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 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
28 contain a provider side internal key.
32 * `ASN1_verify()`, `ASN1_digest()` and `ASN1_sign()` have been deprecated.
33 They are old functions that we don't use, and that you could disable with
34 the macro `NO_ASN1_OLD`. This goes all the way back to OpenSSL 0.9.7.
38 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
39 have been converted to Markdown with the goal to produce documents
40 which not only look pretty when viewed online in the browser, but
41 remain well readable inside a plain text editor.
43 To achieve this goal, a 'minimalistic' Markdown style has been applied
44 which avoids formatting elements that interfere too much with the
45 reading flow in the text file. For example, it
47 * avoids [ATX headings][] and uses [setext headings][] instead
48 (which works for `<h1>` and `<h2>` headings only).
49 * avoids [inline links][] and uses [reference links][] instead.
50 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
52 [ATX headings]: https://github.github.com/gfm/#atx-headings
53 [setext headings]: https://github.github.com/gfm/#setext-headings
54 [inline links]: https://github.github.com/gfm/#inline-link
55 [reference links]: https://github.github.com/gfm/#reference-link
56 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
57 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
62 * The test suite is changed to preserve results of each test recipe.
63 A new directory test-runs/ with subdirectories named like the
64 test recipes are created in the build tree for this purpose.
68 * The command line utilities ecparam and ec have been deprecated. Instead
69 use the pkeyparam, pkey and genpkey programs.
73 * All of the low level RSA functions have been deprecated including:
75 RSA_new_method, RSA_bits, RSA_size, RSA_security_bits,
76 RSA_get0_pss_params, RSA_get_version, RSA_get0_engine,
77 RSA_generate_key_ex, RSA_generate_multi_prime_key,
78 RSA_X931_derive_ex, RSA_X931_generate_key_ex, RSA_check_key,
79 RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
80 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
81 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
82 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
83 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING,
84 RSA_blinding_on, RSA_blinding_off, RSA_setup_blinding,
85 RSA_padding_add_PKCS1_type_1, RSA_padding_check_PKCS1_type_1,
86 RSA_padding_add_PKCS1_type_2, RSA_padding_check_PKCS1_type_2,
87 PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP, RSA_padding_check_PKCS1_OAEP,
88 RSA_padding_add_PKCS1_OAEP_mgf1, RSA_padding_check_PKCS1_OAEP_mgf1,
89 RSA_padding_add_SSLv23, RSA_padding_check_SSLv23,
90 RSA_padding_add_none, RSA_padding_check_none, RSA_padding_add_X931,
91 RSA_padding_check_X931, RSA_X931_hash_id, RSA_verify_PKCS1_PSS,
92 RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
93 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
94 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
95 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
96 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
97 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
98 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
99 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
100 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
101 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
102 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
103 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
104 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
106 Use of these low level functions has been informally discouraged for a long
107 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
108 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
109 L<EVP_PKEY_decrypt(3)>.
113 * X509 certificates signed using SHA1 are no longer allowed at security
115 In TLS/SSL the default security level is 1. It can be set either
116 using the cipher string with @SECLEVEL, or calling
117 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
118 a call to SSL_CTX_use_certificate() will fail if the security level is not
120 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
121 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
126 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
127 deprecated. Instead use the pkeyparam, pkey, genpkey and pkeyparam
128 programs respectively.
132 * All of the low level DH functions have been deprecated including:
134 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
135 DH_new_method, DH_bits, DH_size, DH_security_bits, DH_get_ex_new_index,
136 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
137 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
138 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
139 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
140 DH_KDF_X9_42, DH_get0_engine, DH_get_length, DH_set_length, DH_meth_new,
141 DH_meth_free, DH_meth_dup, DH_meth_get0_name, DH_meth_set1_name,
142 DH_meth_get_flags, DH_meth_set_flags, DH_meth_get0_app_data,
143 DH_meth_set0_app_data, DH_meth_get_generate_key,
144 DH_meth_set_generate_key, DH_meth_get_compute_key,
145 DH_meth_set_compute_key, DH_meth_get_bn_mod_exp,
146 DH_meth_set_bn_mod_exp, DH_meth_get_init, DH_meth_set_init,
147 DH_meth_get_finish, DH_meth_set_finish, DH_meth_get_generate_params
148 and DH_meth_set_generate_params.
150 Use of these low level functions has been informally discouraged for a long
151 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
152 and L<EVP_PKEY_derive(3)>.
156 * All of the low level DSA functions have been deprecated including:
158 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
159 DSA_get_default_method, DSA_set_method, DSA_get_method, DSA_new_method,
160 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
161 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
162 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
163 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
164 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
165 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
166 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
167 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
168 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
169 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
170 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
172 Use of these low level functions has been informally discouraged for a long
173 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
174 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
178 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
179 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
180 This means that applications don't have to look at the curve NID and
181 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
182 However, they still can, that EVP_PKEY_set_alias_type() call acts as
183 a no-op when the EVP_PKEY is already of the given type.
185 Parameter and key generation is also reworked to make it possible
186 to generate EVP_PKEY_SM2 parameters and keys without having to go
187 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
188 However, code that does the latter will still work as before.
192 * Deprecated low level ECDH and ECDSA functions. These include:
194 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
195 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
198 Use of these low level functions has been informally discouraged for a long
199 time. Instead applications should use the EVP_PKEY_derive(3),
200 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
204 * Deprecated the EC_KEY_METHOD functions. These include:
206 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
207 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
208 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
209 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
210 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
211 EC_KEY_METHOD_get_verify.
213 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
217 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
218 and EVP_PKEY_decrypt() instead.
219 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
220 and EVP_PKEY_encrypt() instead.
224 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
225 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
226 a new formulation to include all the things it can be used for,
227 as well as words of caution.
231 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
232 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
236 * All of the low level HMAC functions have been deprecated including:
238 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
239 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
242 Use of these low level functions has been informally discouraged for a long
243 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
244 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
245 and L<EVP_MAC_final(3)>.
