
12435 'openssl ca -updatedb' updates the expiry status of certificates
12438 *Massimiliano Pala <madwolf@comune.modena.it>*
12440 * New '-newreq-nodes' command option to CA.pl. This is like
12441 '-newreq', but calls 'openssl req' with the '-nodes' option
12442 so that the resulting key is not encrypted.
12444 *Damien Miller <djm@mindrot.org>*
12446 * New configuration for the GNU Hurd.
12448 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12450 * Initial code to implement OCSP basic response verify. This
12451 is currently incomplete. Currently just finds the signer's
12452 certificate and verifies the signature on the response.
12456 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12457 value of OPENSSLDIR. This is available via the new '-d' option
12458 to 'openssl version', and is also included in 'openssl version -a'.
12462 * Allowing defining memory allocation callbacks that will be given
12463 file name and line number information in additional arguments
12464 (a const char* and an int). The basic functionality remains, as
12465 well as the original possibility to just replace malloc(),
12466 realloc() and free() by functions that do not know about these
12467 additional arguments. To register and find out the current
12468 settings for extended allocation functions, the following
12469 functions are provided:
12471 CRYPTO_set_mem_ex_functions
12472 CRYPTO_set_locked_mem_ex_functions
12473 CRYPTO_get_mem_ex_functions
12474 CRYPTO_get_locked_mem_ex_functions
12476 These work the same way as CRYPTO_set_mem_functions and friends.
12477 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
12478 extended allocation function is enabled.
12479 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
12480 a conventional allocation function is enabled.
12482 *Richard Levitte, Bodo Moeller*
12484 * Finish off removing the remaining LHASH function pointer casts.
12485 There should no longer be any prototype-casting required when using
12486 the LHASH abstraction, and any casts that remain are "bugs". See
12487 the callback types and macros at the head of lhash.h for details
12488 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12492 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12493 If /dev/[u]random devices are not available or do not return enough
12494 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12496 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12497 /etc/entropy will be queried once each in this sequence, querying stops
12498 when enough entropy was collected without querying more sockets.
12502 * Change the Unix RAND_poll() variant to be able to poll several
12503 random devices, as specified by DEVRANDOM, until a sufficient amount
12504 of data has been collected. We spend at most 10 ms on each file
12505 (select timeout) and read in non-blocking mode. DEVRANDOM now
12506 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12507 (previously it was just the string "/dev/urandom"), so on typical
12508 platforms the 10 ms delay will never occur.
12509 Also separate out the Unix variant to its own file, rand_unix.c.
12510 For VMS, there's a currently-empty rand_vms.c.
12514 * Move OCSP client related routines to ocsp_cl.c. These
12515 provide utility functions which an application needing
12516 to issue a request to an OCSP responder and analyse the
12517 response will typically need: as opposed to those which an
12518 OCSP responder itself would need which will be added later.
12520 OCSP_request_sign() signs an OCSP request with an API similar
12521 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12522 response. OCSP_response_get1_basic() extracts basic response
12523 from response. OCSP_resp_find_status(): finds and extracts status
12524 information from an OCSP_CERTID structure (which will be created
12525 when the request structure is built). These are built from lower
12526 level functions which work on OCSP_SINGLERESP structures but
12527 won't normally be used unless the application wishes to examine
12528 extensions in the OCSP response for example.
12530 Replace nonce routines with a pair of functions.
12531 OCSP_request_add1_nonce() adds a nonce value and optionally
12532 generates a random value. OCSP_check_nonce() checks the
12533 validity of the nonce in an OCSP response.
12537 * Change function OCSP_request_add() to OCSP_request_add0_id().
12538 This doesn't copy the supplied OCSP_CERTID and avoids the
12539 need to free up the newly created id. Change return type
12540 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12541 This can then be used to add extensions to the request.
12542 Deleted OCSP_request_new(), since most of its functionality
12543 is now in OCSP_REQUEST_new() (and the case insensitive name
12544 clash) apart from the ability to set the request name which
12545 will be added elsewhere.
12549 * Update OCSP API. Remove obsolete extensions argument from
12550 various functions. Extensions are now handled using the new
12551 OCSP extension code. New simple OCSP HTTP function which
12552 can be used to send requests and parse the response.
12556 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12557 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12558 uses the special reorder version of SET OF to sort the attributes
12559 and reorder them to match the encoded order. This resolves a long
12560 standing problem: a verify on a PKCS7 structure just after signing
12561 it used to fail because the attribute order did not match the
12562 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12563 it uses the received order. This is necessary to tolerate some broken
12564 software that does not order SET OF. This is handled by encoding
12565 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12566 to produce the required SET OF.
12570 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12571 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12572 files to get correct declarations of the ASN.1 item variables.
12576 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12577 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12578 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12579 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12580 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12581 ASN1_ITEM and no wrapper functions.
12585 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12586 replace the old function pointer based I/O routines. Change most of
12587 the *_d2i_bio() and *_d2i_fp() functions to use these.
12591 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12592 lines, recognize more "algorithms" that can be deselected, and make
12593 it complain about algorithm deselection that isn't recognised.
12597 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12598 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12599 to use new functions. Add NO_ASN1_OLD which can be set to remove
12600 some old style ASN1 functions: this can be used to determine if old
12601 code will still work when these eventually go away.
12605 * New extension functions for OCSP structures, these follow the
12606 same conventions as certificates and CRLs.
12610 * New function X509V3_add1_i2d(). This automatically encodes and
12611 adds an extension. Its behaviour can be customised with various
12612 flags to append, replace or delete. Various wrappers added for
12613 certificates and CRLs.
12617 * Fix to avoid calling the underlying ASN1 print routine when
12618 an extension cannot be parsed. Correct a typo in the
12619 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12623 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12624 entries for variables.
12628 * Add functionality to apps/openssl.c for detecting locking
12629 problems: As the program is single-threaded, all we have
12630 to do is register a locking callback using an array for
12631 storing which locks are currently held by the program.
12635 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12636 SSL_get_ex_data_X509_STORE_idx(), which is used in
12637 ssl_verify_cert_chain() and thus can be called at any time
12638 during TLS/SSL handshakes so that thread-safety is essential.
12639 Unfortunately, the ex_data design is not at all suited
12640 for multi-threaded use, so it probably should be abolished.
12644 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12646 *Broadcom, tweaked and integrated by Geoff Thorpe*
12648 * Move common extension printing code to new function
12649 X509V3_print_extensions(). Reorganise OCSP print routines and
12650 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12654 * New function X509_signature_print() to remove duplication in some
12659 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12660 set (this was treated exactly the same as SET OF previously). This
12661 is used to reorder the STACK representing the structure to match the
12662 encoding. This will be used to get round a problem where a PKCS7
12663 structure which was signed could not be verified because the STACK
12664 order did not reflect the encoded order.
12668 * Reimplement the OCSP ASN1 module using the new code.
12672 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12673 for its ASN1 operations. The old style function pointers still exist
12674 for now but they will eventually go away.
12678 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12679 completely replaces the old ASN1 functionality with a table driven
12680 encoder and decoder which interprets an ASN1_ITEM structure describing
12681 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12682 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12683 has also been converted to the new form.
12687 * Change BN_mod_exp_recp so that negative moduli are tolerated
12688 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12689 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12690 for negative moduli.
12694 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12695 of not touching the result's sign bit.
12699 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12704 * Changed the LHASH code to use prototypes for callbacks, and created
12705 macros to declare and implement thin (optionally static) functions
12706 that provide type-safety and avoid function pointer casting for the
12707 type-specific callbacks.
12711 * Added Kerberos Cipher Suites to be used with TLS, as written in
12713 *Veers Staats <staatsvr@asc.hpc.mil>,
12714 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12716 * Reformat the FAQ so the different questions and answers can be divided
12717 in sections depending on the subject.
12721 * Have the zlib compression code load ZLIB.DLL dynamically under
12726 * New function BN_mod_sqrt for computing square roots modulo a prime
12727 (using the probabilistic Tonelli-Shanks algorithm unless
12728 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12729 be handled deterministically).
12731 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12733 * Make BN_mod_inverse faster by explicitly handling small quotients
12734 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12735 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12739 * New function BN_kronecker.
12743 * Fix BN_gcd so that it works on negative inputs; the result is
12744 positive unless both parameters are zero.
12745 Previously something reasonably close to an infinite loop was
12746 possible because numbers could be growing instead of shrinking
12747 in the implementation of Euclid's algorithm.
12751 * Fix BN_is_word() and BN_is_one() macros to take into account the
12752 sign of the number in question.
12754 Fix BN_is_word(a,w) to work correctly for w == 0.
12756 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12757 because its test if the absolute value of 'a' equals 'w'.
12758 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12759 it exists mostly for use in the implementations of BN_is_zero(),
12760 BN_is_one(), and BN_is_word().
12764 * New function BN_swap.
12768 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12769 the exponentiation functions are more likely to produce reasonable
12770 results on negative inputs.
12774 * Change BN_mod_mul so that the result is always non-negative.
12775 Previously, it could be negative if one of the factors was negative;
12776 I don't think anyone really wanted that behaviour.
12780 * Move BN_mod_... functions into new file crypto/bn/bn_mod.c
12781 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12782 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12783 and add new functions:
12792 BN_mod_lshift1_quick
12794 BN_mod_lshift_quick
12796 These functions always generate non-negative results.
12798 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12799 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12801 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12802 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12803 be reduced modulo m.
12805 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12808 The following entry accidentally appeared in the CHANGES file
12809 distributed with OpenSSL 0.9.7. The modifications described in
12810 it do *not* apply to OpenSSL 0.9.7.
12812 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12813 was actually never needed) and in BN_mul(). The removal in BN_mul()
12814 required a small change in bn_mul_part_recursive() and the addition
12815 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12816 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12817 bn_sub_words() and bn_add_words() except they take arrays with
12823 * In 'openssl passwd', verify passwords read from the terminal
12824 unless the '-salt' option is used (which usually means that
12825 verification would just waste user's time since the resulting
12826 hash is going to be compared with some given password hash)
12827 or the new '-noverify' option is used.
12829 This is an incompatible change, but it does not affect
12830 non-interactive use of 'openssl passwd' (passwords on the command
12831 line, '-stdin' option, '-in ...' option) and thus should not
12832 cause any problems.
12836 * Remove all references to RSAref, since there's no more need for it.
12840 * Make DSO load along a path given through an environment variable
12841 (SHLIB_PATH) with shl_load().
12845 * Constify the ENGINE code as a result of BIGNUM constification.
12846 Also constify the RSA code and most things related to it. In a
12847 few places, most notable in the depth of the ASN.1 code, ugly
12848 casts back to non-const were required (to be solved at a later
12853 * Make it so the openssl application has all engines loaded by default.
12857 * Constify the BIGNUM routines a little more.
12861 * Add the following functions:
12863 ENGINE_load_cswift()
12865 ENGINE_load_atalla()
12866 ENGINE_load_nuron()
12867 ENGINE_load_builtin_engines()
12869 That way, an application can itself choose if external engines that
12870 are built-in in OpenSSL shall ever be used or not. The benefit is
12871 that applications won't have to be linked with libdl or other dso
12872 libraries unless it's really needed.
12874 Changed 'openssl engine' to load all engines on demand.
12875 Changed the engine header files to avoid the duplication of some
12876 declarations (they differed!).
12880 * 'openssl engine' can now list capabilities.
12884 * Better error reporting in 'openssl engine'.
12888 * Never call load_dh_param(NULL) in s_server.
12892 * Add engine application. It can currently list engines by name and
12893 identity, and test if they are actually available.
12897 * Improve RPM specification file by forcing symbolic linking and making
12898 sure the installed documentation is also owned by root.root.
12900 *Damien Miller <djm@mindrot.org>*
12902 * Give the OpenSSL applications more possibilities to make use of
12903 keys (public as well as private) handled by engines.
12907 * Add OCSP code that comes from CertCo.
12911 * Add VMS support for the Rijndael code.
12915 * Added untested support for Nuron crypto accelerator.
12919 * Add support for external cryptographic devices. This code was
12920 previously distributed separately as the "engine" branch.
12922 *Geoff Thorpe, Richard Levitte*
12924 * Rework the filename-translation in the DSO code. It is now possible to
12925 have far greater control over how a "name" is turned into a filename
12926 depending on the operating environment and any oddities about the
12927 different shared library filenames on each system.
12931 * Support threads on FreeBSD-elf in Configure.
12935 * Fix for SHA1 assembly problem with MASM: it produces
12936 warnings about corrupt line number information when assembling
12937 with debugging information. This is caused by the overlapping
12940 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
12943 NCONF_get_number() has no error checking at all. As a replacement,
12944 NCONF_get_number_e() is defined (_e for "error checking") and is
12945 promoted strongly. The old NCONF_get_number is kept around for
12946 binary backward compatibility.
12947 Make it possible for methods to load from something other than a BIO,
12948 by providing a function pointer that is given a name instead of a BIO.
12949 For example, this could be used to load configuration data from an
12954 * Fix for non blocking accept BIOs. Added new I/O special reason
12955 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
12956 with non blocking I/O was not possible because no retry code was
12957 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
12962 * Added the beginnings of Rijndael support.
12966 * Fix for bug in DirectoryString mask setting. Add support for
12967 X509_NAME_print_ex() in 'req' and X509_print_ex() function
12968 to allow certificate printing to more controllable, additional
12969 'certopt' option to 'x509' to allow new printing options to be
12974 * Clean old EAY MD5 hack from e_os.h.
12978 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004] ###
12980 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12981 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
12983 *Joe Orton, Steve Henson*
12985 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003] ###
12987 * Fix additional bug revealed by the NISCC test suite:
12989 Stop bug triggering large recursion when presented with
12990 certain ASN.1 tags [CVE-2003-0851][]
12994 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003] ###
12996 * Fix various bugs revealed by running the NISCC test suite:
12998 Stop out of bounds reads in the ASN1 code when presented with
12999 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13001 If verify callback ignores invalid public key errors don't try to check
13002 certificate signature with the NULL public key.
13007 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13008 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13013 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13014 extra data after the compression methods not only for TLS 1.0
13015 but also for SSL 3.0 (as required by the specification).
13017 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13019 * Change X509_certificate_type() to mark the key as exported/exportable
13020 when it's 512 *bits* long, not 512 bytes.
13024 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003] ###
13026 * Countermeasure against the Klima-Pokorny-Rosa extension of
13027 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13028 a protocol version number mismatch like a decryption error
13029 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13033 * Turn on RSA blinding by default in the default implementation
13034 to avoid a timing attack. Applications that don't want it can call
13035 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13036 They would be ill-advised to do so in most cases.
13038 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13040 * Change RSA blinding code so that it works when the PRNG is not
13041 seeded (in this case, the secret RSA exponent is abused as
13042 an unpredictable seed -- if it is not unpredictable, there
13043 is no point in blinding anyway). Make RSA blinding thread-safe
13044 by remembering the creator's thread ID in rsa->blinding and
13045 having all other threads use local one-time blinding factors
13046 (this requires more computation than sharing rsa->blinding, but
13047 avoids excessive locking; and if an RSA object is not shared
13048 between threads, blinding will still be very fast).
13052 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003] ###
13054 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13055 via timing by performing a MAC computation even if incorrect
13056 block cipher padding has been found. This is a countermeasure
13057 against active attacks where the attacker has to distinguish
13058 between bad padding and a MAC verification error. [CVE-2003-0078][]
13060 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13061 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13062 Martin Vuagnoux (EPFL, Ilion)*
13064 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002] ###
13066 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13067 memory from its contents. This is done with a counter that will
13068 place alternating values in each byte. This can be used to solve
13069 two issues: 1) the removal of calls to memset() by highly optimizing
13070 compilers, and 2) cleansing with other values than 0, since those can
13071 be read through on certain media, for example a swap space on disk.
13075 * Bugfix: client side session caching did not work with external caching,
13076 because the session->cipher setting was not restored when reloading
13077 from the external cache. This problem was masked, when
13078 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13079 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13083 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13084 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13086 *Zeev Lieber <zeev-l@yahoo.com>*
13088 * Undo an undocumented change introduced in 0.9.6e which caused
13089 repeated calls to OpenSSL_add_all_ciphers() and
13090 OpenSSL_add_all_digests() to be ignored, even after calling
13095 * Change the default configuration reader to deal with last line not
13096 being properly terminated.
13100 * Change X509_NAME_cmp() so it applies the special rules on handling
13101 DN values that are of type PrintableString, as well as RDNs of type
13102 emailAddress where the value has the type ia5String.
13104 *stefank@valicert.com via Richard Levitte*
13106 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13107 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13108 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13109 the bitwise-OR of the two for use by the majority of applications
13110 wanting this behaviour, and update the docs. The documented
13111 behaviour and actual behaviour were inconsistent and had been
13112 changing anyway, so this is more a bug-fix than a behavioural
13115 *Geoff Thorpe, diagnosed by Nadav Har'El*
13117 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13118 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13122 * Fix initialization code race conditions in
13123 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13124 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13125 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13126 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13127 ssl2_get_cipher_by_char(),
13128 ssl3_get_cipher_by_char().
13130 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13132 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13133 the cached sessions are flushed, as the remove_cb() might use ex_data
13134 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13135 (see [openssl.org #212]).
13137 *Geoff Thorpe, Lutz Jaenicke*
13139 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13140 length, instead of the encoding length to d2i_ASN1_OBJECT.
13144 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002] ###
13146 * [In 0.9.6g-engine release:]
13147 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
13149 *Lynn Gazis <lgazis@rainbow.com>*
13151 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002] ###
13153 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13154 and get fix the header length calculation.
13155 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13156 Alon Kantor <alonk@checkpoint.com> (and others),
13159 * Use proper error handling instead of 'assertions' in buffer
13160 overflow checks added in 0.9.6e. This prevents DoS (the
13161 assertions could call abort()).
13163 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13165 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002] ###
13167 * Add various sanity checks to asn1_get_length() to reject
13168 the ASN1 length bytes if they exceed sizeof(long), will appear
13169 negative or the content length exceeds the length of the
13172 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13174 * Fix cipher selection routines: ciphers without encryption had no flags
13175 for the cipher strength set and where therefore not handled correctly
13176 by the selection routines (PR #130).