249 * Over two thousand fixes were made to the documentation, including:
250 - Common options (such as -rand/-writerand, TLS version control, etc)
251 were refactored and point to newly-enhanced descriptions in openssl.pod.
252 - Added style conformance for all options (with help from Richard Levitte),
253 documented all reported missing options, added a CI build to check
254 that all options are documented and that no unimplemented options
256 - Documented some internals, such as all use of environment variables.
257 - Addressed all internal broken L<> references.
261 * All of the low level CMAC functions have been deprecated including:
263 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
264 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
266 Use of these low level functions has been informally discouraged for a long
267 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
268 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
269 and L<EVP_MAC_final(3)>.
273 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
274 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
277 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
278 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
279 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
280 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
281 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
282 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
283 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
284 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
285 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
286 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
288 Use of these low level functions has been informally discouraged
289 for a long time. Applications should use the EVP_DigestInit_ex(3),
290 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
294 * Corrected the documentation of the return values from the EVP_DigestSign*
295 set of functions. The documentation mentioned negative values for some
296 errors, but this was never the case, so the mention of negative values
299 Code that followed the documentation and thereby check with something
300 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
304 * All of the low level cipher functions have been deprecated including:
306 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
307 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
308 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
309 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
310 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
311 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
312 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
313 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
314 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
315 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
316 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
317 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
318 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
319 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
320 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
321 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
322 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
323 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
324 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
325 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
326 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
327 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
328 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
329 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
330 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
331 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
332 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
333 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
334 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
336 Use of these low level functions has been informally discouraged for
337 a long time. Applications should use the high level EVP APIs, e.g.
338 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
339 equivalently named decrypt functions instead.
341 *Matt Caswell and Paul Dale*
343 * Removed include/openssl/opensslconf.h.in and replaced it with
344 include/openssl/configuration.h.in, which differs in not including
345 <openssl/macros.h>. A short header include/openssl/opensslconf.h
346 was added to include both.
348 This allows internal hacks where one might need to modify the set
349 of configured macros, for example this if deprecated symbols are
350 still supposed to be available internally:
352 #include <openssl/configuration.h>
354 #undef OPENSSL_NO_DEPRECATED
355 #define OPENSSL_SUPPRESS_DEPRECATED
357 #include <openssl/macros.h>
359 This should not be used by applications that use the exported
360 symbols, as that will lead to linking errors.
364 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
365 used in exponentiation with 512-bit moduli. No EC algorithms are
366 affected. Analysis suggests that attacks against 2-prime RSA1024,
367 3-prime RSA1536, and DSA1024 as a result of this defect would be very
368 difficult to perform and are not believed likely. Attacks against DH512
369 are considered just feasible. However, for an attack the target would
370 have to re-use the DH512 private key, which is not recommended anyway.
371 Also applications directly using the low level API BN_mod_exp may be
372 affected if they use BN_FLG_CONSTTIME.
377 * Most memory-debug features have been deprecated, and the functionality
378 replaced with no-ops.
382 * Introduced a new method type and API, OSSL_SERIALIZER, to
383 represent generic serializers. An implementation is expected to
384 be able to serialize an object associated with a given name (such
385 as an algorithm name for an asymmetric key) into forms given by
386 implementation properties.
388 Serializers are primarily used from inside libcrypto, through
389 calls to functions like EVP_PKEY_print_private(),
390 PEM_write_bio_PrivateKey() and similar.
392 Serializers are specified in such a way that they can be made to
393 directly handle the provider side portion of an object, if this
394 provider side part comes from the same provider as the serializer
395 itself, but can also be made to handle objects in parametrized
396 form (as an OSSL_PARAM array of data). This allows a provider to
397 offer generic serializers as a service for any other provider.
401 * Added a .pragma directive to the syntax of configuration files, to
402 allow varying behavior in a supported and predictable manner.
403 Currently added pragma:
407 This allows dollar signs to be a keyword character unless it's
408 followed by a opening brace or parenthesis. This is useful for
409 platforms where dollar signs are commonly used in names, such as
410 volume names and system directory names on VMS.
414 * Added functionality to create an EVP_PKEY from user data. This
415 is effectively the same as creating a RSA, DH or DSA object and
416 then assigning them to an EVP_PKEY, but directly using algorithm
417 agnostic EVP functions. A benefit is that this should be future
418 proof for public key algorithms to come.
422 * Change the interpretation of the '--api' configuration option to
423 mean that this is a desired API compatibility level with no
424 further meaning. The previous interpretation, that this would
425 also mean to remove all deprecated symbols up to and including
426 the given version, no requires that 'no-deprecated' is also used
427 in the configuration.
429 When building applications, the desired API compatibility level
430 can be set with the OPENSSL_API_COMPAT macro like before. For
431 API compatibility version below 3.0, the old style numerical
432 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
433 For version 3.0 and on, the value is expected to be the decimal
434 value calculated from the major and minor version like this:
436 MAJOR * 10000 + MINOR * 100
440 -DOPENSSL_API_COMPAT=30000 For 3.0
441 -DOPENSSL_API_COMPAT=30200 For 3.2
443 To hide declarations that are deprecated up to and including the
444 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
445 given when building the application as well.
449 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
450 access to certificate and CRL stores via URIs and OSSL_STORE
453 This adds the following functions:
455 - X509_LOOKUP_store()
456 - X509_STORE_load_file()
457 - X509_STORE_load_path()
458 - X509_STORE_load_store()
459 - SSL_add_store_cert_subjects_to_stack()
460 - SSL_CTX_set_default_verify_store()
461 - SSL_CTX_load_verify_file()
462 - SSL_CTX_load_verify_dir()
463 - SSL_CTX_load_verify_store()
465 Also, the following functions are now deprecated:
467 - X509_STORE_load_locations() (use X509_STORE_load_file(),
468 X509_STORE_load_path() or X509_STORE_load_store() instead)
469 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
470 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
474 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
475 The presence of this system service is determined at run-time.
479 * Added functionality to create an EVP_PKEY context based on data
480 for methods from providers. This takes an algorithm name and a
481 property query string and simply stores them, with the intent
482 that any operation that uses this context will use those strings
483 to fetch the needed methods implicitly, thereby making the port
484 of application written for pre-3.0 OpenSSL easier.