13180 * Fix EVP_dsa_sha macro.
13185 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13186 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13187 that was added in OpenSSL 0.9.6d.
13189 As the countermeasure turned out to be incompatible with some
13190 broken SSL implementations, the new option is part of SSL_OP_ALL.
13191 SSL_OP_ALL is usually employed when compatibility with weird SSL
13192 implementations is desired (e.g. '-bugs' option to 's_client' and
13193 's_server'), so the new option is automatically set in many
13198 * Changes in security patch:
13200 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13201 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13202 Air Force Materiel Command, USAF, under agreement number
13205 * Add various sanity checks to asn1_get_length() to reject
13206 the ASN1 length bytes if they exceed sizeof(long), will appear
13207 negative or the content length exceeds the length of the
13208 supplied buffer. [CVE-2002-0659][]
13210 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13212 * Assertions for various potential buffer overflows, not known to
13213 happen in practice.
13215 *Ben Laurie (CHATS)*
13217 * Various temporary buffers to hold ASCII versions of integers were
13218 too small for 64 bit platforms. [CVE-2002-0655][]
13219 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
13221 * Remote buffer overflow in SSL3 protocol - an attacker could
13222 supply an oversized session ID to a client. [CVE-2002-0656][]
13224 *Ben Laurie (CHATS)*
13226 * Remote buffer overflow in SSL2 protocol - an attacker could
13227 supply an oversized client master key. [CVE-2002-0656][]
13229 *Ben Laurie (CHATS)*
13231 ### Changes between 0.9.6c and 0.9.6d [9 May 2002] ###
13233 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13234 encoded as NULL) with id-dsa-with-sha1.
13236 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13238 * Check various X509_...() return values in apps/req.c.
13240 *Nils Larsch <nla@trustcenter.de>*
13242 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13243 an end-of-file condition would erroneously be flagged, when the CRLF
13244 was just at the end of a processed block. The bug was discovered when
13245 processing data through a buffering memory BIO handing the data to a
13246 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13247 <ptsekov@syntrex.com> and Nedelcho Stanev.
13251 * Implement a countermeasure against a vulnerability recently found
13252 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13253 before application data chunks to avoid the use of known IVs
13254 with data potentially chosen by the attacker.
13258 * Fix length checks in ssl3_get_client_hello().
13262 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13263 to prevent ssl3_read_internal() from incorrectly assuming that
13264 ssl3_read_bytes() found application data while handshake
13265 processing was enabled when in fact s->s3->in_read_app_data was
13266 merely automatically cleared during the initial handshake.
13268 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13270 * Fix object definitions for Private and Enterprise: they were not
13271 recognized in their shortname (=lowercase) representation. Extend
13272 obj_dat.pl to issue an error when using undefined keywords instead
13273 of silently ignoring the problem (Svenning Sorensen
13274 <sss@sss.dnsalias.net>).
13278 * Fix DH_generate_parameters() so that it works for 'non-standard'
13279 generators, i.e. generators other than 2 and 5. (Previously, the
13280 code did not properly initialise the 'add' and 'rem' values to
13281 BN_generate_prime().)
13283 In the new general case, we do not insist that 'generator' is
13284 actually a primitive root: This requirement is rather pointless;
13285 a generator of the order-q subgroup is just as good, if not
13290 * Map new X509 verification errors to alerts. Discovered and submitted by
13291 Tom Wu <tom@arcot.com>.
13295 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13296 returning non-zero before the data has been completely received
13297 when using non-blocking I/O.
13299 *Bodo Moeller; problem pointed out by John Hughes*
13301 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13303 *Ben Laurie, Lutz Jaenicke*
13305 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13306 Yoram Zahavi <YoramZ@gilian.com>).
13310 * Add information about CygWin 1.3 and on, and preserve proper
13311 configuration for the versions before that.
13313 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13315 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13316 check whether we deal with a copy of a session and do not delete from
13317 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13318 <izhar@checkpoint.com>.
13322 * Do not store session data into the internal session cache, if it
13323 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13324 flag is set). Proposed by Aslam <aslam@funk.com>.
13328 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13333 * [In 0.9.6d-engine release:]
13334 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13336 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13338 * Add the configuration target linux-s390x.
13340 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13342 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13343 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13344 variable as an indication that a ClientHello message has been
13345 received. As the flag value will be lost between multiple
13346 invocations of ssl3_accept when using non-blocking I/O, the
13347 function may not be aware that a handshake has actually taken
13348 place, thus preventing a new session from being added to the
13351 To avoid this problem, we now set s->new_session to 2 instead of
13352 using a local variable.
13354 *Lutz Jaenicke, Bodo Moeller*
13356 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13357 if the SSL_R_LENGTH_MISMATCH error is detected.
13359 *Geoff Thorpe, Bodo Moeller*
13361 * New 'shared_ldflag' column in Configure platform table.
13365 * Fix EVP_CIPHER_mode macro.
13367 *"Dan S. Camper" <dan@bti.net>*
13369 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13370 type, we must throw them away by setting rr->length to 0.
13372 *D P Chang <dpc@qualys.com>*
13374 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001] ###
13376 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13377 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13378 worked incorrectly for those cases where range = 10..._2 and
13379 3*range is two bits longer than range.)
13383 * Only add signing time to PKCS7 structures if it is not already
13388 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13389 OBJ_ld_ce should be OBJ_id_ce.
13390 Also some ip-pda OIDs in crypto/objects/objects.txt were
13391 incorrect (cf. RFC 3039).
13393 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13395 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13396 returns early because it has nothing to do.
13398 *Andy Schneider <andy.schneider@bjss.co.uk>*
13400 * [In 0.9.6c-engine release:]
13401 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13403 *Andy Schneider <andy.schneider@bjss.co.uk>*
13405 * [In 0.9.6c-engine release:]
13406 Add support for Cryptographic Appliance's keyserver technology.
13407 (Use engine 'keyclient')
13409 *Cryptographic Appliances and Geoff Thorpe*
13411 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13412 is called via tools/c89.sh because arguments have to be
13413 rearranged (all '-L' options must appear before the first object
13416 *Richard Shapiro <rshapiro@abinitio.com>*
13418 * [In 0.9.6c-engine release:]
13419 Add support for Broadcom crypto accelerator cards, backported
13422 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13424 * [In 0.9.6c-engine release:]
13425 Add support for SureWare crypto accelerator cards from
13426 Baltimore Technologies. (Use engine 'sureware')
13428 *Baltimore Technologies and Mark Cox*
13430 * [In 0.9.6c-engine release:]
13431 Add support for crypto accelerator cards from Accelerated
13432 Encryption Processing, www.aep.ie. (Use engine 'aep')
13434 *AEP Inc. and Mark Cox*
13436 * Add a configuration entry for gcc on UnixWare.
13438 *Gary Benson <gbenson@redhat.com>*
13440 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13441 messages are stored in a single piece (fixed-length part and
13442 variable-length part combined) and fix various bugs found on the way.
13446 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13447 instead. BIO_gethostbyname() does not know what timeouts are
13448 appropriate, so entries would stay in cache even when they have
13450 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
13452 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13453 faced with a pathologically small ClientHello fragment that does
13454 not contain client_version: Instead of aborting with an error,
13455 simply choose the highest available protocol version (i.e.,
13456 TLS 1.0 unless it is disabled). In practice, ClientHello
13457 messages are never sent like this, but this change gives us
13458 strictly correct behaviour at least for TLS.
13462 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13463 never resets s->method to s->ctx->method when called from within
13464 one of the SSL handshake functions.
13466 *Bodo Moeller; problem pointed out by Niko Baric*
13468 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13469 (sent using the client's version number) if client_version is
13470 smaller than the protocol version in use. Also change
13471 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13472 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13473 the client will at least see that alert.
13477 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13482 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13483 client receives HelloRequest while in a handshake.
13485 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13487 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13488 should end in 'break', not 'goto end' which circumvents various
13489 cleanups done in state SSL_ST_OK. But session related stuff
13490 must be disabled for SSL_ST_OK in the case that we just sent a
13493 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13494 before just sending a HelloRequest.
13496 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13498 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13499 reveal whether illegal block cipher padding was found or a MAC
13500 verification error occurred. (Neither SSLerr() codes nor alerts
13501 are directly visible to potential attackers, but the information
13502 may leak via logfiles.)
13504 Similar changes are not required for the SSL 2.0 implementation
13505 because the number of padding bytes is sent in clear for SSL 2.0,
13506 and the extra bytes are just ignored. However ssl/s2_pkt.c
13507 failed to verify that the purported number of padding bytes is in
13512 * Add OpenUNIX-8 support including shared libraries
13513 (Boyd Lynn Gerber <gerberb@zenez.com>).
13517 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13518 'wristwatch attack' using huge encoding parameters (cf.
13519 James H. Manger's CRYPTO 2001 paper). Note that the
13520 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13521 encoding parameters and hence was not vulnerable.
13525 * BN_sqr() bug fix.
13527 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13529 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13530 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13531 followed by modular reduction.
13533 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13535 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13536 equivalent based on BN_pseudo_rand() instead of BN_rand().
13540 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13541 This function was broken, as the check for a new client hello message
13542 to handle SGC did not allow these large messages.
13543 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13547 * Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
13551 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13552 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13556 * Rework the configuration and shared library support for Tru64 Unix.
13557 The configuration part makes use of modern compiler features and
13558 still retains old compiler behavior for those that run older versions
13559 of the OS. The shared library support part includes a variant that
13560 uses the RPATH feature, and is available through the special
13561 configuration target "alpha-cc-rpath", which will never be selected
13564 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13566 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13567 with the same message size as in ssl3_get_certificate_request().
13568 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13569 messages might inadvertently be reject as too long.
13571 *Petr Lampa <lampa@fee.vutbr.cz>*
13573 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13577 * Modified SSL library such that the verify_callback that has been set
13578 specifically for an SSL object with SSL_set_verify() is actually being
13579 used. Before the change, a verify_callback set with this function was
13580 ignored and the verify_callback() set in the SSL_CTX at the time of
13581 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13582 to allow the necessary settings.
13586 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13587 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13588 done automatically (in contradiction to the requirements of the C
13589 standard). This made problems when used from OpenSSH.
13593 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13594 dh->length and always used
13596 BN_rand_range(priv_key, dh->p).
13598 BN_rand_range() is not necessary for Diffie-Hellman, and this
13599 specific range makes Diffie-Hellman unnecessarily inefficient if
13600 dh->length (recommended exponent length) is much smaller than the
13601 length of dh->p. We could use BN_rand_range() if the order of
13602 the subgroup was stored in the DH structure, but we only have
13607 BN_rand(priv_key, l, ...)
13609 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13616 RSA_eay_public_encrypt
13617 RSA_eay_private_decrypt
13618 RSA_eay_private_encrypt (signing)
13619 RSA_eay_public_decrypt (signature verification)
13621 (default implementations for RSA_public_encrypt,
13622 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13623 always reject numbers >= n.
13627 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13628 to synchronize access to 'locking_thread'. This is necessary on
13629 systems where access to 'locking_thread' (an 'unsigned long'
13630 variable) is not atomic.
13634 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13635 *before* setting the 'crypto_lock_rand' flag. The previous code had
13636 a race condition if 0 is a valid thread ID.
13638 *Travis Vitek <vitek@roguewave.com>*
13640 * Add support for shared libraries under Irix.
13642 *Albert Chin-A-Young <china@thewrittenword.com>*
13644 * Add configuration option to build on Linux on both big-endian and
13645 little-endian MIPS.
13647 *Ralf Baechle <ralf@uni-koblenz.de>*
13649 * Add the possibility to create shared libraries on HP-UX.
13653 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001] ###
13655 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13656 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13657 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13658 PRNG state recovery was possible based on the output of
13659 one PRNG request appropriately sized to gain knowledge on
13660 'md' followed by enough consecutive 1-byte PRNG requests
13661 to traverse all of 'state'.
13663 1. When updating 'md_local' (the current thread's copy of 'md')
13664 during PRNG output generation, hash all of the previous
13665 'md_local' value, not just the half used for PRNG output.
13667 2. Make the number of bytes from 'state' included into the hash
13668 independent from the number of PRNG bytes requested.
13670 The first measure alone would be sufficient to avoid
13671 Markku-Juhani's attack. (Actually it had never occurred
13672 to me that the half of 'md_local' used for chaining was the
13673 half from which PRNG output bytes were taken -- I had always
13674 assumed that the secret half would be used.) The second
13675 measure makes sure that additional data from 'state' is never
13676 mixed into 'md_local' in small portions; this heuristically
13677 further strengthens the PRNG.
13681 * Fix crypto/bn/asm/mips3.s.
13685 * When only the key is given to "enc", the IV is undefined. Print out
13686 an error message in this case.
13690 * Handle special case when X509_NAME is empty in X509 printing routines.
13694 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13695 positive and less than q.
13699 * Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
13700 used: it isn't thread safe and the add_lock_callback should handle
13703 *Paul Rose <Paul.Rose@bridge.com>*
13705 * Verify that incoming data obeys the block size in
13706 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13712 *Ulf Möller, Bodo Möller*
13714 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13715 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13716 when fixing the server behaviour for backwards-compatible 'client
13717 hello' messages. (Note that the attack is impractical against
13718 SSL 3.0 and TLS 1.0 anyway because length and version checking
13719 means that the probability of guessing a valid ciphertext is
13720 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13723 Before 0.9.5, the countermeasure (hide the error by generating a
13724 random 'decryption result') did not work properly because
13725 ERR_clear_error() was missing, meaning that SSL_get_error() would
13726 detect the supposedly ignored error.
13728 Both problems are now fixed.
13732 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13733 (previously it was 1024).
13737 * Fix for compatibility mode trust settings: ignore trust settings
13738 unless some valid trust or reject settings are present.
13742 * Fix for blowfish EVP: its a variable length cipher.
13746 * Fix various bugs related to DSA S/MIME verification. Handle missing
13747 parameters in DSA public key structures and return an error in the
13748 DSA routines if parameters are absent.
13752 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13753 in the current directory if neither $RANDFILE nor $HOME was set.
13754 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13755 caused some confusion to Windows users who haven't defined $HOME.
13756 Thus RAND_file_name() is changed again: e_os.h can define a
13757 DEFAULT_HOME, which will be used if $HOME is not set.
13758 For Windows, we use "C:"; on other platforms, we still require
13759 environment variables.
13761 * Move 'if (!initialized) RAND_poll()' into regions protected by
13762 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13763 having multiple threads call RAND_poll() concurrently.
13767 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13768 combination of a flag and a thread ID variable.
13769 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13770 flag), *other* threads can enter ssleay_add_bytes without obeying
13771 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13772 that they do not hold after the first thread unsets add_do_not_lock).
13776 * Change bctest again: '-x' expressions are not available in all
13777 versions of 'test'.
13781 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001] ###
13783 * Fix a couple of memory leaks in PKCS7_dataDecode()
13785 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13787 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13788 the default extension for executables, if any. Also, make the perl
13789 scripts that use symlink() to test if it really exists and use "cp"
13790 if it doesn't. All this made OpenSSL compilable and installable in
13795 * Fix for asn1_GetSequence() for indefinite length constructed data.
13796 If SEQUENCE is length is indefinite just set c->slen to the total
13797 amount of data available.
13799 *Steve Henson, reported by shige@FreeBSD.org*
13801 *This change does not apply to 0.9.7.*
13803 * Change bctest to avoid here-documents inside command substitution
13804 (workaround for FreeBSD /bin/sh bug).
13805 For compatibility with Ultrix, avoid shell functions (introduced
13806 in the bctest version that searches along $PATH).
13810 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13811 with des_encrypt() defined on some operating systems, like Solaris
13816 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13817 On the Importance of Eliminating Errors in Cryptographic
13818 Computations, J. Cryptology 14 (2001) 2, 101-119,
13819 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
13823 * MIPS assembler BIGNUM division bug fix.
13827 * Disabled incorrect Alpha assembler code.
13831 * Fix PKCS#7 decode routines so they correctly update the length
13832 after reading an EOC for the EXPLICIT tag.
13836 *This change does not apply to 0.9.7.*
13838 * Fix bug in PKCS#12 key generation routines. This was triggered
13839 if a 3DES key was generated with a 0 initial byte. Include
13840 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13841 (but broken) behaviour.
13845 * Enhance bctest to search for a working bc along $PATH and print
13848 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13850 * Fix memory leaks in err.c: free err_data string if necessary;
13851 don't write to the wrong index in ERR_set_error_data.
13855 * Implement ssl23_peek (analogous to ssl23_read), which previously
13860 * Replace rdtsc with _emit statements for VC++ version 5.
13862 *Jeremy Cooper <jeremy@baymoo.org>*
13864 * Make it possible to reuse SSLv2 sessions.
13868 * In copy_email() check for >= 0 as a return value for
13869 X509_NAME_get_index_by_NID() since 0 is a valid index.
13871 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13873 * Avoid coredump with unsupported or invalid public keys by checking if
13874 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13875 PKCS7_verify() fails with non detached data.
13879 * Don't use getenv in library functions when run as setuid/setgid.
13880 New function OPENSSL_issetugid().
13884 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13885 due to incorrect handling of multi-threading:
13887 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13889 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13891 3. Count how many times MemCheck_off() has been called so that
13892 nested use can be treated correctly. This also avoids
13893 inband-signalling in the previous code (which relied on the
13894 assumption that thread ID 0 is impossible).
13898 * Add "-rand" option also to s_client and s_server.
13902 * Fix CPU detection on Irix 6.x.
13903 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13904 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13906 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13911 *This change does not apply to 0.9.7.*
13913 * Use the cached encoding of an X509_NAME structure rather than
13914 copying it. This is apparently the reason for the libsafe "errors"
13915 but the code is actually correct.
13919 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13920 Bleichenbacher's DSA attack.
13921 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13922 to be set and top=0 forces the highest bit to be set; top=-1 is new
13923 and leaves the highest bit random.
13925 *Ulf Moeller, Bodo Moeller*
13927 * In the NCONF_...-based implementations for CONF_... queries
13928 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13929 a temporary CONF structure with the data component set to NULL
13930 (which gives segmentation faults in lh_retrieve).
13931 Instead, use NULL for the CONF pointer in CONF_get_string and
13932 CONF_get_number (which may use environment variables) and directly
13933 return NULL from CONF_get_section.
13937 * Fix potential buffer overrun for EBCDIC.
13941 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
13942 keyUsage if basicConstraints absent for a CA.
13946 * Make SMIME_write_PKCS7() write mail header values with a format that
13947 is more generally accepted (no spaces before the semicolon), since
13948 some programs can't parse those values properly otherwise. Also make
13949 sure BIO's that break lines after each write do not create invalid
13954 * Make the CRL encoding routines work with empty SEQUENCE OF. The
13955 macros previously used would not encode an empty SEQUENCE OF
13956 and break the signature.
13960 *This change does not apply to 0.9.7.*
13962 * Zero the premaster secret after deriving the master secret in
13967 * Add some EVP_add_digest_alias registrations (as found in
13968 OpenSSL_add_all_digests()) to SSL_library_init()
13969 aka OpenSSL_add_ssl_algorithms(). This provides improved
13970 compatibility with peers using X.509 certificates
13971 with unconventional AlgorithmIdentifier OIDs.
13975 * Fix for Irix with NO_ASM.
13977 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13979 * ./config script fixes.
13981 *Ulf Moeller, Richard Levitte*
13983 * Fix 'openssl passwd -1'.
13987 * Change PKCS12_key_gen_asc() so it can cope with non null
13988 terminated strings whose length is passed in the passlen
13989 parameter, for example from PEM callbacks. This was done
13990 by adding an extra length parameter to asc2uni().
13992 *Steve Henson, reported by <oddissey@samsung.co.kr>*
13994 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
13995 call failed, free the DSA structure.
13999 * Fix to uni2asc() to cope with zero length Unicode strings.
14000 These are present in some PKCS#12 files.
14004 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14005 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14006 when writing a 32767 byte record.
14008 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14010 * In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
14011 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
14013 (RSA objects have a reference count access to which is protected
14014 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14015 so they are meant to be shared between threads.)
14016 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14017 "Reddie, Steven" <Steven.Reddie@ca.com>*
14019 * Fix a deadlock in CRYPTO_mem_leaks().
14023 * Use better test patterns in bntest.
14027 * rand_win.c fix for Borland C.
14031 * BN_rshift bugfix for n == 0.
14035 * Add a 'bctest' script that checks for some known 'bc' bugs
14036 so that 'make test' does not abort just because 'bc' is broken.
14040 * Store verify_result within SSL_SESSION also for client side to
14041 avoid potential security hole. (Re-used sessions on the client side
14042 always resulted in verify_result==X509_V_OK, not using the original
14043 result of the server certificate verification.)
14047 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14048 SSL3_RT_APPLICATION_DATA, return 0.
14049 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14054 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14055 releases, have been re-implemented by renaming the previous
14056 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14057 and ssl3_read_internal, respectively, and adding 'peek' parameters
14058 to them. The new ssl[23]_{read,peek} functions are calls to
14059 ssl[23]_read_internal with the 'peek' flag set appropriately.
14060 A 'peek' parameter has also been added to ssl3_read_bytes, which
14061 does the actual work for ssl3_read_internal.
14065 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14066 the method-specific "init()" handler. Also clean up ex_data after
14067 calling the method-specific "finish()" handler. Previously, this was
14068 happening the other way round.
14072 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14073 The previous value, 12, was not always sufficient for BN_mod_exp().
14077 * Make sure that shared libraries get the internal name engine with
14078 the full version number and not just 0. This should mark the
14079 shared libraries as not backward compatible. Of course, this should
14080 be changed again when we can guarantee backward binary compatibility.
14084 * Fix typo in get_cert_by_subject() in by_dir.c
14086 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14088 * Rework the system to generate shared libraries:
14090 - Make note of the expected extension for the shared libraries and
14091 if there is a need for symbolic links from for example libcrypto.so.0
14092 to libcrypto.so.0.9.7. There is extended info in Configure for
14095 - Make as few rebuilds of the shared libraries as possible.
14097 - Still avoid linking the OpenSSL programs with the shared libraries.
14099 - When installing, install the shared libraries separately from the
14104 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14106 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14107 and not in SSL_clear because the latter is also used by the
14108 accept/connect functions; previously, the settings made by
14109 SSL_set_read_ahead would be lost during the handshake.
14111 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14113 * Correct util/mkdef.pl to be selective about disabled algorithms.
14114 Previously, it would create entries for disabled algorithms no
14119 * Added several new manual pages for SSL_* function.
14123 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000] ###
14125 * In ssl23_get_client_hello, generate an error message when faced
14126 with an initial SSL 3.0/TLS record that is too small to contain the
14127 first two bytes of the ClientHello message, i.e. client_version.
14128 (Note that this is a pathologic case that probably has never happened
14129 in real life.) The previous approach was to use the version number
14130 from the record header as a substitute; but our protocol choice
14131 should not depend on that one because it is not authenticated
14132 by the Finished messages.
14136 * More robust randomness gathering functions for Windows.
14138 *Jeffrey Altman <jaltman@columbia.edu>*
14140 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14141 not set then we don't setup the error code for issuer check errors
14142 to avoid possibly overwriting other errors which the callback does
14143 handle. If an application does set the flag then we assume it knows
14144 what it is doing and can handle the new informational codes
14149 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14150 a general "ANY" type, as such it should be able to decode anything
14151 including tagged types. However it didn't check the class so it would
14152 wrongly interpret tagged types in the same way as their universal
14153 counterpart and unknown types were just rejected. Changed so that the
14154 tagged and unknown types are handled in the same way as a SEQUENCE:
14155 that is the encoding is stored intact. There is also a new type
14156 "V_ASN1_OTHER" which is used when the class is not universal, in this
14157 case we have no idea what the actual type is so we just lump them all
14162 * On VMS, stdout may very well lead to a file that is written to
14163 in a record-oriented fashion. That means that every write() will
14164 write a separate record, which will be read separately by the
14165 programs trying to read from it. This can be very confusing.
14167 The solution is to put a BIO filter in the way that will buffer
14168 text until a linefeed is reached, and then write everything a
14169 line at a time, so every record written will be an actual line,
14170 not chunks of lines and not (usually doesn't happen, but I've
14171 seen it once) several lines in one record. BIO_f_linebuffer() is
14174 Currently, it's a VMS-only method, because that's where it has
14175 been tested well enough.
14179 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14180 it can return incorrect results.
14181 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14182 but it was in 0.9.6-beta[12].)
14186 * Disable the check for content being present when verifying detached
14187 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14188 include zero length content when signing messages.
14192 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14193 BIO_ctrl (for BIO pairs).
14197 * Add DSO method for VMS.
14201 * Bug fix: Montgomery multiplication could produce results with the
14206 * Add RPM specification openssl.spec and modify it to build three
14207 packages. The default package contains applications, application
14208 documentation and run-time libraries. The devel package contains
14209 include files, static libraries and function documentation. The
14210 doc package contains the contents of the doc directory. The original
14211 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14215 * Add a large number of documentation files for many SSL routines.
14217 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14219 * Add a configuration entry for Sony News 4.
14221 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14223 * Don't set the two most significant bits to one when generating a
14224 random number < q in the DSA library.
14228 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14229 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14230 the underlying transport is blocking) if a handshake took place.
14231 (The default behaviour is needed by applications such as s_client
14232 and s_server that use select() to determine when to use SSL_read;
14233 but for applications that know in advance when to expect data, it
14234 just makes things more complicated.)
14238 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14243 * Add a few more EBCDIC conditionals that make `req' and `x509'
14244 work better on such systems.
14246 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14248 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14249 Update PKCS12_parse() so it copies the friendlyName and the
14250 keyid to the certificates aux info.
14254 * Fix bug in PKCS7_verify() which caused an infinite loop
14255 if there was more than one signature.
14257 *Sven Uszpelkat <su@celocom.de>*
14259 * Major change in util/mkdef.pl to include extra information
14260 about each symbol, as well as presenting variables as well
14261 as functions. This change means that there's n more need
14262 to rebuild the .num files when some algorithms are excluded.
14266 * Allow the verify time to be set by an application,
14267 rather than always using the current time.
14271 * Phase 2 verify code reorganisation. The certificate
14272 verify code now looks up an issuer certificate by a
14273 number of criteria: subject name, authority key id
14274 and key usage. It also verifies self signed certificates
14275 by the same criteria. The main comparison function is
14276 X509_check_issued() which performs these checks.
14278 Lot of changes were necessary in order to support this
14279 without completely rewriting the lookup code.
14281 Authority and subject key identifier are now cached.
14283 The LHASH 'certs' is X509_STORE has now been replaced
14284 by a STACK_OF(X509_OBJECT). This is mainly because an
14285 LHASH can't store or retrieve multiple objects with
14286 the same hash value.
14288 As a result various functions (which were all internal
14289 use only) have changed to handle the new X509_STORE
14290 structure. This will break anything that messed round
14291 with X509_STORE internally.
14293 The functions X509_STORE_add_cert() now checks for an
14294 exact match, rather than just subject name.
14296 The X509_STORE API doesn't directly support the retrieval
14297 of multiple certificates matching a given criteria, however
14298 this can be worked round by performing a lookup first
14299 (which will fill the cache with candidate certificates)
14300 and then examining the cache for matches. This is probably
14301 the best we can do without throwing out X509_LOOKUP
14302 entirely (maybe later...).
14304 The X509_VERIFY_CTX structure has been enhanced considerably.
14306 All certificate lookup operations now go via a get_issuer()
14307 callback. Although this currently uses an X509_STORE it
14308 can be replaced by custom lookups. This is a simple way
14309 to bypass the X509_STORE hackery necessary to make this
14310 work and makes it possible to use more efficient techniques
14311 in future. A very simple version which uses a simple
14312 STACK for its trusted certificate store is also provided
14313 using X509_STORE_CTX_trusted_stack().
14315 The verify_cb() and verify() callbacks now have equivalents
14316 in the X509_STORE_CTX structure.
14318 X509_STORE_CTX also has a 'flags' field which can be used
14319 to customise the verify behaviour.
14323 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14324 excludes S/MIME capabilities.
14328 * When a certificate request is read in keep a copy of the
14329 original encoding of the signed data and use it when outputting
14330 again. Signatures then use the original encoding rather than
14331 a decoded, encoded version which may cause problems if the
14332 request is improperly encoded.
14336 * For consistency with other BIO_puts implementations, call
14337 buffer_write(b, ...) directly in buffer_puts instead of calling
14340 In BIO_puts, increment b->num_write as in BIO_write.
14342 *Peter.Sylvester@EdelWeb.fr*
14344 * Fix BN_mul_word for the case where the word is 0. (We have to use
14345 BN_zero, we may not return a BIGNUM with an array consisting of
14346 words set to zero.)
14350 * Avoid calling abort() from within the library when problems are
14351 detected, except if preprocessor symbols have been defined
14352 (such as REF_CHECK, BN_DEBUG etc.).
14356 * New openssl application 'rsautl'. This utility can be
14357 used for low level RSA operations. DER public key
14358 BIO/fp routines also added.
14362 * New Configure entry and patches for compiling on QNX 4.
14364 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14366 * A demo state-machine implementation was sponsored by
14367 Nuron (http://www.nuron.com/) and is now available in
14368 demos/state_machine.
14372 * New options added to the 'dgst' utility for signature
14373 generation and verification.
14377 * Unrecognized PKCS#7 content types are now handled via a
14378 catch all ASN1_TYPE structure. This allows unsupported
14379 types to be stored as a "blob" and an application can
14380 encode and decode it manually.
14384 * Fix various signed/unsigned issues to make a_strex.c
14385 compile under VC++.
14387 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14389 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14390 length if passed a buffer. ASN1_INTEGER_to_BN failed
14391 if passed a NULL BN and its argument was negative.
14393 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14395 * Modification to PKCS#7 encoding routines to output definite
14396 length encoding. Since currently the whole structures are in
14397 memory there's not real point in using indefinite length
14398 constructed encoding. However if OpenSSL is compiled with
14399 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14403 * Added BIO_vprintf() and BIO_vsnprintf().
14407 * Added more prefixes to parse for in the strings written
14408 through a logging bio, to cover all the levels that are available
14409 through syslog. The prefixes are now:
14411 PANIC, EMERG, EMR => LOG_EMERG
14412 ALERT, ALR => LOG_ALERT
14413 CRIT, CRI => LOG_CRIT
14414 ERROR, ERR => LOG_ERR
14415 WARNING, WARN, WAR => LOG_WARNING
14416 NOTICE, NOTE, NOT => LOG_NOTICE
14417 INFO, INF => LOG_INFO
14418 DEBUG, DBG => LOG_DEBUG
14420 and as before, if none of those prefixes are present at the
14421 beginning of the string, LOG_ERR is chosen.
14423 On Win32, the LOG_* levels are mapped according to this:
14425 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14426 LOG_WARNING => EVENTLOG_WARNING_TYPE
14427 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14432 * Made it possible to reconfigure with just the configuration
14433 argument "reconf" or "reconfigure". The command line arguments
14434 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14435 and are retrieved from there when reconfiguring.
14441 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14443 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14447 * The obj_dat.pl script was messing up the sorting of object
14448 names. The reason was that it compared the quoted version
14449 of strings as a result "OCSP" > "OCSP Signing" because
14450 " > SPACE. Changed script to store unquoted versions of
14451 names and add quotes on output. It was also omitting some
14452 names from the lookup table if they were given a default
14453 value (that is if SN is missing it is given the same
14454 value as LN and vice versa), these are now added on the
14455 grounds that if an object has a name we should be able to
14456 look it up. Finally added warning output when duplicate
14457 short or long names are found.
14461 * Changes needed for Tandem NSK.
14463 *Scott Uroff <scott@xypro.com>*
14465 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14466 RSA_padding_check_SSLv23(), special padding was never detected
14467 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14468 version rollback attacks was not effective.
14470 In s23_clnt.c, don't use special rollback-attack detection padding
14471 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14472 client; similarly, in s23_srvr.c, don't do the rollback check if
14473 SSL 2.0 is the only protocol enabled in the server.
14477 * Make it possible to get hexdumps of unprintable data with 'openssl
14478 asn1parse'. By implication, the functions ASN1_parse_dump() and
14479 BIO_dump_indent() are added.
14483 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14484 these print out strings and name structures based on various
14485 flags including RFC2253 support and proper handling of
14486 multibyte characters. Added options to the 'x509' utility
14487 to allow the various flags to be set.
14491 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14492 Also change the functions X509_cmp_current_time() and
14493 X509_gmtime_adj() work with an ASN1_TIME structure,
14494 this will enable certificates using GeneralizedTime in validity
14495 dates to be checked.
14499 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14500 negative public key encodings) on by default,
14501 NO_NEG_PUBKEY_BUG can be set to disable it.
14505 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14506 content octets. An i2c_ASN1_OBJECT is unnecessary because
14507 the encoding can be trivially obtained from the structure.
14511 * crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
14512 not read locks (CRYPTO_r_[un]lock).
14516 * A first attempt at creating official support for shared
14517 libraries through configuration. I've kept it so the
14518 default is static libraries only, and the OpenSSL programs
14519 are always statically linked for now, but there are
14520 preparations for dynamic linking in place.
14521 This has been tested on Linux and Tru64.
14525 * Randomness polling function for Win9x, as described in:
14526 Peter Gutmann, Software Generation of Practically Strong
14531 * Fix so PRNG is seeded in req if using an already existing
14536 * New options to smime application. -inform and -outform
14537 allow alternative formats for the S/MIME message including
14538 PEM and DER. The -content option allows the content to be
14539 specified separately. This should allow things like Netscape
14540 form signing output easier to verify.
14544 * Fix the ASN1 encoding of tags using the 'long form'.
14548 * New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
14549 STRING types. These convert content octets to and from the
14550 underlying type. The actual tag and length octets are
14551 already assumed to have been read in and checked. These
14552 are needed because all other string types have virtually
14553 identical handling apart from the tag. By having versions
14554 of the ASN1 functions that just operate on content octets
14555 IMPLICIT tagging can be handled properly. It also allows
14556 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14557 and ASN1_INTEGER are identical apart from the tag.
14561 * Change the handling of OID objects as follows:
14563 - New object identifiers are inserted in objects.txt, following
14564 the syntax given in objects.README.
14565 - objects.pl is used to process obj_mac.num and create a new
14567 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14570 This is currently kind of a hack, and the perl code in objects.pl
14571 isn't very elegant, but it works as I intended. The simplest way
14572 to check that it worked correctly is to look in obj_dat.h and
14573 check the array nid_objs and make sure the objects haven't moved
14574 around (this is important!). Additions are OK, as well as
14575 consistent name changes.
14579 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14583 * Addition of the command line parameter '-rand file' to 'openssl req'.
14584 The given file adds to whatever has already been seeded into the
14585 random pool through the RANDFILE configuration file option or
14586 environment variable, or the default random state file.
14590 * mkstack.pl now sorts each macro group into lexical order.
14591 Previously the output order depended on the order the files
14592 appeared in the directory, resulting in needless rewriting
14597 * Patches to make OpenSSL compile under Win32 again. Mostly
14598 work arounds for the VC++ problem that it treats func() as
14599 func(void). Also stripped out the parts of mkdef.pl that
14600 added extra typesafe functions: these no longer exist.
14604 * Reorganisation of the stack code. The macros are now all
14605 collected in safestack.h . Each macro is defined in terms of
14606 a "stack macro" of the form SKM_<name>(type, a, b). The
14607 DEBUG_SAFESTACK is now handled in terms of function casts,
14608 this has the advantage of retaining type safety without the
14609 use of additional functions. If DEBUG_SAFESTACK is not defined
14610 then the non typesafe macros are used instead. Also modified the
14611 mkstack.pl script to handle the new form. Needs testing to see
14612 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14613 the default if no major problems. Similar behaviour for ASN1_SET_OF
14614 and PKCS12_STACK_OF.