488 * The undocumented function NCONF_WIN32() has been deprecated; for
489 conversion details see the HISTORY section of doc/man5/config.pod
493 * Introduced the new functions EVP_DigestSignInit_ex() and
494 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
495 EVP_DigestVerifyUpdate() have been converted to functions. See the man
496 pages for further details.
500 * Over two thousand fixes were made to the documentation, including:
501 adding missing command flags, better style conformance, documentation
504 *Rich Salz, Richard Levitte*
506 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
507 X25519, X448, Ed25519 and Ed448.
511 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
516 * Deprecated the public definition of ERR_STATE as well as the function
517 ERR_get_state(). This is done in preparation of making ERR_STATE an
522 * Added ERR functionality to give callers access to the stored function
523 names that have replaced the older function code based functions.
525 New functions are ERR_get_error_func(), ERR_peek_error_func(),
526 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
527 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
528 and ERR_peek_last_error_all().
530 These functions have become deprecated: ERR_get_error_line_data(),
531 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
532 ERR_func_error_string().
536 * Extended testing to be verbose for failing tests only. The make variables
537 VERBOSE_FAILURE or VF can be used to enable this:
539 $ make VF=1 test # Unix
540 $ mms /macro=(VF=1) test ! OpenVMS
541 $ nmake VF=1 test # Windows
546 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
547 used even when parsing explicit parameters, when loading a serialized key
548 or calling `EC_GROUP_new_from_ecpkparameters()`/
549 `EC_GROUP_new_from_ecparameters()`.
550 This prevents bypass of security hardening and performance gains,
551 especially for curves with specialized EC_METHODs.
552 By default, if a key encoded with explicit parameters is loaded and later
553 serialized, the output is still encoded with explicit parameters, even if
554 internally a "named" EC_GROUP is used for computation.
558 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
559 this change, EC_GROUP_set_generator would accept order and/or cofactor as
560 NULL. After this change, only the cofactor parameter can be NULL. It also
561 does some minimal sanity checks on the passed order.
566 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
567 An attack is simple, if the first CMS_recipientInfo is valid but the
568 second CMS_recipientInfo is chosen ciphertext. If the second
569 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
570 encryption key will be replaced by garbage, and the message cannot be
571 decoded, but if the RSA decryption fails, the correct encryption key is
572 used and the recipient will not notice the attack.
573 As a work around for this potential attack the length of the decrypted
574 key must be equal to the cipher default key length, in case the
575 certifiate is not given and all recipientInfo are tried out.
576 The old behaviour can be re-enabled in the CMS code by setting the
577 CMS_DEBUG_DECRYPT flag.
581 * Early start up entropy quality from the DEVRANDOM seed source has been
582 improved for older Linux systems. The RAND subsystem will wait for
583 /dev/random to be producing output before seeding from /dev/urandom.
584 The seeded state is stored for future library initialisations using
585 a system global shared memory segment. The shared memory identifier
586 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
587 the desired value. The default identifier is 114.
591 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
592 when primes for RSA keys are computed.
593 Since we previously always generated primes == 2 (mod 3) for RSA keys,
594 the 2-prime and 3-prime RSA modules were easy to distinguish, since
595 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
596 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
597 This avoids possible fingerprinting of newly generated RSA modules.
601 * Correct the extended master secret constant on EBCDIC systems. Without this
602 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
603 negotiate EMS will fail. Unfortunately this also means that TLS connections
604 between EBCDIC systems with this fix, and EBCDIC systems without this
605 fix will fail if they negotiate EMS.
609 * Changed the library initialisation so that the config file is now loaded
610 by default. This was already the case for libssl. It now occurs for both
611 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
612 OPENSSL_init_crypto() to suppress automatic loading of a config file.
616 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
617 where the former acts as a replacement for ERR_put_error(), and the
618 latter replaces the combination ERR_put_error()+ERR_add_error_data().
619 ERR_raise_data() adds more flexibility by taking a format string and
620 an arbitrary number of arguments following it, to be processed with
625 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
626 to check if a named provider is loaded and available. When called, it
627 will also activate all fallback providers if such are still present.
631 * Enforce a minimum DH modulus size of 512 bits.
635 * Changed DH parameters to generate the order q subgroup instead of 2q.
636 Previously generated DH parameters are still accepted by DH_check
637 but DH_generate_key works around that by clearing bit 0 of the
638 private key for those. This avoids leaking bit 0 of the private key.
642 * Significantly reduce secure memory usage by the randomness pools.
646 * {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
651 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
652 algorithms. An implementation of a key exchange algorithm can be obtained
653 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
654 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
655 the older EVP_PKEY_derive_init() function. See the man pages for the new
656 functions for further details.
660 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
664 * Removed the function names from error messages and deprecated the
667 * Removed NextStep support and the macro OPENSSL_UNISTD
671 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
672 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
673 Also removed "export var as function" capability; we do not export
674 variables, only functions.
678 * RC5_32_set_key has been changed to return an int type, with 0 indicating
679 an error and 1 indicating success. In previous versions of OpenSSL this
680 was a void type. If a key was set longer than the maximum possible this
685 * Support SM2 signing and verification schemes with X509 certificate.
689 * Use SHA256 as the default digest for TS query in the ts app.
693 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
694 This checks that the salt length is at least 128 bits, the derived key
695 length is at least 112 bits, and that the iteration count is at least 1000.
696 For backwards compatibility these checks are disabled by default in the
697 default provider, but are enabled by default in the fips provider.
698 To enable or disable these checks use the control
699 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
703 * Default cipher lists/suites are now available via a function, the
704 #defines are deprecated.
708 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
709 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
710 for Windows Store apps easier. Also, the "no-uplink" option has been added.
714 * Join the directories crypto/x509 and crypto/x509v3
718 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
719 This changes the size when using the genpkey app when no size is given. It
720 fixes an omission in earlier changes that changed all RSA, DSA and DH
721 generation apps to use 2048 bits by default.
725 * Added command 'openssl kdf' that uses the EVP_KDF API.
729 * Added command 'openssl mac' that uses the EVP_MAC API.
733 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
734 as default directories. Also added the command 'openssl info'
735 for scripting purposes.
739 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
740 deprecated. These undocumented functions were never integrated into the EVP
741 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
742 Bi-directional IGE mode. These modes were never formally standardised and
743 usage of these functions is believed to be very small. In particular
744 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
745 is ever used. The security implications are believed to be minimal, but
746 this issue was never fixed for backwards compatibility reasons. New code
747 should not use these modes.
751 * Add prediction resistance to the DRBG reseeding process.
755 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
756 mandated by IEEE Std 1619-2018.
760 * Added newline escaping functionality to a filename when using openssl dgst.
761 This output format is to replicate the output format found in the '*sum'
762 checksum programs. This aims to preserve backward compatibility.
764 *Matt Eaton, Richard Levitte, and Paul Dale*
766 * Removed the heartbeat message in DTLS feature, as it has very
767 little usage and doesn't seem to fulfill a valuable purpose.
768 The configuration option is now deprecated.
772 * Changed the output of 'openssl {digestname} < file' to display the
773 digest name in its output.
777 * Added a new generic trace API which provides support for enabling
778 instrumentation through trace output. This feature is mainly intended
779 as an aid for developers and is disabled by default. To utilize it,
780 OpenSSL needs to be configured with the `enable-trace` option.
782 If the tracing API is enabled, the application can activate trace output
783 by registering BIOs as trace channels for a number of tracing and debugging
786 The 'openssl' application has been expanded to enable any of the types
787 available via environment variables defined by the user, and serves as
788 one possible example on how to use this functionality.
790 *Richard Levitte & Matthias St. Pierre*
792 * Added build tests for C++. These are generated files that only do one
793 thing, to include one public OpenSSL head file each. This tests that
794 the public header files can be usefully included in a C++ application.
796 This test isn't enabled by default. It can be enabled with the option
797 'enable-buildtest-c++'.
801 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
805 * Add KMAC to EVP_MAC.
809 * Added property based algorithm implementation selection framework to
814 * Added SCA hardening for modular field inversion in EC_GROUP through
815 a new dedicated field_inv() pointer in EC_METHOD.
816 This also addresses a leakage affecting conversions from projective
817 to affine coordinates.
819 *Billy Bob Brumley, Nicola Tuveri*
821 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
822 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
823 those algorithms that were already supported through the EVP_PKEY API
824 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
825 and scrypt are now wrappers that call EVP_KDF.
829 * Build devcrypto engine as a dynamic engine.
833 * Add keyed BLAKE2 to EVP_MAC.
837 * Fix a bug in the computation of the endpoint-pair shared secret used
838 by DTLS over SCTP. This breaks interoperability with older versions
839 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
840 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
841 interoperability with such broken implementations. However, enabling
842 this switch breaks interoperability with correct implementations.
844 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
845 re-used X509_PUBKEY object if the second PUBKEY is malformed.
849 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
853 * Change the license to the Apache License v2.0.
857 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
859 - Major releases (indicated by incrementing the MAJOR release number)
860 may introduce incompatible API/ABI changes.
861 - Minor releases (indicated by incrementing the MINOR release number)
862 may introduce new features but retain API/ABI compatibility.
863 - Patch releases (indicated by incrementing the PATCH number)
864 are intended for bug fixes and other improvements of existing
865 features only (like improving performance or adding documentation)
866 and retain API/ABI compatibility.
870 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
874 * Remove the 'dist' target and add a tarball building script. The
875 'dist' target has fallen out of use, and it shouldn't be
876 necessary to configure just to create a source distribution.
880 * Recreate the OS390-Unix config target. It no longer relies on a
881 special script like it did for OpenSSL pre-1.1.0.
885 * Instead of having the source directories listed in Configure, add
886 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
891 * Add GMAC to EVP_MAC.
895 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
899 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
900 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
901 to facilitate the continued use of MACs through raw private keys in
902 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
906 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
907 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
911 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
912 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
913 are retained for backwards compatibility.
917 * AES-XTS mode now enforces that its two keys are different to mitigate
918 the attacked described in "Efficient Instantiations of Tweakable
919 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
920 Details of this attack can be obtained from:
921 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
925 * Rename the object files, i.e. give them other names than in previous
926 versions. Their names now include the name of the final product, as
927 well as its type mnemonic (bin, lib, shlib).
931 * Added new option for 'openssl list', '-objects', which will display the
932 list of built in objects, i.e. OIDs with names.
936 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
937 improves application performance by removing data copies and providing
938 applications with zero-copy system calls such as sendfile and splice.
946 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx] ###
949 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020] ###
951 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
952 while reading in libssl then we would report an error back to the
953 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
954 an error to the stack (which means we instead return SSL_ERROR_SSL) and
955 therefore give a hint as to what went wrong.
959 * Check that ed25519 and ed448 are allowed by the security level. Previously
960 signature algorithms not using an MD were not being checked that they were
961 allowed by the security level.
965 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
966 was not quite right. The behaviour was not consistent between resumption
967 and normal handshakes, and also not quite consistent with historical
968 behaviour. The behaviour in various scenarios has been clarified and
969 it has been updated to make it match historical behaviour as closely as
974 * *[VMS only]* The header files that the VMS compilers include automatically,
975 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
976 that the C++ compiler doesn't understand. This is a shortcoming in the
977 compiler, but can be worked around with `__cplusplus` guards.