14618 * When some versions of IIS use the 'NET' form of private key the
14619 key derivation algorithm is different. Normally MD5(password) is
14620 used as a 128 bit RC4 key. In the modified case
14621 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14622 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14623 as the old Netscape_RSA functions except they have an additional
14624 'sgckey' parameter which uses the modified algorithm. Also added
14625 an -sgckey command line option to the rsa utility. Thanks to
14626 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14627 algorithm to openssl-dev.
14631 * The evp_local.h macros were using 'c.##kname' which resulted in
14632 invalid expansion on some systems (SCO 5.0.5 for example).
14633 Corrected to 'c.kname'.
14635 *Phillip Porch <root@theporch.com>*
14637 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14638 a STACK of email addresses from a certificate or request, these look
14639 in the subject name and the subject alternative name extensions and
14640 omit any duplicate addresses.
14644 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14645 This makes DSA verification about 2 % faster.
14649 * Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
14650 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14651 plus overhead for 1024 bit moduli).
14652 This makes exponentiations about 0.5 % faster for 1024 bit
14653 exponents (as measured by "openssl speed rsa2048").
14657 * Rename memory handling macros to avoid conflicts with other
14659 Malloc => OPENSSL_malloc
14660 Malloc_locked => OPENSSL_malloc_locked
14661 Realloc => OPENSSL_realloc
14662 Free => OPENSSL_free
14666 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14667 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14671 * CygWin32 support.
14673 *John Jarvie <jjarvie@newsguy.com>*
14675 * The type-safe stack code has been rejigged. It is now only compiled
14676 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14677 by default all type-specific stack functions are "#define"d back to
14678 standard stack functions. This results in more streamlined output
14679 but retains the type-safety checking possibilities of the original
14684 * The STACK code has been cleaned up, and certain type declarations
14685 that didn't make a lot of sense have been brought in line. This has
14686 also involved a cleanup of sorts in safestack.h to more correctly
14687 map type-safe stack functions onto their plain stack counterparts.
14688 This work has also resulted in a variety of "const"ifications of
14689 lots of the code, especially "_cmp" operations which should normally
14690 be prototyped with "const" parameters anyway.
14694 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14695 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14696 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14697 where all of 'md' is used each time the PRNG is used, but 'state'
14698 is used only indexed by a cyclic counter. As entropy may not be
14699 well distributed from the beginning, 'md' is important as a
14700 chaining variable. However, the output function chains only half
14701 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14702 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14703 in all of 'state' being rewritten, with the new values depending
14704 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14708 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14709 the handshake is continued after ssl_verify_cert_chain();
14710 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14711 can lead to 'unexplainable' connection aborts later.
14713 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14715 * Major EVP API cipher revision.
14716 Add hooks for extra EVP features. This allows various cipher
14717 parameters to be set in the EVP interface. Support added for variable
14718 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14719 setting of RC2 and RC5 parameters.
14721 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14724 Remove lots of duplicated code from the EVP library. For example *every*
14725 cipher init() function handles the 'iv' in the same way according to the
14726 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14727 for CFB and OFB modes they zero ctx->num.
14729 New functionality allows removal of S/MIME code RC2 hack.
14731 Most of the routines have the same form and so can be declared in terms
14734 By shifting this to the top level EVP_CipherInit() it can be removed from
14735 all individual ciphers. If the cipher wants to handle IVs or keys
14736 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14739 Change lots of functions like EVP_EncryptUpdate() to now return a
14740 value: although software versions of the algorithms cannot fail
14741 any installed hardware versions can.
14745 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14746 this option is set, tolerate broken clients that send the negotiated
14747 protocol version number instead of the requested protocol version
14752 * Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
14753 i.e. non-zero for export ciphersuites, zero otherwise.
14754 Previous versions had this flag inverted, inconsistent with
14755 rsa_tmp_cb (..._TMP_RSA_CB).
14757 *Bodo Moeller; problem reported by Amit Chopra*
14759 * Add missing DSA library text string. Work around for some IIS
14760 key files with invalid SEQUENCE encoding.
14764 * Add a document (doc/standards.txt) that list all kinds of standards
14765 and so on that are implemented in OpenSSL.
14769 * Enhance c_rehash script. Old version would mishandle certificates
14770 with the same subject name hash and wouldn't handle CRLs at all.
14771 Added -fingerprint option to crl utility, to support new c_rehash
14776 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14780 * Fix for SSL server purpose checking. Server checking was
14781 rejecting certificates which had extended key usage present
14782 but no ssl client purpose.
14784 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14786 * Make PKCS#12 code work with no password. The PKCS#12 spec
14787 is a little unclear about how a blank password is handled.
14788 Since the password in encoded as a BMPString with terminating
14789 double NULL a zero length password would end up as just the
14790 double NULL. However no password at all is different and is
14791 handled differently in the PKCS#12 key generation code. NS
14792 treats a blank password as zero length. MSIE treats it as no
14793 password on export: but it will try both on import. We now do
14794 the same: PKCS12_parse() tries zero length and no password if
14795 the password is set to "" or NULL (NULL is now a valid password:
14796 it wasn't before) as does the pkcs12 application.
14800 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14801 perror when PEM_read_bio_X509_REQ fails, the error message must
14802 be obtained from the error queue.
14806 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14807 it in ERR_remove_state if appropriate, and change ERR_get_state
14808 accordingly to avoid race conditions (this is necessary because
14809 thread_hash is no longer constant once set).
14813 * Bugfix for linux-elf makefile.one.
14817 * RSA_get_default_method() will now cause a default
14818 RSA_METHOD to be chosen if one doesn't exist already.
14819 Previously this was only set during a call to RSA_new()
14820 or RSA_new_method(NULL) meaning it was possible for
14821 RSA_get_default_method() to return NULL.
14825 * Added native name translation to the existing DSO code
14826 that will convert (if the flag to do so is set) filenames
14827 that are sufficiently small and have no path information
14828 into a canonical native form. Eg. "blah" converted to
14829 "libblah.so" or "blah.dll" etc.
14833 * New function ERR_error_string_n(e, buf, len) which is like
14834 ERR_error_string(e, buf), but writes at most 'len' bytes
14835 including the 0 terminator. For ERR_error_string_n, 'buf'
14838 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14840 * CONF library reworked to become more general. A new CONF
14841 configuration file reader "class" is implemented as well as a
14842 new functions (NCONF_*, for "New CONF") to handle it. The now
14843 old CONF_* functions are still there, but are reimplemented to
14844 work in terms of the new functions. Also, a set of functions
14845 to handle the internal storage of the configuration data is
14846 provided to make it easier to write new configuration file
14847 reader "classes" (I can definitely see something reading a
14848 configuration file in XML format, for example), called _CONF_*,
14849 or "the configuration storage API"...
14851 The new configuration file reading functions are:
14853 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14854 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14856 NCONF_default, NCONF_WIN32
14858 NCONF_dump_fp, NCONF_dump_bio
14860 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14861 NCONF_new creates a new CONF object. This works in the same way
14862 as other interfaces in OpenSSL, like the BIO interface.
14863 NCONF_dump_* dump the internal storage of the configuration file,
14864 which is useful for debugging. All other functions take the same
14865 arguments as the old CONF_* functions with the exception of the
14866 first that must be a `CONF *' instead of a `LHASH *'.
14868 To make it easier to use the new classes with the old CONF_* functions,
14869 the function CONF_set_default_method is provided.
14873 * Add '-tls1' option to 'openssl ciphers', which was already
14874 mentioned in the documentation but had not been implemented.
14875 (This option is not yet really useful because even the additional
14876 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14880 * Initial DSO code added into libcrypto for letting OpenSSL (and
14881 OpenSSL-based applications) load shared libraries and bind to
14882 them in a portable way.
14884 *Geoff Thorpe, with contributions from Richard Levitte*
14886 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000] ###
14888 * Make sure _lrotl and _lrotr are only used with MSVC.
14890 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14891 (the default implementation of RAND_status).
14893 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14894 to '-clrext' (= clear extensions), as intended and documented.
14895 *Bodo Moeller; inconsistency pointed out by Michael Attili
14896 <attili@amaxo.com>*
14898 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14899 was larger than the MD block size.
14901 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14903 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14904 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14905 using the passed key: if the passed key was a private key the result
14906 of X509_print(), for example, would be to print out all the private key
14911 * des_quad_cksum() byte order bug fix.
14912 *Ulf Möller, using the problem description in krb4-0.9.7, where
14913 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14915 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14918 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14920 * For easily testing in shell scripts whether some command
14921 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14922 returns with exit code 0 iff no command of the given name is available.
14923 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14924 the output goes to stdout and nothing is printed to stderr.
14925 Additional arguments are always ignored.
14927 Since for each cipher there is a command of the same name,
14928 the 'no-cipher' compilation switches can be tested this way.
14930 ('openssl no-XXX' is not able to detect pseudo-commands such
14931 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
14935 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
14939 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
14940 is set; it will be thrown away anyway because each handshake creates
14942 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
14943 to parameters -- in previous versions (since OpenSSL 0.9.3) the
14944 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
14945 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
14949 * New s_client option -ign_eof: EOF at stdin is ignored, and
14950 'Q' and 'R' lose their special meanings (quit/renegotiate).
14951 This is part of what -quiet does; unlike -quiet, -ign_eof
14952 does not suppress any output.
14956 * Add compatibility options to the purpose and trust code. The
14957 purpose X509_PURPOSE_ANY is "any purpose" which automatically
14958 accepts a certificate or CA, this was the previous behaviour,
14959 with all the associated security issues.
14961 X509_TRUST_COMPAT is the old trust behaviour: only and
14962 automatically trust self signed roots in certificate store. A
14963 new trust setting X509_TRUST_DEFAULT is used to specify that
14964 a purpose has no associated trust setting and it should instead
14965 use the value in the default purpose.
14969 * Fix the PKCS#8 DSA private key code so it decodes keys again
14970 and fix a memory leak.
14974 * In util/mkerr.pl (which implements 'make errors'), preserve
14975 reason strings from the previous version of the .c file, as
14976 the default to have only downcase letters (and digits) in
14977 automatically generated reasons codes is not always appropriate.
14981 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
14982 using strerror. Previously, ERR_reason_error_string() returned
14983 library names as reason strings for SYSerr; but SYSerr is a special
14984 case where small numbers are errno values, not library numbers.
14988 * Add '-dsaparam' option to 'openssl dhparam' application. This
14989 converts DSA parameters into DH parameters. (When creating parameters,
14990 DSA_generate_parameters is used.)
14994 * Include 'length' (recommended exponent length) in C code generated
14995 by 'openssl dhparam -C'.
14999 * The second argument to set_label in perlasm was already being used
15000 so couldn't be used as a "file scope" flag. Moved to third argument
15005 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15006 instead of RAND_bytes for encryption IVs and salts.
15010 * Include RAND_status() into RAND_METHOD instead of implementing
15011 it only for md_rand.c Otherwise replacing the PRNG by calling
15012 RAND_set_rand_method would be impossible.
15016 * Don't let DSA_generate_key() enter an infinite loop if the random
15017 number generation fails.
15021 * New 'rand' application for creating pseudo-random output.
15025 * Added configuration support for Linux/IA64
15027 *Rolf Haberrecker <rolf@suse.de>*
15029 * Assembler module support for Mingw32.
15033 * Shared library support for HPUX (in shlib/).
15035 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15037 * Shared library support for Solaris gcc.
15039 *Lutz Behnke <behnke@trustcenter.de>*
15041 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000] ###
15043 * PKCS7_encrypt() was adding text MIME headers twice because they
15044 were added manually and by SMIME_crlf_copy().
15048 * In bntest.c don't call BN_rand with zero bits argument.
15050 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15052 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15053 case was implemented. This caused BN_div_recp() to fail occasionally.
15057 * Add an optional second argument to the set_label() in the perl
15058 assembly language builder. If this argument exists and is set
15059 to 1 it signals that the assembler should use a symbol whose
15060 scope is the entire file, not just the current function. This
15061 is needed with MASM which uses the format label:: for this scope.
15063 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15065 * Change the ASN1 types so they are typedefs by default. Before
15066 almost all types were #define'd to ASN1_STRING which was causing
15067 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15072 * Change names of new functions to the new get1/get0 naming
15073 convention: After 'get1', the caller owns a reference count
15074 and has to call ..._free; 'get0' returns a pointer to some
15075 data structure without incrementing reference counters.
15076 (Some of the existing 'get' functions increment a reference
15077 counter, some don't.)
15078 Similarly, 'set1' and 'add1' functions increase reference
15079 counters or duplicate objects.
15083 * Allow for the possibility of temp RSA key generation failure:
15084 the code used to assume it always worked and crashed on failure.
15088 * Fix potential buffer overrun problem in BIO_printf().
15089 *Ulf Möller, using public domain code by Patrick Powell; problem
15090 pointed out by David Sacerdote <das33@cornell.edu>*
15092 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15093 RAND_egd() and RAND_status(). In the command line application,
15094 the EGD socket can be specified like a seed file using RANDFILE
15099 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15100 Some CAs (e.g. Verisign) distribute certificates in this form.
15104 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15105 list to exclude them. This means that no special compilation option
15106 is needed to use anonymous DH: it just needs to be included in the
15111 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15112 EVP_MD_type. The old functionality is available in a new macro called
15113 EVP_MD_md(). Change code that uses it and update docs.
15117 * ..._ctrl functions now have corresponding ..._callback_ctrl functions
15118 where the 'void *' argument is replaced by a function pointer argument.
15119 Previously 'void *' was abused to point to functions, which works on
15120 many platforms, but is not correct. As these functions are usually
15121 called by macros defined in OpenSSL header files, most source code
15122 should work without changes.
15126 * <openssl/opensslconf.h> (which is created by Configure) now contains
15127 sections with information on -D... compiler switches used for
15128 compiling the library so that applications can see them. To enable
15129 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
15130 must be defined. E.g.,
15131 #define OPENSSL_ALGORITHM_DEFINES
15132 #include <openssl/opensslconf.h>
15133 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
15135 *Richard Levitte, Ulf and Bodo Möller*
15137 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15142 * Change the 'other' type in certificate aux info to a STACK_OF
15143 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15144 the required ASN1 format: arbitrary types determined by an OID.
15148 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15149 argument to 'req'. This is not because the function is newer or
15150 better than others it just uses the work 'NEW' in the certificate
15151 request header lines. Some software needs this.
15155 * Reorganise password command line arguments: now passwords can be
15156 obtained from various sources. Delete the PEM_cb function and make
15157 it the default behaviour: i.e. if the callback is NULL and the
15158 usrdata argument is not NULL interpret it as a null terminated pass
15159 phrase. If usrdata and the callback are NULL then the pass phrase
15160 is prompted for as usual.
15164 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15165 the support is automatically enabled. The resulting binaries will
15166 autodetect the card and use it if present.
15168 *Ben Laurie and Compaq Inc.*
15170 * Work around for Netscape hang bug. This sends certificate request
15171 and server done in one record. Since this is perfectly legal in the
15172 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15173 the bugs/SSLv3 entry for more info.
15177 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15181 * Add -rand argument to smime and pkcs12 applications and read/write
15186 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15190 * Add command line password options to the remaining applications.
15194 * Bug fix for BN_div_recp() for numerators with an even number of
15199 * More tests in bntest.c, and changed test_bn output.
15203 * ./config recognizes MacOS X now.
15207 * Bug fix for BN_div() when the first words of num and divisor are
15208 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
15212 * Add support for various broken PKCS#8 formats, and command line
15213 options to produce them.
15217 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15218 get temporary BIGNUMs from a BN_CTX.
15222 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15227 * Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
15228 include a #define from the old name to the new. The original intent
15229 was that statically linked binaries could for example just call
15230 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15231 link with digests. This never worked because SSLeay_add_all_digests()
15232 and SSLeay_add_all_ciphers() were in the same source file so calling
15233 one would link with the other. They are now in separate source files.
15237 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15241 * Use a less unusual form of the Miller-Rabin primality test (it used
15242 a binary algorithm for exponentiation integrated into the Miller-Rabin
15243 loop, our standard modexp algorithms are faster).
15247 * Support for the EBCDIC character set completed.
15249 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15251 * Source code cleanups: use const where appropriate, eliminate casts,
15252 use void * instead of char * in lhash.
15256 * Bugfix: ssl3_send_server_key_exchange was not restartable
15257 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15258 this the server could overwrite ephemeral keys that the client
15263 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15264 using 50 iterations of the Rabin-Miller test.
15266 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15267 iterations of the Rabin-Miller test as required by the appendix
15268 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15269 As BN_is_prime_fasttest includes trial division, DSA parameter
15270 generation becomes much faster.
15272 This implies a change for the callback functions in DSA_is_prime
15273 and DSA_generate_parameters: The callback function is called once
15274 for each positive witness in the Rabin-Miller test, not just
15275 occasionally in the inner loop; and the parameters to the
15276 callback function now provide an iteration count for the outer
15277 loop rather than for the current invocation of the inner loop.
15278 DSA_generate_parameters additionally can call the callback
15279 function with an 'iteration count' of -1, meaning that a
15280 candidate has passed the trial division test (when q is generated
15281 from an application-provided seed, trial division is skipped).
15285 * New function BN_is_prime_fasttest that optionally does trial
15286 division before starting the Rabin-Miller test and has
15287 an additional BN_CTX * argument (whereas BN_is_prime always
15288 has to allocate at least one BN_CTX).
15289 'callback(1, -1, cb_arg)' is called when a number has passed the
15290 trial division stage.
15294 * Fix for bug in CRL encoding. The validity dates weren't being handled
15299 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15303 * New function BN_pseudo_rand().
15307 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15308 bignum version of BN_from_montgomery() with the working code from
15309 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15314 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15315 made it impossible to use the same SSL_SESSION data structure in
15316 SSL2 clients in multiple threads.
15320 * The return value of RAND_load_file() no longer counts bytes obtained
15321 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15322 to seed the PRNG (previously an explicit byte count was required).