979 C++ applications that use OpenSSL libraries must be compiled using the
980 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
981 functions. Otherwise, only functions with symbols of less than 31
982 characters can be used, as the linker will not be able to successfully
983 resolve symbols with longer names.
987 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
988 The presence of this system service is determined at run-time.
992 * Added newline escaping functionality to a filename when using openssl dgst.
993 This output format is to replicate the output format found in the '*sum'
994 checksum programs. This aims to preserve backward compatibility.
996 *Matt Eaton, Richard Levitte, and Paul Dale*
998 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1003 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019] ###
1005 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1006 number generator (RNG). This was intended to include protection in the
1007 event of a fork() system call in order to ensure that the parent and child
1008 processes did not share the same RNG state. However this protection was not
1009 being used in the default case.
1011 A partial mitigation for this issue is that the output from a high
1012 precision timer is mixed into the RNG state so the likelihood of a parent
1013 and child process sharing state is significantly reduced.
1015 If an application already calls OPENSSL_init_crypto() explicitly using
1016 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1019 *Matthias St. Pierre*
1021 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1022 used even when parsing explicit parameters, when loading a serialized key
1023 or calling `EC_GROUP_new_from_ecpkparameters()`/
1024 `EC_GROUP_new_from_ecparameters()`.
1025 This prevents bypass of security hardening and performance gains,
1026 especially for curves with specialized EC_METHODs.
1027 By default, if a key encoded with explicit parameters is loaded and later
1028 serialized, the output is still encoded with explicit parameters, even if
1029 internally a "named" EC_GROUP is used for computation.
1033 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1034 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1035 NULL. After this change, only the cofactor parameter can be NULL. It also
1036 does some minimal sanity checks on the passed order.
1041 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1042 An attack is simple, if the first CMS_recipientInfo is valid but the
1043 second CMS_recipientInfo is chosen ciphertext. If the second
1044 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1045 encryption key will be replaced by garbage, and the message cannot be
1046 decoded, but if the RSA decryption fails, the correct encryption key is
1047 used and the recipient will not notice the attack.
1048 As a work around for this potential attack the length of the decrypted
1049 key must be equal to the cipher default key length, in case the
1050 certifiate is not given and all recipientInfo are tried out.
1051 The old behaviour can be re-enabled in the CMS code by setting the
1052 CMS_DEBUG_DECRYPT flag.
1057 * Early start up entropy quality from the DEVRANDOM seed source has been
1058 improved for older Linux systems. The RAND subsystem will wait for
1059 /dev/random to be producing output before seeding from /dev/urandom.
1060 The seeded state is stored for future library initialisations using
1061 a system global shared memory segment. The shared memory identifier
1062 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1063 the desired value. The default identifier is 114.
1067 * Correct the extended master secret constant on EBCDIC systems. Without this
1068 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1069 negotiate EMS will fail. Unfortunately this also means that TLS connections
1070 between EBCDIC systems with this fix, and EBCDIC systems without this
1071 fix will fail if they negotiate EMS.
1075 * Use Windows installation paths in the mingw builds
1077 Mingw isn't a POSIX environment per se, which means that Windows
1078 paths should be used for installation.
1083 * Changed DH_check to accept parameters with order q and 2q subgroups.
1084 With order 2q subgroups the bit 0 of the private key is not secret
1085 but DH_generate_key works around that by clearing bit 0 of the
1086 private key for those. This avoids leaking bit 0 of the private key.
1090 * Significantly reduce secure memory usage by the randomness pools.
1094 * Revert the DEVRANDOM_WAIT feature for Linux systems
1096 The DEVRANDOM_WAIT feature added a select() call to wait for the
1097 /dev/random device to become readable before reading from the
1098 /dev/urandom device.
1100 It turned out that this change had negative side effects on
1101 performance which were not acceptable. After some discussion it
1102 was decided to revert this feature and leave it up to the OS
1103 resp. the platform maintainer to ensure a proper initialization
1104 during early boot time.
1106 *Matthias St. Pierre*
1108 ### Changes between 1.1.1b and 1.1.1c [28 May 2019] ###
1110 * Add build tests for C++. These are generated files that only do one
1111 thing, to include one public OpenSSL head file each. This tests that
1112 the public header files can be usefully included in a C++ application.
1114 This test isn't enabled by default. It can be enabled with the option
1115 'enable-buildtest-c++'.
1119 * Enable SHA3 pre-hashing for ECDSA and DSA.
1123 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1124 This changes the size when using the genpkey app when no size is given. It
1125 fixes an omission in earlier changes that changed all RSA, DSA and DH
1126 generation apps to use 2048 bits by default.
1130 * Reorganize the manual pages to consistently have RETURN VALUES,
1131 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1132 util/fix-doc-nits accordingly.
1134 *Paul Yang, Joshua Lock*
1136 * Add the missing accessor EVP_PKEY_get0_engine()
1140 * Have apps like 's_client' and 's_server' output the signature scheme
1141 along with other cipher suite parameters when debugging.
1145 * Make OPENSSL_config() error agnostic again.
1149 * Do the error handling in RSA decryption constant time.
1153 * Prevent over long nonces in ChaCha20-Poly1305.
1155 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1156 for every encryption operation. RFC 7539 specifies that the nonce value
1157 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1158 and front pads the nonce with 0 bytes if it is less than 12
1159 bytes. However it also incorrectly allows a nonce to be set of up to 16
1160 bytes. In this case only the last 12 bytes are significant and any
1161 additional leading bytes are ignored.
1163 It is a requirement of using this cipher that nonce values are
1164 unique. Messages encrypted using a reused nonce value are susceptible to
1165 serious confidentiality and integrity attacks. If an application changes
1166 the default nonce length to be longer than 12 bytes and then makes a
1167 change to the leading bytes of the nonce expecting the new value to be a
1168 new unique nonce then such an application could inadvertently encrypt
1169 messages with a reused nonce.