15324 *Ulf Möller, Bodo Möller*
15326 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15327 used (char *) instead of (void *) and had casts all over the place.
15331 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15335 * Retain source code compatibility for BN_prime_checks macro:
15336 BN_is_prime(..., BN_prime_checks, ...) now uses
15337 BN_prime_checks_for_size to determine the appropriate number of
15338 Rabin-Miller iterations.
15342 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15343 DH_CHECK_P_NOT_SAFE_PRIME.
15344 (Check if this is true? OpenPGP calls them "strong".)
15348 * Merge the functionality of "dh" and "gendh" programs into a new program
15349 "dhparam". The old programs are retained for now but will handle DH keys
15350 (instead of parameters) in future.
15354 * Make the ciphers, s_server and s_client programs check the return values
15355 when a new cipher list is set.
15359 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15360 ciphers. Before when the 56bit ciphers were enabled the sorting was
15363 The syntax for the cipher sorting has been extended to support sorting by
15364 cipher-strength (using the strength_bits hard coded in the tables).
15365 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15367 Fix a bug in the cipher-command parser: when supplying a cipher command
15368 string with an "undefined" symbol (neither command nor alphanumeric
15369 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15370 an error is flagged.
15372 Due to the strength-sorting extension, the code of the
15373 ssl_create_cipher_list() function was completely rearranged. I hope that
15374 the readability was also increased :-)
15376 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15378 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15379 for the first serial number and places 2 in the serial number file. This
15380 avoids problems when the root CA is created with serial number zero and
15381 the first user certificate has the same issuer name and serial number
15386 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15387 the new code. Add documentation for this stuff.
15391 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15392 X509_*() to X509at_*() on the grounds that they don't handle X509
15393 structures and behave in an analogous way to the X509v3 functions:
15394 they shouldn't be called directly but wrapper functions should be used
15397 So we also now have some wrapper functions that call the X509at functions
15398 when passed certificate requests. (TO DO: similar things can be done with
15399 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15400 things. Some of these need some d2i or i2d and print functionality
15401 because they handle more complex structures.)
15405 * Add missing #ifndefs that caused missing symbols when building libssl
15406 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15407 NO_RSA in ssl/s2*.c.
15409 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15411 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15412 has a return value which indicates the quality of the random data
15413 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15414 error queue. New function RAND_pseudo_bytes() generates output that is
15415 guaranteed to be unique but not unpredictable. RAND_add is like
15416 RAND_seed, but takes an extra argument for an entropy estimate
15417 (RAND_seed always assumes full entropy).
15421 * Do more iterations of Rabin-Miller probable prime test (specifically,
15422 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15423 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15424 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15425 false-positive rate of at most 2^-80 for random input.
15429 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15433 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15434 in the 0.9.5 release), this returns the chain
15435 from an X509_CTX structure with a dup of the stack and all
15436 the X509 reference counts upped: so the stack will exist
15437 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15440 Also make SSL_SESSION_print() print out the verify return
15445 * Add manpage for the pkcs12 command. Also change the default
15446 behaviour so MAC iteration counts are used unless the new
15447 -nomaciter option is used. This improves file security and
15448 only older versions of MSIE (4.0 for example) need it.
15452 * Honor the no-xxx Configure options when creating .DEF files.
15456 * Add PKCS#10 attributes to field table: challengePassword,
15457 unstructuredName and unstructuredAddress. These are taken from
15458 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15459 international characters are used.
15461 More changes to X509_ATTRIBUTE code: allow the setting of types
15462 based on strings. Remove the 'loc' parameter when adding
15463 attributes because these will be a SET OF encoding which is sorted
15468 * Initial changes to the 'req' utility to allow request generation
15469 automation. This will allow an application to just generate a template
15470 file containing all the field values and have req construct the
15473 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15474 used all over the place including certificate requests and PKCS#7
15475 structures. They are currently handled manually where necessary with
15476 some primitive wrappers for PKCS#7. The new functions behave in a
15477 manner analogous to the X509 extension functions: they allow
15478 attributes to be looked up by NID and added.
15480 Later something similar to the X509V3 code would be desirable to
15481 automatically handle the encoding, decoding and printing of the
15482 more complex types. The string types like challengePassword can
15483 be handled by the string table functions.
15485 Also modified the multi byte string table handling. Now there is
15486 a 'global mask' which masks out certain types. The table itself
15487 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15488 is useful when for example there is only one permissible type
15489 (as in countryName) and using the mask might result in no valid
15494 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15495 SSL_get_peer_finished to allow applications to obtain the latest
15496 Finished messages sent to the peer or expected from the peer,
15497 respectively. (SSL_get_peer_finished is usually the Finished message
15498 actually received from the peer, otherwise the protocol will be aborted.)
15500 As the Finished message are message digests of the complete handshake
15501 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15502 be used for external authentication procedures when the authentication
15503 provided by SSL/TLS is not desired or is not enough.
15507 * Enhanced support for Alpha Linux is added. Now ./config checks if
15508 the host supports BWX extension and if Compaq C is present on the
15509 $PATH. Just exploiting of the BWX extension results in 20-30%
15510 performance kick for some algorithms, e.g. DES and RC4 to mention
15511 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15516 * Add support for MS "fast SGC". This is arguably a violation of the
15517 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15518 weak crypto and after checking the certificate is SGC a second one
15519 with strong crypto. MS SGC stops the first handshake after receiving
15520 the server certificate message and sends a second client hello. Since
15521 a server will typically do all the time consuming operations before
15522 expecting any further messages from the client (server key exchange
15523 is the most expensive) there is little difference between the two.
15525 To get OpenSSL to support MS SGC we have to permit a second client
15526 hello message after we have sent server done. In addition we have to
15527 reset the MAC if we do get this second client hello.
15531 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15532 if a DER encoded private key is RSA or DSA traditional format. Changed
15533 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15534 format DER encoded private key. Newer code should use PKCS#8 format which
15535 has the key type encoded in the ASN1 structure. Added DER private key
15536 support to pkcs8 application.
15540 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15541 ciphersuites has been selected (as required by the SSL 3/TLS 1
15542 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15543 is set, we interpret this as a request to violate the specification
15544 (the worst that can happen is a handshake failure, and 'correct'
15545 behaviour would result in a handshake failure anyway).
15549 * In SSL_CTX_add_session, take into account that there might be multiple
15550 SSL_SESSION structures with the same session ID (e.g. when two threads
15551 concurrently obtain them from an external cache).
15552 The internal cache can handle only one SSL_SESSION with a given ID,
15553 so if there's a conflict, we now throw out the old one to achieve
15558 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15559 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15560 some routines that use cipher OIDs: some ciphers do not have OIDs
15561 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15566 * Simplify the trust setting structure and code. Now we just have
15567 two sequences of OIDs for trusted and rejected settings. These will
15568 typically have values the same as the extended key usage extension
15569 and any application specific purposes.
15571 The trust checking code now has a default behaviour: it will just
15572 check for an object with the same NID as the passed id. Functions can
15573 be provided to override either the default behaviour or the behaviour
15574 for a given id. SSL client, server and email already have functions
15575 in place for compatibility: they check the NID and also return "trusted"
15576 if the certificate is self signed.
15580 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15581 traditional format into an EVP_PKEY structure.
15585 * Add a password callback function PEM_cb() which either prompts for
15586 a password if usr_data is NULL or otherwise assumes it is a null
15587 terminated password. Allow passwords to be passed on command line
15588 environment or config files in a few more utilities.
15592 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15593 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15594 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15595 Update documentation.
15599 * Support for ASN1 "NULL" type. This could be handled before by using
15600 ASN1_TYPE but there wasn't any function that would try to read a NULL
15601 and produce an error if it couldn't. For compatibility we also have
15602 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15603 don't allocate anything because they don't need to.
15607 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15610 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15612 * Rebuild of the memory allocation routines used by OpenSSL code and
15613 possibly others as well. The purpose is to make an interface that
15614 provide hooks so anyone can build a separate set of allocation and
15615 deallocation routines to be used by OpenSSL, for example memory
15616 pool implementations, or something else, which was previously hard
15617 since Malloc(), Realloc() and Free() were defined as macros having
15618 the values malloc, realloc and free, respectively (except for Win32
15619 compilations). The same is provided for memory debugging code.
15620 OpenSSL already comes with functionality to find memory leaks, but
15621 this gives people a chance to debug other memory problems.
15623 With these changes, a new set of functions and macros have appeared:
15625 CRYPTO_set_mem_debug_functions() [F]
15626 CRYPTO_get_mem_debug_functions() [F]
15627 CRYPTO_dbg_set_options() [F]
15628 CRYPTO_dbg_get_options() [F]
15629 CRYPTO_malloc_debug_init() [M]
15631 The memory debug functions are NULL by default, unless the library
15632 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15633 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15634 gives the standard debugging functions that come with OpenSSL) or
15635 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15636 provided by the library user) must be used. When the standard
15637 debugging functions are used, CRYPTO_dbg_set_options can be used to
15638 request additional information:
15639 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15640 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15642 Also, things like CRYPTO_set_mem_functions will always give the
15643 expected result (the new set of functions is used for allocation
15644 and deallocation) at all times, regardless of platform and compiler
15647 To finish it up, some functions that were never use in any other
15648 way than through macros have a new API and new semantic:
15650 CRYPTO_dbg_malloc()
15651 CRYPTO_dbg_realloc()
15654 All macros of value have retained their old syntax.
15656 *Richard Levitte and Bodo Moeller*
15658 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15659 ordering of SMIMECapabilities wasn't in "strength order" and there
15660 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15665 * Some ASN1 types with illegal zero length encoding (INTEGER,
15666 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15668 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15670 * Merge in my S/MIME library for OpenSSL. This provides a simple
15671 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15672 functionality to handle multipart/signed properly) and a utility
15673 called 'smime' to call all this stuff. This is based on code I
15674 originally wrote for Celo who have kindly allowed it to be
15675 included in OpenSSL.
15679 * Add variants des_set_key_checked and des_set_key_unchecked of
15680 des_set_key (aka des_key_sched). Global variable des_check_key
15681 decides which of these is called by des_set_key; this way
15682 des_check_key behaves as it always did, but applications and
15683 the library itself, which was buggy for des_check_key == 1,
15684 have a cleaner way to pick the version they need.
15688 * New function PKCS12_newpass() which changes the password of a
15693 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15694 dynamic mix. In both cases the ids can be used as an index into the
15695 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15696 functions so they accept a list of the field values and the
15697 application doesn't need to directly manipulate the X509_TRUST
15702 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15707 * Modify the way the V3 extension code looks up extensions. This now
15708 works in a similar way to the object code: we have some "standard"
15709 extensions in a static table which is searched with OBJ_bsearch()
15710 and the application can add dynamic ones if needed. The file
15711 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15712 updated whenever a new extension is added to the core code and kept
15713 in ext_nid order. There is a simple program 'tabtest.c' which checks
15714 this. New extensions are not added too often so this file can readily
15715 be maintained manually.
15717 There are two big advantages in doing things this way. The extensions
15718 can be looked up immediately and no longer need to be "added" using
15719 X509V3_add_standard_extensions(): this function now does nothing.
15720 *Side note: I get *lots* of email saying the extension code doesn't
15721 work because people forget to call this function*
15722 Also no dynamic allocation is done unless new extensions are added:
15723 so if we don't add custom extensions there is no need to call
15724 X509V3_EXT_cleanup().
15728 * Modify enc utility's salting as follows: make salting the default. Add a
15729 magic header, so unsalted files fail gracefully instead of just decrypting
15730 to garbage. This is because not salting is a big security hole, so people
15731 should be discouraged from doing it.
15735 * Fixes and enhancements to the 'x509' utility. It allowed a message
15736 digest to be passed on the command line but it only used this
15737 parameter when signing a certificate. Modified so all relevant
15738 operations are affected by the digest parameter including the
15739 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15740 DSA key was used because it didn't fix the digest.
15744 * Initial certificate chain verify code. Currently tests the untrusted
15745 certificates for consistency with the verify purpose (which is set
15746 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15748 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15749 this is because it will reject chains with invalid extensions whereas
15750 every previous version of OpenSSL and SSLeay made no checks at all.
15752 Trust code: checks the root CA for the relevant trust settings. Trust
15753 settings have an initial value consistent with the verify purpose: e.g.
15754 if the verify purpose is for SSL client use it expects the CA to be
15755 trusted for SSL client use. However the default value can be changed to
15756 permit custom trust settings: one example of this would be to only trust
15757 certificates from a specific "secure" set of CAs.
15759 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15760 which should be used for version portability: especially since the
15761 verify structure is likely to change more often now.
15763 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15764 to set them. If not set then assume SSL clients will verify SSL servers
15767 Two new options to the verify program: -untrusted allows a set of
15768 untrusted certificates to be passed in and -purpose which sets the
15769 intended purpose of the certificate. If a purpose is set then the
15770 new chain verify code is used to check extension consistency.
15774 * Support for the authority information access extension.
15778 * Modify RSA and DSA PEM read routines to transparently handle
15779 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15780 public keys in a format compatible with certificate
15781 SubjectPublicKeyInfo structures. Unfortunately there were already
15782 functions called *_PublicKey_* which used various odd formats so
15783 these are retained for compatibility: however the DSA variants were
15784 never in a public release so they have been deleted. Changed dsa/rsa
15785 utilities to handle the new format: note no releases ever handled public
15786 keys so we should be OK.
15788 The primary motivation for this change is to avoid the same fiasco
15789 that dogs private keys: there are several incompatible private key
15790 formats some of which are standard and some OpenSSL specific and
15791 require various evil hacks to allow partial transparent handling and
15792 even then it doesn't work with DER formats. Given the option anything
15793 other than PKCS#8 should be dumped: but the other formats have to
15794 stay in the name of compatibility.
15796 With public keys and the benefit of hindsight one standard format
15797 is used which works with EVP_PKEY, RSA or DSA structures: though
15798 it clearly returns an error if you try to read the wrong kind of key.
15800 Added a -pubkey option to the 'x509' utility to output the public key.
15801 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
15802 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
15803 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
15804 that do the same as the EVP_PKEY_assign_*() except they up the
15805 reference count of the added key (they don't "swallow" the
15810 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15811 CRLs would fail if the file contained no certificates or no CRLs:
15812 added a new function to read in both types and return the number
15813 read: this means that if none are read it will be an error. The
15814 DER versions of the certificate and CRL reader would always fail
15815 because it isn't possible to mix certificates and CRLs in DER format
15816 without choking one or the other routine. Changed this to just read
15817 a certificate: this is the best we can do. Also modified the code
15818 in apps/verify.c to take notice of return codes: it was previously
15819 attempting to read in certificates from NULL pointers and ignoring
15820 any errors: this is one reason why the cert and CRL reader seemed
15821 to work. It doesn't check return codes from the default certificate
15822 routines: these may well fail if the certificates aren't installed.
15826 * Code to support otherName option in GeneralName.
15830 * First update to verify code. Change the verify utility
15831 so it warns if it is passed a self signed certificate:
15832 for consistency with the normal behaviour. X509_verify
15833 has been modified to it will now verify a self signed
15834 certificate if *exactly* the same certificate appears
15835 in the store: it was previously impossible to trust a
15836 single self signed certificate. This means that:
15837 openssl verify ss.pem
15838 now gives a warning about a self signed certificate but
15839 openssl verify -CAfile ss.pem ss.pem
15844 * For servers, store verify_result in SSL_SESSION data structure
15845 (and add it to external session representation).
15846 This is needed when client certificate verifications fails,
15847 but an application-provided verification callback (set by
15848 SSL_CTX_set_cert_verify_callback) allows accepting the session
15849 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15850 but returns 1): When the session is reused, we have to set
15851 ssl->verify_result to the appropriate error code to avoid
15854 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15856 * Fix a bug in the new PKCS#7 code: it didn't consider the
15857 case in PKCS7_dataInit() where the signed PKCS7 structure
15858 didn't contain any existing data because it was being created.
15860 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15862 * Add a salt to the key derivation routines in enc.c. This
15863 forms the first 8 bytes of the encrypted file. Also add a
15864 -S option to allow a salt to be input on the command line.
15868 * New function X509_cmp(). Oddly enough there wasn't a function
15869 to compare two certificates. We do this by working out the SHA1
15870 hash and comparing that. X509_cmp() will be needed by the trust
15875 * SSL_get1_session() is like SSL_get_session(), but increments
15876 the reference count in the SSL_SESSION returned.
15878 *Geoff Thorpe <geoff@eu.c2.net>*
15880 * Fix for 'req': it was adding a null to request attributes.
15881 Also change the X509_LOOKUP and X509_INFO code to handle
15882 certificate auxiliary information.
15886 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15891 * Add the possibility to add extra information to the memory leak
15892 detecting output, to form tracebacks, showing from where each
15893 allocation was originated: CRYPTO_push_info("constant string") adds
15894 the string plus current file name and line number to a per-thread
15895 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15896 is like calling CYRPTO_pop_info() until the stack is empty.
15897 Also updated memory leak detection code to be multi-thread-safe.
15901 * Add options -text and -noout to pkcs7 utility and delete the
15902 encryption options which never did anything. Update docs.
15906 * Add options to some of the utilities to allow the pass phrase
15907 to be included on either the command line (not recommended on
15908 OSes like Unix) or read from the environment. Update the
15909 manpages and fix a few bugs.
15913 * Add a few manpages for some of the openssl commands.
15917 * Fix the -revoke option in ca. It was freeing up memory twice,
15918 leaking and not finding already revoked certificates.
15922 * Extensive changes to support certificate auxiliary information.
15923 This involves the use of X509_CERT_AUX structure and X509_AUX
15924 functions. An X509_AUX function such as PEM_read_X509_AUX()
15925 can still read in a certificate file in the usual way but it
15926 will also read in any additional "auxiliary information". By
15927 doing things this way a fair degree of compatibility can be
15928 retained: existing certificates can have this information added
15929 using the new 'x509' options.
15931 Current auxiliary information includes an "alias" and some trust
15932 settings. The trust settings will ultimately be used in enhanced
15933 certificate chain verification routines: currently a certificate
15934 can only be trusted if it is self signed and then it is trusted
15939 * Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
15940 The problem was that one of the replacement routines had not been working
15941 since SSLeay releases. For now the offending routine has been replaced
15942 with non-optimised assembler. Even so, this now gives around 95%
15943 performance improvement for 1024 bit RSA signs.
15947 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
15948 handling. Most clients have the effective key size in bits equal to
15949 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
15950 A few however don't do this and instead use the size of the decrypted key
15951 to determine the RC2 key length and the AlgorithmIdentifier to determine
15952 the effective key length. In this case the effective key length can still
15953 be 40 bits but the key length can be 168 bits for example. This is fixed
15954 by manually forcing an RC2 key into the EVP_PKEY structure because the
15955 EVP code can't currently handle unusual RC2 key sizes: it always assumes
15956 the key length and effective key length are equal.
15960 * Add a bunch of functions that should simplify the creation of
15961 X509_NAME structures. Now you should be able to do:
15962 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
15963 and have it automatically work out the correct field type and fill in
15964 the structures. The more adventurous can try:
15965 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
15966 and it will (hopefully) work out the correct multibyte encoding.
15970 * Change the 'req' utility to use the new field handling and multibyte
15971 copy routines. Before the DN field creation was handled in an ad hoc
15972 way in req, ca, and x509 which was rather broken and didn't support
15973 BMPStrings or UTF8Strings. Since some software doesn't implement
15974 BMPStrings or UTF8Strings yet, they can be enabled using the config file
15975 using the dirstring_type option. See the new comment in the default
15976 openssl.cnf for more info.
15980 * Make crypto/rand/md_rand.c more robust:
15981 - Assure unique random numbers after fork().
15982 - Make sure that concurrent threads access the global counter and
15983 md serializably so that we never lose entropy in them
15984 or use exactly the same state in multiple threads.
15985 Access to the large state is not always serializable because
15986 the additional locking could be a performance killer, and
15987 md should be large enough anyway.
15991 * New file apps/app_rand.c with commonly needed functionality
15992 for handling the random seed file.
15994 Use the random seed file in some applications that previously did not:
15996 dsaparam -genkey (which also ignored its '-rand' option),
15999 x509 (when signing).
16000 Except on systems with /dev/urandom, it is crucial to have a random
16001 seed file at least for key creation, DSA signing, and for DH exchanges;
16002 for RSA signatures we could do without one.
16004 gendh and gendsa (unlike genrsa) used to read only the first byte
16005 of each file listed in the '-rand' option. The function as previously
16006 found in genrsa is now in app_rand.c and is used by all programs
16007 that support '-rand'.
16011 * In RAND_write_file, use mode 0600 for creating files;
16012 don't just chmod when it may be too late.
16016 * Report an error from X509_STORE_load_locations
16017 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16021 * New function ASN1_mbstring_copy() this copies a string in either
16022 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16023 into an ASN1_STRING type. A mask of permissible types is passed
16024 and it chooses the "minimal" type to use or an error if not type
16029 * Add function equivalents to the various macros in asn1.h. The old
16030 macros are retained with an M_ prefix. Code inside the library can
16031 use the M_ macros. External code (including the openssl utility)
16032 should *NOT* in order to be "shared library friendly".
16036 * Add various functions that can check a certificate's extensions
16037 to see if it usable for various purposes such as SSL client,
16038 server or S/MIME and CAs of these types. This is currently
16039 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16040 verification. Also added a -purpose flag to x509 utility to
16041 print out all the purposes.
16045 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16050 * New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
16051 for, obtain and decode and extension and obtain its critical flag.
16052 This allows all the necessary extension code to be handled in a
16053 single function call.
16057 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16058 platforms. See crypto/rc4/rc4_enc.c for further details.
16062 * New -noout option to asn1parse. This causes no output to be produced
16063 its main use is when combined with -strparse and -out to extract data
16064 from a file (which may not be in ASN.1 format).
16068 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16069 when producing the local key id.
16071 *Richard Levitte <levitte@stacken.kth.se>*
16073 * New option -dhparam in s_server. This allows a DH parameter file to be
16074 stated explicitly. If it is not stated then it tries the first server
16075 certificate file. The previous behaviour hard coded the filename
16080 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16081 a public key to be input or output. For example:
16082 openssl rsa -in key.pem -pubout -out pubkey.pem
16083 Also added necessary DSA public key functions to handle this.
16087 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16088 in the message. This was handled by allowing
16089 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16091 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16093 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16094 to the end of the strings whereas this didn't. This would cause problems
16095 if strings read with d2i_ASN1_bytes() were later modified.
16097 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16099 * Fix for base64 decode bug. When a base64 bio reads only one line of
16100 data and it contains EOF it will end up returning an error. This is
16101 caused by input 46 bytes long. The cause is due to the way base64
16102 BIOs find the start of base64 encoded data. They do this by trying a
16103 trial decode on each line until they find one that works. When they
16104 do a flag is set and it starts again knowing it can pass all the
16105 data directly through the decoder. Unfortunately it doesn't reset
16106 the context it uses. This means that if EOF is reached an attempt
16107 is made to pass two EOFs through the context and this causes the
16108 resulting error. This can also cause other problems as well. As is
16109 usual with these problems it takes *ages* to find and the fix is
16110 trivial: move one line.
16112 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) *
16114 * Ugly workaround to get s_client and s_server working under Windows. The
16115 old code wouldn't work because it needed to select() on sockets and the
16116 tty (for keypresses and to see if data could be written). Win32 only
16117 supports select() on sockets so we select() with a 1s timeout on the
16118 sockets and then see if any characters are waiting to be read, if none
16119 are present then we retry, we also assume we can always write data to
16120 the tty. This isn't nice because the code then blocks until we've
16121 received a complete line of data and it is effectively polling the
16122 keyboard at 1s intervals: however it's quite a bit better than not
16123 working at all :-) A dedicated Windows application might handle this
16124 with an event loop for example.
16128 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16129 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16130 will be called when RSA_sign() and RSA_verify() are used. This is useful
16131 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16132 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16133 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16134 This necessitated the support of an extra signature type NID_md5_sha1
16135 for SSL signatures and modifications to the SSL library to use it instead
16136 of calling RSA_public_decrypt() and RSA_private_encrypt().
16140 * Add new -verify -CAfile and -CApath options to the crl program, these
16141 will lookup a CRL issuers certificate and verify the signature in a
16142 similar way to the verify program. Tidy up the crl program so it
16143 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16144 less strict. It will now permit CRL extensions even if it is not
16145 a V2 CRL: this will allow it to tolerate some broken CRLs.
16149 * Initialize all non-automatic variables each time one of the openssl
16150 sub-programs is started (this is necessary as they may be started
16151 multiple times from the "OpenSSL>" prompt).
16153 *Lennart Bang, Bodo Moeller*
16155 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16156 removing all other RSA functionality (this is what NO_RSA does). This
16157 is so (for example) those in the US can disable those operations covered
16158 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16163 * Non-copying interface to BIO pairs.
16164 (still largely untested)
16168 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16169 ASCII string. This was handled independently in various places before.
16173 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16174 UTF8 strings a character at a time.
16178 * Use client_version from client hello to select the protocol
16179 (s23_srvr.c) and for RSA client key exchange verification
16180 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16184 * Add various utility functions to handle SPKACs, these were previously
16185 handled by poking round in the structure internals. Added new function
16186 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16187 print, verify and generate SPKACs. Based on an original idea from
16188 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16192 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16196 * Allow the config file extension section to be overwritten on the
16197 command line. Based on an original idea from Massimiliano Pala
16198 <madwolf@comune.modena.it>. The new option is called -extensions
16199 and can be applied to ca, req and x509. Also -reqexts to override
16200 the request extensions in req and -crlexts to override the crl extensions
16205 * Add new feature to the SPKAC handling in ca. Now you can include
16206 the same field multiple times by preceding it by "XXXX." for example:
16209 this is the same syntax as used in the req config file.
16213 * Allow certificate extensions to be added to certificate requests. These
16214 are specified in a 'req_extensions' option of the req section of the
16215 config file. They can be printed out with the -text option to req but
16216 are otherwise ignored at present.
16220 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16221 data read consists of only the final block it would not decrypted because
16222 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16223 A misplaced 'break' also meant the decrypted final block might not be
16224 copied until the next read.
16228 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16229 a few extra parameters to the DH structure: these will be useful if
16230 for example we want the value of 'q' or implement X9.42 DH.
16234 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16235 provides hooks that allow the default DSA functions or functions on a
16236 "per key" basis to be replaced. This allows hardware acceleration and
16237 hardware key storage to be handled without major modification to the
16238 library. Also added low level modexp hooks and CRYPTO_EX structure and
16239 associated functions.
16243 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16244 as "read only": it can't be written to and the buffer it points to will
16245 not be freed. Reading from a read only BIO is much more efficient than
16246 a normal memory BIO. This was added because there are several times when
16247 an area of memory needs to be read from a BIO. The previous method was
16248 to create a memory BIO and write the data to it, this results in two
16249 copies of the data and an O(n^2) reading algorithm. There is a new
16250 function BIO_new_mem_buf() which creates a read only memory BIO from
16251 an area of memory. Also modified the PKCS#7 routines to use read only
16256 * Bugfix: ssl23_get_client_hello did not work properly when called in
16257 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16258 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16259 but a retry condition occurred while trying to read the rest.
16263 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16264 NID_pkcs7_encrypted by default: this was wrong since this should almost
16265 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16266 the encrypted data type: this is a more sensible place to put it and it
16267 allows the PKCS#12 code to be tidied up that duplicated this
16272 * Changed obj_dat.pl script so it takes its input and output files on
16273 the command line. This should avoid shell escape redirection problems
16278 * Initial support for certificate extension requests, these are included
16279 in things like Xenroll certificate requests. Included functions to allow
16280 extensions to be obtained and added.
16284 * -crlf option to s_client and s_server for sending newlines as
16285 CRLF (as required by many protocols).
16289 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999] ###
16291 * Install libRSAglue.a when OpenSSL is built with RSAref.
16293 *Ralf S. Engelschall*
16295 * A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
16297 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16299 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16304 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16305 DH parameters/keys (q is lost during that conversion, but the resulting
16306 DH parameters contain its length).
16308 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16309 much faster than DH_generate_parameters (which creates parameters
16310 where p = 2*q + 1), and also the smaller q makes DH computations
16311 much more efficient (160-bit exponentiation instead of 1024-bit
16312 exponentiation); so this provides a convenient way to support DHE
16313 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16314 utter importance to use
16315 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16317 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16318 when such DH parameters are used, because otherwise small subgroup
16319 attacks may become possible!
16323 * Avoid memory leak in i2d_DHparams.
16327 * Allow the -k option to be used more than once in the enc program:
16328 this allows the same encrypted message to be read by multiple recipients.
16332 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16333 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16334 it will always use the numerical form of the OID, even if it has a short
16339 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16340 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16341 otherwise bn_mod_exp was called. In the case of hardware keys for example
16342 no private key components need be present and it might store extra data
16343 in the RSA structure, which cannot be accessed from bn_mod_exp.
16344 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16345 private key operations.
16349 * Added support for SPARC Linux.
16353 * pem_password_cb function type incompatibly changed from
16354 typedef int pem_password_cb(char *buf, int size, int rwflag);
16356 ....(char *buf, int size, int rwflag, void *userdata);
16357 so that applications can pass data to their callbacks:
16358 The PEM[_ASN1]_{read,write}... functions and macros now take an
16359 additional void * argument, which is just handed through whenever
16360 the password callback is called.
16362 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16364 New function SSL_CTX_set_default_passwd_cb_userdata.
16366 Compatibility note: As many C implementations push function arguments
16367 onto the stack in reverse order, the new library version is likely to
16368 interoperate with programs that have been compiled with the old
16369 pem_password_cb definition (PEM_whatever takes some data that
16370 happens to be on the stack as its last argument, and the callback
16371 just ignores this garbage); but there is no guarantee whatsoever that
16374 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16375 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16376 problems not only on Windows, but also on some Unix platforms.
16377 To avoid problematic command lines, these definitions are now in an
16378 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16379 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16383 * MIPS III/IV assembler module is reimplemented.
16387 * More DES library cleanups: remove references to srand/rand and
16388 delete an unused file.
16392 * Add support for the free Netwide assembler (NASM) under Win32,
16393 since not many people have MASM (ml) and it can be hard to obtain.
16394 This is currently experimental but it seems to work OK and pass all
16395 the tests. Check out INSTALL.W32 for info.
16399 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16400 without temporary keys kept an extra copy of the server key,
16401 and connections with temporary keys did not free everything in case
16406 * New function RSA_check_key and new openssl rsa option -check
16407 for verifying the consistency of RSA keys.
16409 *Ulf Moeller, Bodo Moeller*
16411 * Various changes to make Win32 compile work:
16412 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16413 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16414 comparison" warnings.
16415 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
16419 * Add a debugging option to PKCS#5 v2 key generation function: when
16420 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16421 derived keys are printed to stderr.
16425 * Copy the flags in ASN1_STRING_dup().
16427 *Roman E. Pavlov <pre@mo.msk.ru>*
16429 * The x509 application mishandled signing requests containing DSA
16430 keys when the signing key was also DSA and the parameters didn't match.
16432 It was supposed to omit the parameters when they matched the signing key:
16433 the verifying software was then supposed to automatically use the CA's
16434 parameters if they were absent from the end user certificate.
16436 Omitting parameters is no longer recommended. The test was also
16437 the wrong way round! This was probably due to unusual behaviour in
16438 EVP_cmp_parameters() which returns 1 if the parameters match.
16439 This meant that parameters were omitted when they *didn't* match and
16440 the certificate was useless. Certificates signed with 'ca' didn't have
16443 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16445 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16446 The interface is as follows:
16447 Applications can use
16448 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16449 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16450 "off" is now the default.
16451 The library internally uses
16452 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16453 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16454 to disable memory-checking temporarily.
16456 Some inconsistent states that previously were possible (and were
16457 even the default) are now avoided.
16459 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16460 with each memory chunk allocated; this is occasionally more helpful
16461 than just having a counter.
16463 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16465 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16470 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16471 which largely parallels "options", but is for changing API behaviour,
16472 whereas "options" are about protocol behaviour.
16473 Initial "mode" flags are:
16475 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16476 a single record has been written.
16477 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16478 retries use the same buffer location.
16479 (But all of the contents must be
16484 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16487 * Fix problems with no-hmac etc.
16489 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16491 * New functions RSA_get_default_method(), RSA_set_method() and
16492 RSA_get_method(). These allows replacement of RSA_METHODs without having
16493 to mess around with the internals of an RSA structure.
16497 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16498 Also really enable memory leak checks in openssl.c and in some
16501 *Chad C. Mulligan, Bodo Moeller*
16503 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16504 up the length of negative integers. This has now been simplified to just
16505 store the length when it is first determined and use it later, rather
16506 than trying to keep track of where data is copied and updating it to
16508 *Steve Henson, reported by Brien Wheeler
16509 <bwheeler@authentica-security.com>*
16511 * Add a new function PKCS7_signatureVerify. This allows the verification
16512 of a PKCS#7 signature but with the signing certificate passed to the
16513 function itself. This contrasts with PKCS7_dataVerify which assumes the
16514 certificate is present in the PKCS#7 structure. This isn't always the
16515 case: certificates can be omitted from a PKCS#7 structure and be
16516 distributed by "out of band" means (such as a certificate database).
16520 * Complete the PEM_* macros with DECLARE_PEM versions to replace the
16521 function prototypes in pem.h, also change util/mkdef.pl to add the
16522 necessary function names.
16526 * mk1mf.pl (used by Windows builds) did not properly read the
16527 options set by Configure in the top level Makefile, and Configure
16528 was not even able to write more than one option correctly.
16529 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16533 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16534 file to be loaded from a BIO or FILE pointer. The BIO version will
16535 for example allow memory BIOs to contain config info.
16539 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16540 Whoever hopes to achieve shared-library compatibility across versions
16541 must use this, not the compile-time macro.
16542 (Exercise 0.9.4: Which is the minimum library version required by
16544 Note: All this applies only to multi-threaded programs, others don't
16549 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16550 through a BIO pair triggered the default case, i.e.
16551 SSLerr(...,SSL_R_UNKNOWN_STATE).
16555 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16556 can use the SSL library even if none of the specific BIOs is
16561 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16562 for the encoded length.
16564 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16566 * Add initial documentation of the X509V3 functions.
16570 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16571 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16572 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16573 secure PKCS#8 private key format with a high iteration count.
16577 * Fix determination of Perl interpreter: A perl or perl5
16578 _directory_ in $PATH was also accepted as the interpreter.
16580 *Ralf S. Engelschall*
16582 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16583 wrong with it but it was very old and did things like calling
16584 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16585 unusual formatting.
16589 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16590 to use the new extension code.
16594 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16595 with macros. This should make it easier to change their form, add extra
16596 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16601 * Add to configuration table a new entry that can specify an alternative
16602 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16603 according to Mark Crispin <MRC@Panda.COM>.
16608 * DES CBC did not update the IV. Weird.
16612 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16613 Changing the behaviour of the former might break existing programs --
16614 where IV updating is needed, des_ncbc_encrypt can be used.
16617 * When bntest is run from "make test" it drives bc to check its
16618 calculations, as well as internally checking them. If an internal check
16619 fails, it needs to cause bc to give a non-zero result or make test carries
16620 on without noticing the failure. Fixed.
16624 * DES library cleanups.
16628 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16629 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16630 ciphers. NOTE: although the key derivation function has been verified
16631 against some published test vectors it has not been extensively tested
16632 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16637 * Instead of "mkdir -p", which is not fully portable, use new
16638 Perl script "util/mkdir-p.pl".
16642 * Rewrite the way password based encryption (PBE) is handled. It used to
16643 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16644 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16645 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16646 the 'parameter' field of the AlgorithmIdentifier is passed to the
16647 underlying key generation function so it must do its own ASN1 parsing.