1171 Additionally the ignored bytes in a long nonce are not covered by the
1172 integrity guarantee of this cipher. Any application that relies on the
1173 integrity of these ignored leading bytes of a long nonce may be further
1174 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1175 is safe because no such use sets such a long nonce value. However user
1176 applications that use this cipher directly and set a non-default nonce
1177 length to be longer than 12 bytes may be vulnerable.
1179 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1185 * Add DEVRANDOM_WAIT feature for Linux systems
1187 On older Linux systems where the getrandom() system call is not available,
1188 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1189 Contrary to getrandom(), the /dev/urandom device will not block during
1190 early boot when the kernel CSPRNG has not been seeded yet.
1192 To mitigate this known weakness, use select() to wait for /dev/random to
1193 become readable before reading from /dev/urandom.
1195 * Ensure that SM2 only uses SM3 as digest algorithm
1199 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019] ###
1201 * Change the info callback signals for the start and end of a post-handshake
1202 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1203 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1204 confused by this and assume that a TLSv1.2 renegotiation has started. This
1205 can break KeyUpdate handling. Instead we no longer signal the start and end
1206 of a post handshake message exchange (although the messages themselves are
1207 still signalled). This could break some applications that were expecting
1208 the old signals. However without this KeyUpdate is not usable for many
1213 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018] ###
1215 * Timing vulnerability in DSA signature generation
1217 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1218 timing side channel attack. An attacker could use variations in the signing
1219 algorithm to recover the private key.
1221 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1226 * Timing vulnerability in ECDSA signature generation
1228 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1229 timing side channel attack. An attacker could use variations in the signing
1230 algorithm to recover the private key.
1232 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1237 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1238 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1239 of two gigabytes and the error handling improved.
1241 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1242 categorized as a normal bug, not a security issue, because the DRBG reseeds
1243 automatically and is fully functional even without additional randomness
1244 provided by the application.
1246 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018] ###
1248 * Add a new ClientHello callback. Provides a callback interface that gives
1249 the application the ability to adjust the nascent SSL object at the
1250 earliest stage of ClientHello processing, immediately after extensions have
1251 been collected but before they have been processed. In particular, this
1252 callback can adjust the supported TLS versions in response to the contents
1257 * Add SM2 base algorithm support.
1261 * s390x assembly pack: add (improved) hardware-support for the following
1262 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1263 aes-cfb/cfb8, aes-ecb.
1267 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1268 parameter is no longer accepted, as it leads to a corrupt table. NULL
1269 pem_str is reserved for alias entries only.
1273 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1274 step for prime curves. The new implementation is based on formulae from
1275 differential addition-and-doubling in homogeneous projective coordinates
1276 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1277 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1278 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1279 to work in projective coordinates.
1281 *Billy Bob Brumley, Nicola Tuveri*
1283 * Change generating and checking of primes so that the error rate of not
1284 being prime depends on the intended use based on the size of the input.
1285 For larger primes this will result in more rounds of Miller-Rabin.
1286 The maximal error rate for primes with more than 1080 bits is lowered
1289 *Kurt Roeckx, Annie Yousar*
1291 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1295 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1296 moving between systems, and to avoid confusion when a Windows build is
1297 done with mingw vs with MSVC. For POSIX installs, there's still a
1298 symlink or copy named 'tsget' to avoid that confusion as well.
1302 * Revert blinding in ECDSA sign and instead make problematic addition
1303 length-invariant. Switch even to fixed-length Montgomery multiplication.
1307 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1308 step for binary curves. The new implementation is based on formulae from
1309 differential addition-and-doubling in mixed Lopez-Dahab projective
1310 coordinates, modified to independently blind the operands.
1312 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1314 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1315 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1316 EC_METHODs to implement their own specialized "ladder step", to take
1317 advantage of more favorable coordinate systems or more efficient
1318 differential addition-and-doubling algorithms.
1320 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1322 * Modified the random device based seed sources to keep the relevant
1323 file descriptors open rather than reopening them on each access.
1324 This allows such sources to operate in a chroot() jail without
1325 the associated device nodes being available. This behaviour can be
1326 controlled using RAND_keep_random_devices_open().
1330 * Numerous side-channel attack mitigations have been applied. This may have
1331 performance impacts for some algorithms for the benefit of improved
1332 security. Specific changes are noted in this change log by their respective
1337 * AIX shared library support overhaul. Switch to AIX "natural" way of
1338 handling shared libraries, which means collecting shared objects of
1339 different versions and bitnesses in one common archive. This allows to
1340 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1341 doesn't affect the way 3rd party applications are linked, only how
1342 multi-version installation is managed.
1346 * Make ec_group_do_inverse_ord() more robust and available to other
1347 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1348 mitigations are applied to the fallback BN_mod_inverse().
1349 When using this function rather than BN_mod_inverse() directly, new
1350 EC cryptosystem implementations are then safer-by-default.
1354 * Add coordinate blinding for EC_POINT and implement projective
1355 coordinate blinding for generic prime curves as a countermeasure to
1356 chosen point SCA attacks.
1358 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1360 * Add blinding to ECDSA and DSA signatures to protect against side channel
1361 attacks discovered by Keegan Ryan (NCC Group).
1365 * Enforce checking in the pkeyutl command line app to ensure that the input
1366 length does not exceed the maximum supported digest length when performing
1367 a sign, verify or verifyrecover operation.
1371 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1372 I/O in combination with something like select() or poll() will hang. This
1373 can be turned off again using SSL_CTX_clear_mode().
1374 Many applications do not properly handle non-application data records, and
1375 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1376 around the problems in those applications, but can also break some.
1377 It's recommended to read the manpages about SSL_read(), SSL_write(),
1378 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1379 SSL_CTX_set_read_ahead() again.
1383 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1384 now allow empty (zero character) pass phrases.