16648 This has also changed the EVP_PBE_CipherInit() function which now has a
16649 'parameter' argument instead of literal salt and iteration count values
16650 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16654 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16655 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16656 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16657 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16658 value was just used as a "magic string" and not used directly its
16659 value doesn't matter.
16663 * Introduce some semblance of const correctness to BN. Shame C doesn't
16668 * "linux-sparc64" configuration (ultrapenguin).
16670 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16671 "linux-sparc" configuration.
16673 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16675 * config now generates no-xxx options for missing ciphers.
16679 * Support the EBCDIC character set (work in progress).
16680 File ebcdic.c not yet included because it has a different license.
16682 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16684 * Support BS2000/OSD-POSIX.
16686 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16688 * Make callbacks for key generation use void * instead of char *.
16692 * Make S/MIME samples compile (not yet tested).
16696 * Additional typesafe stacks.
16700 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16705 ### Changes between 0.9.3 and 0.9.3a [29 May 1999] ###
16707 * New configuration variant "sco5-gcc".
16709 * Updated some demos.
16711 *Sean O Riordain, Wade Scholine*
16713 * Add missing BIO_free at exit of pkcs12 application.
16717 * Fix memory leak in conf.c.
16721 * Updates for Win32 to assembler version of MD5.
16725 * Set #! path to perl in apps/der_chop to where we found it
16726 instead of using a fixed path.
16730 * SHA library changes for irix64-mips4-cc.
16734 * Improvements for VMS support.
16739 ### Changes between 0.9.2b and 0.9.3 [24 May 1999] ###
16741 * Bignum library bug fix. IRIX 6 passes "make test" now!
16742 This also avoids the problems with SC4.2 and unpatched SC5.
16744 *Andy Polyakov <appro@fy.chalmers.se>*
16746 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16747 These are required because of the typesafe stack would otherwise break
16748 existing code. If old code used a structure member which used to be STACK
16749 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16750 sk_num or sk_value it would produce an error because the num, data members
16751 are not present in STACK_OF. Now it just produces a warning. sk_set
16752 replaces the old method of assigning a value to sk_value
16753 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16754 that does this will no longer work (and should use sk_set instead) but
16755 this could be regarded as a "questionable" behaviour anyway.
16759 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16760 correctly handle encrypted S/MIME data.
16764 * Change type of various DES function arguments from des_cblock
16765 (which means, in function argument declarations, pointer to char)
16766 to des_cblock * (meaning pointer to array with 8 char elements),
16767 which allows the compiler to do more typechecking; it was like
16768 that back in SSLeay, but with lots of ugly casts.
16770 Introduce new type const_des_cblock.
16774 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16775 problems: find RecipientInfo structure that matches recipient certificate
16776 and initialise the ASN1 structures properly based on passed cipher.
16780 * Belatedly make the BN tests actually check the results.
16784 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16785 to and from BNs: it was completely broken. New compilation option
16786 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16787 key elements as negative integers.
16791 * Reorganize and speed up MD5.
16793 *Andy Polyakov <appro@fy.chalmers.se>*
16797 *Richard Levitte <richard@levitte.org>*
16799 * New option -out to asn1parse to allow the parsed structure to be
16800 output to a file. This is most useful when combined with the -strparse
16801 option to examine the output of things like OCTET STRINGS.
16805 * Make SSL library a little more fool-proof by not requiring any longer
16806 that SSL_set_{accept,connect}_state be called before
16807 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
16808 in many applications because usually everything *appeared* to work as
16809 intended anyway -- now it really works as intended).
16813 * Move openssl.cnf out of lib/.
16817 * Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
16818 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16819 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
16821 *Ralf S. Engelschall*
16823 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16824 handle PKCS#7 enveloped data properly.
16826 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16828 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16829 copying pointers. The cert_st handling is changed by this in
16830 various ways (and thus what used to be known as ctx->default_cert
16831 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
16832 any longer when s->cert does not give us what we need).
16833 ssl_cert_instantiate becomes obsolete by this change.
16834 As soon as we've got the new code right (possibly it already is?),
16835 we have solved a couple of bugs of the earlier code where s->cert
16836 was used as if it could not have been shared with other SSL structures.
16838 Note that using the SSL API in certain dirty ways now will result
16839 in different behaviour than observed with earlier library versions:
16840 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
16841 does not influence s as it used to.
16843 In order to clean up things more thoroughly, inside SSL_SESSION
16844 we don't use CERT any longer, but a new structure SESS_CERT
16845 that holds per-session data (if available); currently, this is
16846 the peer's certificate chain and, for clients, the server's certificate
16847 and temporary key. CERT holds only those values that can have
16848 meaningful defaults in an SSL_CTX.
16852 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16853 from the internal representation. Various PKCS#7 fixes: remove some
16854 evil casts and set the enc_dig_alg field properly based on the signing
16859 * Allow PKCS#12 password to be set from the command line or the
16860 environment. Let 'ca' get its config file name from the environment
16861 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16866 * Allow certificate policies extension to use an IA5STRING for the
16867 organization field. This is contrary to the PKIX definition but
16868 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16873 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16874 without disallowing inline assembler and the like for non-pedantic builds.
16878 * Support Borland C++ builder.
16880 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16886 * SHA-1 cleanups and performance enhancements.
16888 *Andy Polyakov <appro@fy.chalmers.se>*
16890 * Sparc v8plus assembler for the bignum library.
16892 *Andy Polyakov <appro@fy.chalmers.se>*
16894 * Accept any -xxx and +xxx compiler options in Configure.
16898 * Update HPUX configuration.
16902 * Add missing sk_<type>_unshift() function to safestack.h
16904 *Ralf S. Engelschall*
16906 * New function SSL_CTX_use_certificate_chain_file that sets the
16907 "extra_cert"s in addition to the certificate. (This makes sense
16908 only for "PEM" format files, as chains as a whole are not
16913 * Support verify_depth from the SSL API.
16914 x509_vfy.c had what can be considered an off-by-one-error:
16915 Its depth (which was not part of the external interface)
16916 was actually counting the number of certificates in a chain;
16917 now it really counts the depth.
16921 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16922 instead of X509err, which often resulted in confusing error
16923 messages since the error codes are not globally unique
16924 (e.g. an alleged error in ssl3_accept when a certificate
16925 didn't match the private key).
16927 * New function SSL_CTX_set_session_id_context that allows to set a default
16928 value (so that you don't need SSL_set_session_id_context for each
16929 connection using the SSL_CTX).
16933 * OAEP decoding bug fix.
16937 * Support INSTALL_PREFIX for package builders, as proposed by
16942 * New Configure options "threads" and "no-threads". For systems
16943 where the proper compiler options are known (currently Solaris
16944 and Linux), "threads" is the default.
16948 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
16952 * Install various scripts to $(OPENSSLDIR)/misc, not to
16953 $(INSTALLTOP)/bin -- they shouldn't clutter directories
16954 such as /usr/local/bin.
16958 * "make linux-shared" to build shared libraries.
16960 *Niels Poppe <niels@netbox.org>*
16962 * New Configure option no-<cipher> (rsa, idea, rc5, ...).
16966 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
16967 extension adding in x509 utility.
16971 * Remove NOPROTO sections and error code comments.
16975 * Partial rewrite of the DEF file generator to now parse the ANSI
16980 * New Configure options --prefix=DIR and --openssldir=DIR.
16984 * Complete rewrite of the error code script(s). It is all now handled
16985 by one script at the top level which handles error code gathering,
16986 header rewriting and C source file generation. It should be much better
16987 than the old method: it now uses a modified version of Ulf's parser to
16988 read the ANSI prototypes in all header files (thus the old K&R definitions
16989 aren't needed for error creation any more) and do a better job of
16990 translating function codes into names. The old 'ASN1 error code embedded
16991 in a comment' is no longer necessary and it doesn't use .err files which
16992 have now been deleted. Also the error code call doesn't have to appear all
16993 on one line (which resulted in some large lines...).
16997 * Change #include filenames from <foo.h> to <openssl/foo.h>.
17001 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17002 0 (which usually indicates a closed connection), but continue reading.
17006 * Fix some race conditions.
17010 * Add support for CRL distribution points extension. Add Certificate
17011 Policies and CRL distribution points documentation.
17015 * Move the autogenerated header file parts to crypto/opensslconf.h.
17019 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17020 8 of keying material. Merlin has also confirmed interop with this fix
17021 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17023 *Merlin Hughes <merlin@baltimore.ie>*
17025 * Fix lots of warnings.
17027 *Richard Levitte <levitte@stacken.kth.se>*
17029 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17030 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17032 *Richard Levitte <levitte@stacken.kth.se>*
17034 * Fix problems with sizeof(long) == 8.
17036 *Andy Polyakov <appro@fy.chalmers.se>*
17038 * Change functions to ANSI C.
17042 * Fix typos in error codes.
17044 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17046 * Remove defunct assembler files from Configure.
17050 * SPARC v8 assembler BIGNUM implementation.
17052 *Andy Polyakov <appro@fy.chalmers.se>*
17054 * Support for Certificate Policies extension: both print and set.
17055 Various additions to support the r2i method this uses.
17059 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17060 return a const string when you are expecting an allocated buffer.
17064 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17065 types DirectoryString and DisplayText.
17069 * Add code to allow r2i extensions to access the configuration database,
17070 add an LHASH database driver and add several ctx helper functions.
17074 * Fix an evil bug in bn_expand2() which caused various BN functions to
17075 fail when they extended the size of a BIGNUM.
17079 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17080 support typesafe stack.
17084 * Fix typo in SSL_[gs]et_options().
17086 *Nils Frostberg <nils@medcom.se>*
17088 * Delete various functions and files that belonged to the (now obsolete)
17089 old X509V3 handling code.
17093 * New Configure option "rsaref".
17097 * Don't auto-generate pem.h.
17101 * Introduce type-safe ASN.1 SETs.
17105 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17107 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17109 * Introduce type-safe STACKs. This will almost certainly break lots of code
17110 that links with OpenSSL (well at least cause lots of warnings), but fear
17111 not: the conversion is trivial, and it eliminates loads of evil casts. A
17112 few STACKed things have been converted already. Feel free to convert more.
17113 In the fullness of time, I'll do away with the STACK type altogether.
17117 * Add `openssl ca -revoke <certfile>' facility which revokes a certificate
17118 specified in <certfile> by updating the entry in the index.txt file.
17119 This way one no longer has to edit the index.txt file manually for
17120 revoking a certificate. The -revoke option does the gory details now.
17122 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17124 * Fix `openssl crl -noout -text' combination where `-noout' killed the
17125 `-text' option at all and this way the `-noout -text' combination was
17126 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
17128 *Ralf S. Engelschall*
17130 * Make sure a corresponding plain text error message exists for the
17131 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17132 verify callback function determined that a certificate was revoked.
17134 *Ralf S. Engelschall*
17136 * Bugfix: In test/testenc, don't test "openssl <cipher>" for
17137 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17138 all available ciphers including rc5, which was forgotten until now.
17139 In order to let the testing shell script know which algorithms
17140 are available, a new (up to now undocumented) command
17141 "openssl list-cipher-commands" is used.
17145 * Bugfix: s_client occasionally would sleep in select() when
17146 it should have checked SSL_pending() first.
17150 * New functions DSA_do_sign and DSA_do_verify to provide access to
17151 the raw DSA values prior to ASN.1 encoding.
17155 * Tweaks to Configure
17157 *Niels Poppe <niels@netbox.org>*
17159 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17164 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17168 * New config option to avoid instructions that are illegal on the 80386.
17169 The default code is faster, but requires at least a 486.
17173 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17174 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17175 same as SSL2_VERSION anyway.
17179 * New "-showcerts" option for s_client.
17183 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17184 application. Various cleanups and fixes.
17188 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17189 modify error routines to work internally. Add error codes and PBE init
17190 to library startup routines.
17194 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17195 packing functions to asn1 and evp. Changed function names and error
17196 codes along the way.
17200 * PKCS12 integration: and so it begins... First of several patches to
17201 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17202 objects to objects.h
17206 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17207 and display support for Thawte strong extranet extension.
17211 * Add LinuxPPC support.
17213 *Jeff Dubrule <igor@pobox.org>*
17215 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17216 bn_div_words in alpha.s.
17218 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17220 * Make sure the RSA OAEP test is skipped under -DRSAref because
17221 OAEP isn't supported when OpenSSL is built with RSAref.
17223 *Ulf Moeller <ulf@fitug.de>*
17225 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17226 so they no longer are missing under -DNOPROTO.
17228 *Soren S. Jorvang <soren@t.dk>*
17231 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999] ###
17233 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17234 doesn't work when the session is reused. Coming soon!
17238 * Fix a security hole, that allows sessions to be reused in the wrong
17239 context thus bypassing client cert protection! All software that uses
17240 client certs and session caches in multiple contexts NEEDS PATCHING to
17241 allow session reuse! A fuller solution is in the works.
17243 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17245 * Some more source tree cleanups (removed obsolete files
17246 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17247 permission on "config" script to be executable) and a fix for the INSTALL
17250 *Ulf Moeller <ulf@fitug.de>*
17252 * Remove some legacy and erroneous uses of malloc, free instead of
17255 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17257 * Make rsa_oaep_test return non-zero on error.
17259 *Ulf Moeller <ulf@fitug.de>*
17261 * Add support for native Solaris shared libraries. Configure
17262 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17263 if someone would make that last step automatic.
17265 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17267 * ctx_size was not built with the right compiler during "make links". Fixed.
17271 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17272 except NULL ciphers". This means the default cipher list will no longer
17273 enable NULL ciphers. They need to be specifically enabled e.g. with
17274 the string "DEFAULT:eNULL".
17278 * Fix to RSA private encryption routines: if p < q then it would
17279 occasionally produce an invalid result. This will only happen with
17280 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17284 * Be less restrictive and allow also `perl util/perlpath.pl
17285 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17286 because this way one can also use an interpreter named `perl5' (which is
17287 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17288 installed as `perl').
17290 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17292 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17294 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17296 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17297 advapi32.lib to Win32 build and change the pem test comparison
17298 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17299 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17300 and crypto/des/ede_cbcm_enc.c.
17304 * DES quad checksum was broken on big-endian architectures. Fixed.
17308 * Comment out two functions in bio.h that aren't implemented. Fix up the
17309 Win32 test batch file so it (might) work again. The Win32 test batch file
17310 is horrible: I feel ill....
17314 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17315 in e_os.h. Audit of header files to check ANSI and non ANSI
17316 sections: 10 functions were absent from non ANSI section and not exported
17317 from Windows DLLs. Fixed up libeay.num for new functions.
17321 * Make `openssl version' output lines consistent.
17323 *Ralf S. Engelschall*
17325 * Fix Win32 symbol export lists for BIO functions: Added
17326 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17327 to ms/libeay{16,32}.def.
17329 *Ralf S. Engelschall*
17331 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17332 fine under Unix and passes some trivial tests I've now added. But the
17333 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17334 added to make sure no one expects that this stuff really works in the
17335 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17336 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17339 *Ralf S. Engelschall*
17341 * Fix the generation of two part addresses in perl.
17343 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17345 * Add config entry for Linux on MIPS.
17347 *John Tobey <jtobey@channel1.com>*
17349 * Make links whenever Configure is run, unless we are on Windoze.
17353 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17354 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17359 * Add a useful kludge to allow package maintainers to specify compiler and
17360 other platforms details on the command line without having to patch the
17361 Configure script every time: One now can use ``perl Configure
17362 <id>:<details>'', i.e. platform ids are allowed to have details appended
17363 to them (separated by colons). This is treated as there would be a static
17364 pre-configured entry in Configure's %table under key <id> with value
17365 <details> and ``perl Configure <id>'' is called. So, when you want to
17366 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17367 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
17368 now, which overrides the FreeBSD-elf entry on-the-fly.
17370 *Ralf S. Engelschall*
17372 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17376 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17377 on the `perl Configure ...' command line. This way one can compile
17378 OpenSSL libraries with Position Independent Code (PIC) which is needed
17379 for linking it into DSOs.
17381 *Ralf S. Engelschall*
17383 * Remarkably, export ciphers were totally broken and no-one had noticed!
17388 * Cleaned up the LICENSE document: The official contact for any license
17389 questions now is the OpenSSL core team under openssl-core@openssl.org.
17390 And add a paragraph about the dual-license situation to make sure people
17391 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17392 to the OpenSSL toolkit.
17394 *Ralf S. Engelschall*
17396 * General source tree makefile cleanups: Made `making xxx in yyy...'
17397 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17398 Additionally cleaned up the `make links' target: Remove unnecessary
17399 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17400 to speed processing and no longer clutter the display with confusing
17401 stuff. Instead only the actually done links are displayed.
17403 *Ralf S. Engelschall*
17405 * Permit null encryption ciphersuites, used for authentication only. It used
17406 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17407 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17412 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17413 signed attributes when verifying signatures (this would break them),
17414 the detached data encoding was wrong and public keys obtained using
17415 X509_get_pubkey() weren't freed.
17419 * Add text documentation for the BUFFER functions. Also added a work around
17420 to a Win95 console bug. This was triggered by the password read stuff: the
17421 last character typed gets carried over to the next fread(). If you were
17422 generating a new cert request using 'req' for example then the last
17423 character of the passphrase would be CR which would then enter the first
17428 * Added the new `Includes OpenSSL Cryptography Software' button as
17429 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17430 button and can be used by applications based on OpenSSL to show the
17431 relationship to the OpenSSL project.
17433 *Ralf S. Engelschall*
17435 * Remove confusing variables in function signatures in files
17436 ssl/ssl_lib.c and ssl/ssl.h.
17438 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17440 * Don't install bss_file.c under PREFIX/include/
17442 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17444 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17445 functions that return function pointers and has support for NT specific
17446 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17447 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17448 unsigned to signed types: this was killing the Win32 compile.
17452 * Add new certificate file to stack functions,
17453 SSL_add_dir_cert_subjects_to_stack() and
17454 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17455 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17456 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17457 This means that Apache-SSL and similar packages don't have to mess around
17458 to add as many CAs as they want to the preferred list.