1388 * Apply blinding to binary field modular inversion and remove patent
1389 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1393 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1394 binary and prime elliptic curves.
1398 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1399 constant time fixed point multiplication.
1403 * Revise elliptic curve scalar multiplication with timing attack
1404 defenses: ec_wNAF_mul redirects to a constant time implementation
1405 when computing fixed point and variable point multiplication (which
1406 in OpenSSL are mostly used with secret scalars in keygen, sign,
1407 ECDH derive operations).
1408 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1411 * Updated CONTRIBUTING
1415 * Updated DRBG / RAND to request nonce and additional low entropy
1416 randomness from the system.
1418 *Matthias St. Pierre*
1420 * Updated 'openssl rehash' to use OpenSSL consistent default.
1424 * Moved the load of the ssl_conf module to libcrypto, which helps
1425 loading engines that libssl uses before libssl is initialised.
1429 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1433 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1435 *Ingo Schwarze, Rich Salz*
1437 * Added output of accepting IP address and port for 'openssl s_server'
1441 * Added a new API for TLSv1.3 ciphersuites:
1442 SSL_CTX_set_ciphersuites()
1443 SSL_set_ciphersuites()
1447 * Memory allocation failures consistently add an error to the error
1452 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1453 in libcrypto when run as setuid/setgid.
1457 * Load any config file by default when libssl is used.
1461 * Added new public header file <openssl/rand_drbg.h> and documentation
1462 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1464 *Matthias St. Pierre*
1466 * QNX support removed (cannot find contributors to get their approval
1467 for the license change).
1471 * TLSv1.3 replay protection for early data has been implemented. See the
1472 SSL_read_early_data() man page for further details.
1476 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1477 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1478 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1479 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1480 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1481 configuration has been separated out. See the ciphers man page or the
1482 SSL_CTX_set_ciphersuites() man page for more information.
1486 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1487 in responder mode now supports the new "-multi" option, which
1488 spawns the specified number of child processes to handle OCSP
1489 requests. The "-timeout" option now also limits the OCSP
1490 responder's patience to wait to receive the full client request
1491 on a newly accepted connection. Child processes are respawned
1492 as needed, and the CA index file is automatically reloaded
1493 when changed. This makes it possible to run the "ocsp" responder
1494 as a long-running service, making the OpenSSL CA somewhat more
1495 feature-complete. In this mode, most diagnostic messages logged
1496 after entering the event loop are logged via syslog(3) rather than
1501 * Added support for X448 and Ed448. Heavily based on original work by
1506 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1507 objects loaded. This adds the functions OSSL_STORE_expect() and
1508 OSSL_STORE_find() as well as needed tools to construct searches and
1509 get the search data out of them.
1513 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1514 version of OpenSSL should review their configuration settings to ensure
1515 that they are still appropriate for TLSv1.3. For further information see:
1516 https://wiki.openssl.org/index.php/TLS1.3
1520 * Grand redesign of the OpenSSL random generator
1522 The default RAND method now utilizes an AES-CTR DRBG according to
1523 NIST standard SP 800-90Ar1. The new random generator is essentially
1524 a port of the default random generator from the OpenSSL FIPS 2.0
1525 object module. It is a hybrid deterministic random bit generator
1526 using an AES-CTR bit stream and which seeds and reseeds itself
1527 automatically using trusted system entropy sources.
1529 Some of its new features are:
1530 - Support for multiple DRBG instances with seed chaining.
1531 - The default RAND method makes use of a DRBG.
1532 - There is a public and private DRBG instance.
1533 - The DRBG instances are fork-safe.
1534 - Keep all global DRBG instances on the secure heap if it is enabled.
1535 - The public and private DRBG instance are per thread for lock free
1538 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1540 * Changed Configure so it only says what it does and doesn't dump
1541 so much data. Instead, ./configdata.pm should be used as a script
1542 to display all sorts of configuration data.
1546 * Added processing of "make variables" to Configure.
1550 * Added SHA512/224 and SHA512/256 algorithm support.
1554 * The last traces of Netware support, first removed in 1.1.0, have
1559 * Get rid of Makefile.shared, and in the process, make the processing
1560 of certain files (rc.obj, or the .def/.map/.opt files produced from
1561 the ordinal files) more visible and hopefully easier to trace and
1562 debug (or make silent).
1566 * Make it possible to have environment variable assignments as
1567 arguments to config / Configure.
1571 * Add multi-prime RSA (RFC 8017) support.
1575 * Add SM3 implemented according to GB/T 32905-2016
1576 * Jack Lloyd <jack.lloyd@ribose.com>,
1577 Ronald Tse <ronald.tse@ribose.com>,
1578 Erick Borsboom <erick.borsboom@ribose.com> *
1580 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1581 as documented in RFC6066.
1582 Based on a patch from Tomasz Moń
1584 *Filipe Raimundo da Silva*
1586 * Add SM4 implemented according to GB/T 32907-2016.
1587 * Jack Lloyd <jack.lloyd@ribose.com>,
1588 Ronald Tse <ronald.tse@ribose.com>,
1589 Erick Borsboom <erick.borsboom@ribose.com> *
1591 * Reimplement -newreq-nodes and ERR_error_string_n; the
1592 original author does not agree with the license change.
1596 * Add ARIA AEAD TLS support.
1600 * Some macro definitions to support VS6 have been removed. Visual
1601 Studio 6 has not worked since 1.1.0
1605 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1606 without clearing the errors.
1610 * Add "atfork" functions. If building on a system that without
1611 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1612 requirements. The RAND facility now uses/requires this.
1620 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1621 not possible to disable entirely. However, it's still possible to
1622 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1625 To disable, configure with 'no-ui-console'. 'no-ui' is still
1626 possible to use as an alias. Check at compile time with the
1627 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1628 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1632 * Add a STORE module, which implements a uniform and URI based reader of
1633 stores that can contain keys, certificates, CRLs and numerous other
1634 objects. The main API is loosely based on a few stdio functions,
1635 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1636 OSSL_STORE_error and OSSL_STORE_close.