17462 * Experiment with doxygen documentation. Currently only partially applied to
17464 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
17465 openssl.doxy as the configuration file.
17469 * Get rid of remaining C++-style comments which strict C compilers hate.
17471 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17473 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17474 compiled in by default: it has problems with large keys.
17478 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17479 DH private keys and/or callback functions which directly correspond to
17480 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17481 is needed for applications which have to configure certificates on a
17482 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17484 For the RSA certificate situation is makes no difference, but
17485 for the DSA certificate situation this fixes the "no shared cipher"
17486 problem where the OpenSSL cipher selection procedure failed because the
17487 temporary keys were not overtaken from the context and the API provided
17488 no way to reconfigure them.
17489 The new functions now let applications reconfigure the stuff and they
17490 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17491 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17492 non-public-API function ssl_cert_instantiate() is used as a helper
17493 function and also to reduce code redundancy inside ssl_rsa.c.
17495 *Ralf S. Engelschall*
17497 * Move s_server -dcert and -dkey options out of the undocumented feature
17498 area because they are useful for the DSA situation and should be
17499 recognized by the users.
17501 *Ralf S. Engelschall*
17503 * Fix the cipher decision scheme for export ciphers: the export bits are
17504 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17505 SSL_EXP_MASK. So, the original variable has to be used instead of the
17506 already masked variable.
17508 *Richard Levitte <levitte@stacken.kth.se>*
17510 * Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
17512 *Richard Levitte <levitte@stacken.kth.se>*
17514 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17515 from `int' to `unsigned int' because it's a length and initialized by
17516 EVP_DigestFinal() which expects an `unsigned int *'.
17518 *Richard Levitte <levitte@stacken.kth.se>*
17520 * Don't hard-code path to Perl interpreter on shebang line of Configure
17521 script. Instead use the usual Shell->Perl transition trick.
17523 *Ralf S. Engelschall*
17525 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17526 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17527 -noout -modulus' as it's already the case for `openssl rsa -noout
17528 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17529 currently the public key is printed (a decision which was already done by
17530 `openssl dsa -modulus' in the past) which serves a similar purpose.
17531 Additionally the NO_RSA no longer completely removes the whole -modulus
17532 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17535 *Ralf S. Engelschall*
17537 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17538 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17540 *Arne Ansper <arne@ats.cyber.ee>*
17542 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17543 to be added. Now both 'req' and 'ca' can use new objects defined in the
17548 * Add cool BIO that does syslog (or event log on NT).
17550 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17552 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17553 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17554 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17555 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17559 * Add preliminary config info for new extension code.
17563 * Make RSA_NO_PADDING really use no padding.
17565 *Ulf Moeller <ulf@fitug.de>*
17567 * Generate errors when private/public key check is done.
17571 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17572 for some CRL extensions and new objects added.
17576 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17577 key usage extension and fuller support for authority key id.
17581 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17582 padding method for RSA, which is recommended for new applications in PKCS
17583 #1 v2.0 (RFC 2437, October 1998).
17584 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17585 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17586 against Bleichbacher's attack on RSA.
17587 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17590 * Updates to the new SSL compression code
17592 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17594 * Fix so that the version number in the master secret, when passed
17595 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17596 (because the server will not accept higher), that the version number
17597 is 0x03,0x01, not 0x03,0x00
17599 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17601 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17602 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17603 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17607 * Support for RAW extensions where an arbitrary extension can be
17608 created by including its DER encoding. See apps/openssl.cnf for
17613 * Make sure latest Perl versions don't interpret some generated C array
17614 code as Perl array code in the crypto/err/err_genc.pl script.
17616 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17618 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17619 not many people have the assembler. Various Win32 compilation fixes and
17620 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17621 build instructions.
17625 * Modify configure script 'Configure' to automatically create crypto/date.h
17626 file under Win32 and also build pem.h from pem.org. New script
17627 util/mkfiles.pl to create the MINFO file on environments that can't do a
17628 'make files': perl util/mkfiles.pl >MINFO should work.
17632 * Major rework of DES function declarations, in the pursuit of correctness
17633 and purity. As a result, many evil casts evaporated, and some weirdness,
17634 too. You may find this causes warnings in your code. Zapping your evil
17635 casts will probably fix them. Mostly.
17639 * Fix for a typo in asn1.h. Bug fix to object creation script
17640 obj_dat.pl. It considered a zero in an object definition to mean
17641 "end of object": none of the objects in objects.h have any zeros
17642 so it wasn't spotted.
17644 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17646 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17647 Masking (CBCM). In the absence of test vectors, the best I have been able
17648 to do is check that the decrypt undoes the encrypt, so far. Send me test
17649 vectors if you have them.
17653 * Correct calculation of key length for export ciphers (too much space was
17654 allocated for null ciphers). This has not been tested!
17658 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17659 message is now correct (it understands "crypto" and "ssl" on its
17660 command line). There is also now an "update" option. This will update
17661 the util/ssleay.num and util/libeay.num files with any new functions.
17663 perl util/mkdef.pl crypto ssl update
17664 it will update them.
17668 * Overhauled the Perl interface (perl/*):
17669 - ported BN stuff to OpenSSL's different BN library
17670 - made the perl/ source tree CVS-aware
17671 - renamed the package from SSLeay to OpenSSL (the files still contain
17672 their history because I've copied them in the repository)
17673 - removed obsolete files (the test scripts will be replaced
17674 by better Test::Harness variants in the future)
17676 *Ralf S. Engelschall*
17678 * First cut for a very conservative source tree cleanup:
17679 1. merge various obsolete readme texts into doc/ssleay.txt
17680 where we collect the old documents and readme texts.
17681 2. remove the first part of files where I'm already sure that we no
17682 longer need them because of three reasons: either they are just temporary
17683 files which were left by Eric or they are preserved original files where
17684 I've verified that the diff is also available in the CVS via "cvs diff
17685 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17686 the crypto/md/ stuff).
17688 *Ralf S. Engelschall*
17690 * More extension code. Incomplete support for subject and issuer alt
17691 name, issuer and authority key id. Change the i2v function parameters
17692 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17693 what that's for :-) Fix to ASN1 macro which messed up
17694 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17698 * Preliminary support for ENUMERATED type. This is largely copied from the
17703 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17705 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17707 * Make sure `make rehash' target really finds the `openssl' program.
17709 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17711 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17712 like to hear about it if this slows down other processors.
17716 * Add CygWin32 platform information to Configure script.
17718 *Alan Batie <batie@aahz.jf.intel.com>*
17720 * Fixed ms/32all.bat script: `no_asm' -> `no-asm'
17722 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17724 * New program nseq to manipulate netscape certificate sequences
17728 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17733 * Fixes to BN code. Previously the default was to define BN_RECURSION
17734 but the BN code had some problems that would cause failures when
17735 doing certificate verification and some other functions.
17737 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17739 * Add ASN1 and PEM code to support netscape certificate sequences.
17743 * Add ASN1 and PEM code to support netscape certificate sequences.
17747 * Add several PKIX and private extended key usage OIDs.
17751 * Modify the 'ca' program to handle the new extension code. Modify
17752 openssl.cnf for new extension format, add comments.
17756 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17757 and add a sample to openssl.cnf so req -x509 now adds appropriate
17762 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17763 error code, add initial support to X509_print() and x509 application.
17767 * Takes a deep breath and start adding X509 V3 extension support code. Add
17768 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17769 stuff is currently isolated and isn't even compiled yet.
17773 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17774 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17775 Removed the versions check from X509 routines when loading extensions:
17776 this allows certain broken certificates that don't set the version
17777 properly to be processed.
17781 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17782 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17783 can still be regenerated with "make depend".
17787 * Spelling mistake in C version of CAST-128.
17789 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17791 * Changes to the error generation code. The perl script err-code.pl
17792 now reads in the old error codes and retains the old numbers, only
17793 adding new ones if necessary. It also only changes the .err files if new
17794 codes are added. The makefiles have been modified to only insert errors
17795 when needed (to avoid needlessly modifying header files). This is done
17796 by only inserting errors if the .err file is newer than the auto generated
17797 C file. To rebuild all the error codes from scratch (the old behaviour)
17798 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17799 or delete all the .err files.
17803 * CAST-128 was incorrectly implemented for short keys. The C version has
17804 been fixed, but is untested. The assembler versions are also fixed, but
17805 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17806 to regenerate it if needed.
17807 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17808 Hagino <itojun@kame.net>*
17810 * File was opened incorrectly in randfile.c.
17812 *Ulf Möller <ulf@fitug.de>*
17814 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17815 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17816 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17817 al: it's just almost always a UTCTime. Note this patch adds new error
17818 codes so do a "make errors" if there are problems.
17822 * Correct Linux 1 recognition in config.
17824 *Ulf Möller <ulf@fitug.de>*
17826 * Remove pointless MD5 hash when using DSA keys in ca.
17828 *Anonymous <nobody@replay.com>*
17830 * Generate an error if given an empty string as a cert directory. Also
17831 generate an error if handed NULL (previously returned 0 to indicate an
17832 error, but didn't set one).
17834 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17836 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17840 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17841 parameters. This was causing a warning which killed off the Win32 compile.
17845 * Remove C++ style comments from crypto/bn/bn_local.h.
17847 *Neil Costigan <neil.costigan@celocom.com>*
17849 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17850 based on a text string, looking up short and long names and finally
17851 "dot" format. The "dot" format stuff didn't work. Added new function
17852 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17853 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17854 OID is not part of the table.
17858 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17859 X509_LOOKUP_by_alias().
17863 * Sort openssl functions by name.
17867 * Get the gendsa program working (hopefully) and add it to app list. Remove
17868 encryption from sample DSA keys (in case anyone is interested the password
17873 * Make _all_ *_free functions accept a NULL pointer.
17875 *Frans Heymans <fheymans@isaserver.be>*
17877 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17880 *Anonymous <nobody@replay.com>*
17882 * s_server should send the CAfile as acceptable CAs, not its own cert.
17884 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17886 * Don't blow it for numeric -newkey arguments to apps/req.
17888 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17890 * Temp key "for export" tests were wrong in s3_srvr.c.
17892 *Anonymous <nobody@replay.com>*
17894 * Add prototype for temp key callback functions
17895 SSL_CTX_set_tmp_{rsa,dh}_callback().
17899 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17900 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17904 * X509_name_add_entry() freed the wrong thing after an error.
17906 *Arne Ansper <arne@ats.cyber.ee>*
17908 * rsa_eay.c would attempt to free a NULL context.
17910 *Arne Ansper <arne@ats.cyber.ee>*
17912 * BIO_s_socket() had a broken should_retry() on Windoze.
17914 *Arne Ansper <arne@ats.cyber.ee>*
17916 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17918 *Arne Ansper <arne@ats.cyber.ee>*
17920 * Make sure the already existing X509_STORE->depth variable is initialized
17921 in X509_STORE_new(), but document the fact that this variable is still
17922 unused in the certificate verification process.
17924 *Ralf S. Engelschall*
17926 * Fix the various library and apps files to free up pkeys obtained from
17927 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17931 * Fix reference counting in X509_PUBKEY_get(). This makes
17932 demos/maurice/example2.c work, amongst others, probably.
17934 *Steve Henson and Ben Laurie*
17936 * First cut of a cleanup for apps/. First the `ssleay' program is now named
17937 `openssl' and second, the shortcut symlinks for the `openssl <command>'
17938 are no longer created. This way we have a single and consistent command
17939 line interface `openssl <command>', similar to `cvs <command>'.
17941 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
17943 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
17944 BIT STRING wrapper always have zero unused bits.
17948 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
17952 * Make the top-level INSTALL documentation easier to understand.
17956 * Makefiles updated to exit if an error occurs in a sub-directory
17957 make (including if user presses ^C) [Paul Sutton]
17959 * Make Montgomery context stuff explicit in RSA data structure.
17963 * Fix build order of pem and err to allow for generated pem.h.
17967 * Fix renumbering bug in X509_NAME_delete_entry().
17971 * Enhanced the err-ins.pl script so it makes the error library number
17972 global and can add a library name. This is needed for external ASN1 and
17973 other error libraries.
17977 * Fixed sk_insert which never worked properly.
17981 * Fix ASN1 macros so they can handle indefinite length constructed
17982 EXPLICIT tags. Some non standard certificates use these: they can now
17987 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
17988 into a single doc/ssleay.txt bundle. This way the information is still
17989 preserved but no longer messes up this directory. Now it's new room for
17990 the new set of documentation files.
17992 *Ralf S. Engelschall*
17994 * SETs were incorrectly DER encoded. This was a major pain, because they
17995 shared code with SEQUENCEs, which aren't coded the same. This means that
17996 almost everything to do with SETs or SEQUENCEs has either changed name or
17997 number of arguments.
17999 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18001 * Fix test data to work with the above.
18005 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18006 was already fixed by Eric for 0.9.1 it seems.
18008 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18010 * Autodetect FreeBSD3.
18014 * Fix various bugs in Configure. This affects the following platforms:
18018 unixware-2.0-pentium
18023 * Eliminate generated files from CVS. Reorder tests to regenerate files
18024 before they are needed.
18028 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18033 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998] ###
18035 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18036 changed SSLeay to OpenSSL in version strings.
18038 *Ralf S. Engelschall*
18040 * Some fixups to the top-level documents.
18044 * Fixed the nasty bug where rsaref.h was not found under compile-time
18045 because the symlink to include/ was missing.
18047 *Ralf S. Engelschall*
18049 * Incorporated the popular no-RSA/DSA-only patches
18050 which allow to compile a RSA-free SSLeay.
18052 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18054 * Fixed nasty rehash problem under `make -f Makefile.ssl links'
18055 when "ssleay" is still not found.
18057 *Ralf S. Engelschall*
18059 * Added more platforms to Configure: Cray T3E, HPUX 11,
18061 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18063 * Updated the README file.
18065 *Ralf S. Engelschall*
18067 * Added various .cvsignore files in the CVS repository subdirs
18068 to make a "cvs update" really silent.
18070 *Ralf S. Engelschall*
18072 * Recompiled the error-definition header files and added
18073 missing symbols to the Win32 linker tables.
18075 *Ralf S. Engelschall*
18077 * Cleaned up the top-level documents;
18078 o new files: CHANGES and LICENSE
18079 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18080 o merged COPYRIGHT into LICENSE
18081 o removed obsolete TODO file
18082 o renamed MICROSOFT to INSTALL.W32
18084 *Ralf S. Engelschall*
18086 * Removed dummy files from the 0.9.1b source tree:
18087 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18088 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18089 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18090 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18091 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18093 *Ralf S. Engelschall*
18095 * Added various platform portability fixes.
18099 * The Genesis of the OpenSSL rpject:
18100 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18101 Young and Tim J. Hudson created while they were working for C2Net until
18104 *The OpenSSL Project*
18107 ### Changes between 0.9.0b and 0.9.1b [not released] ###
18109 * Updated a few CA certificates under certs/
18113 * Changed some BIGNUM api stuff.
18117 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18118 DGUX x86, Linux Alpha, etc.
18122 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18123 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18128 * Add -strparse option to asn1pars program which parses nested
18131 *Dr Stephen Henson <shenson@bigfoot.com>*
18133 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18137 * DSA fix for "ca" program.
18141 * Added "-genkey" option to "dsaparam" program.
18145 * Added RIPE MD160 (rmd160) message digest.
18149 * Added -a (all) option to "ssleay version" command.
18153 * Added PLATFORM define which is the id given to Configure.
18157 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18161 * Extended the ASN.1 parser routines.
18165 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18169 * Added a BN_CTX to the BN library.
18173 * Fixed the weak key values in DES library
18177 * Changed API in EVP library for cipher aliases.
18181 * Added support for RC2/64bit cipher.
18185 * Converted the lhash library to the crypto/mem.c functions.
18189 * Added more recognized ASN.1 object ids.
18193 * Added more RSA padding checks for SSL/TLS.
18197 * Added BIO proxy/filter functionality.
18201 * Added extra_certs to SSL_CTX which can be used
18202 send extra CA certificates to the client in the CA cert chain sending
18203 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18207 * Now Fortezza is denied in the authentication phase because
18208 this is key exchange mechanism is not supported by SSLeay at all.
18212 * Additional PKCS1 checks.
18216 * Support the string "TLSv1" for all TLS v1 ciphers.
18220 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18221 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18225 * Fixed a few memory leaks.
18229 * Fixed various code and comment typos.
18233 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18234 bytes sent in the client random.
18236 *Edward Bishop <ebishop@spyglass.com>*
18241 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18242 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18243 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18244 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18245 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18246 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18247 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18248 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18249 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18250 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18251 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18252 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18253 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18254 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18255 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18256 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18257 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18258 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18259 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18260 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18261 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18262 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18263 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18264 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18265 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18266 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18267 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18268 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18269 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18270 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18271 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18272 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18273 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18274 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18275 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18276 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18277 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18278 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18279 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18280 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18281 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18282 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18283 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18284 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18285 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18286 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18287 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18288 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18289 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18290 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18291 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18292 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18293 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18294 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18295 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18296 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18297 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18298 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18299 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18300 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18301 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18302 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18303 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18304 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18305 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18306 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18307 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18308 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18309 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18310 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18311 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18312 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18313 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18314 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18315 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18316 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18317 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18318 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18319 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18320 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18321 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18322 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18323 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18324 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18325 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18326 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18327 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18328 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18329 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18330 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18331 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18332 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18333 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18334 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18335 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18336 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18337 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18338 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18339 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18340 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18341 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18342 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18343 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18344 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18345 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18346 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18347 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18348 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18349 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18350 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18351 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18352 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18353 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18354 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18355 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18356 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18357 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18358 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18359 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18360 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18361 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18362 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18363 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18364 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18365 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18366 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18367 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18368 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18369 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18370 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18371 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18372 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18373 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18374 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18375 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18376 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18377 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18378 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18379 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18380 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18381 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18382 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18383 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18384 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18385 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18386 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18387 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18388 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18389 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18390 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18391 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18392 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18393 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18394 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18395 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18396 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18397 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18398 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18399 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18400 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655