1637 The implementation uses backends called "loaders" to implement arbitrary
1638 URI schemes. There is one built in "loader" for the 'file' scheme.
1642 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1643 then adjusted to work on FreeBSD 8.4 as well.
1644 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1645 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1649 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1650 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1651 error code calls like this:
1653 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1655 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1656 that can be encoded in C. For the foreseeable future, this will only
1659 *Richard Levitte and Tim Hudson*
1661 * Removed BSD cryptodev engine.
1665 * Add a build target 'build_all_generated', to build all generated files
1666 and only that. This can be used to prepare everything that requires
1667 things like perl for a system that lacks perl and then move everything
1668 to that system and do the rest of the build there.
1672 * In the UI interface, make it possible to duplicate the user data. This
1673 can be used by engines that need to retain the data for a longer time
1674 than just the call where this user data is passed.
1678 * Ignore the '-named_curve auto' value for compatibility of applications
1681 *Tomas Mraz <tmraz@fedoraproject.org>*
1683 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1684 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1685 alerts across multiple records (some of which could be empty). In practice
1686 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1687 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1688 support this at all. Supporting it adds significant complexity to the
1689 record layer, and its removal is unlikely to cause interoperability
1694 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1695 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1696 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1701 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1702 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1704 *Richard Levitte, Andy Polyakov*
1706 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1711 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1712 platform rather than 'mingw'.
1716 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1717 success if they are asked to add an object which already exists
1718 in the store. This change cascades to other functions which load
1719 certificates and CRLs.
1723 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1724 facilitate stack unwinding even from assembly subroutines.
1728 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1729 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1733 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1734 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1735 which is the minimum version we support.
1739 * Certificate time validation (X509_cmp_time) enforces stricter
1740 compliance with RFC 5280. Fractional seconds and timezone offsets
1741 are no longer allowed.
1745 * Add support for ARIA
1749 * s_client will now send the Server Name Indication (SNI) extension by
1750 default unless the new "-noservername" option is used. The server name is
1751 based on the host provided to the "-connect" option unless overridden by
1752 using "-servername".
1756 * Add support for SipHash
1760 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1761 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1762 prevent issues where no progress is being made and the peer continually
1763 sends unrecognised record types, using up resources processing them.
1767 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1768 using the algorithm defined in
1769 https://www.akkadia.org/drepper/SHA-crypt.txt
1773 * Heartbeat support has been removed; the ABI is changed for now.
1775 *Richard Levitte, Rich Salz*
1777 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1781 * The RSA "null" method, which was partially supported to avoid patent
1782 issues, has been replaced to always returns NULL.
1790 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019] ###
1792 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1793 used even when parsing explicit parameters, when loading a serialized key
1794 or calling `EC_GROUP_new_from_ecpkparameters()`/
1795 `EC_GROUP_new_from_ecparameters()`.
1796 This prevents bypass of security hardening and performance gains,
1797 especially for curves with specialized EC_METHODs.
1798 By default, if a key encoded with explicit parameters is loaded and later
1799 serialized, the output is still encoded with explicit parameters, even if
1800 internally a "named" EC_GROUP is used for computation.
1804 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1805 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1806 NULL. After this change, only the cofactor parameter can be NULL. It also
1807 does some minimal sanity checks on the passed order.
1812 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1813 An attack is simple, if the first CMS_recipientInfo is valid but the
1814 second CMS_recipientInfo is chosen ciphertext. If the second
1815 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1816 encryption key will be replaced by garbage, and the message cannot be
1817 decoded, but if the RSA decryption fails, the correct encryption key is
1818 used and the recipient will not notice the attack.
1819 As a work around for this potential attack the length of the decrypted
1820 key must be equal to the cipher default key length, in case the
1821 certifiate is not given and all recipientInfo are tried out.
1822 The old behaviour can be re-enabled in the CMS code by setting the
1823 CMS_DEBUG_DECRYPT flag.
1828 * Use Windows installation paths in the mingw builds
1830 Mingw isn't a POSIX environment per se, which means that Windows
1831 paths should be used for installation.
1836 ### Changes between 1.1.0j and 1.1.0k [28 May 2019] ###
1838 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1839 This changes the size when using the genpkey app when no size is given. It
1840 fixes an omission in earlier changes that changed all RSA, DSA and DH
1841 generation apps to use 2048 bits by default.
1845 * Prevent over long nonces in ChaCha20-Poly1305.
1847 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1848 for every encryption operation. RFC 7539 specifies that the nonce value
1849 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1850 and front pads the nonce with 0 bytes if it is less than 12
1851 bytes. However it also incorrectly allows a nonce to be set of up to 16
1852 bytes. In this case only the last 12 bytes are significant and any
1853 additional leading bytes are ignored.
1855 It is a requirement of using this cipher that nonce values are
1856 unique. Messages encrypted using a reused nonce value are susceptible to
1857 serious confidentiality and integrity attacks. If an application changes
1858 the default nonce length to be longer than 12 bytes and then makes a
1859 change to the leading bytes of the nonce expecting the new value to be a
1860 new unique nonce then such an application could inadvertently encrypt
1861 messages with a reused nonce.
1863 Additionally the ignored bytes in a long nonce are not covered by the
1864 integrity guarantee of this cipher. Any application that relies on the
1865 integrity of these ignored leading bytes of a long nonce may be further
1866 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1867 is safe because no such use sets such a long nonce value. However user
1868 applications that use this cipher directly and set a non-default nonce
1869 length to be longer than 12 bytes may be vulnerable.
1871 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1877 * Added SCA hardening for modular field inversion in EC_GROUP through
1878 a new dedicated field_inv() pointer in EC_METHOD.
1879 This also addresses a leakage affecting conversions from projective
1880 to affine coordinates.
1882 *Billy Bob Brumley, Nicola Tuveri*
1884 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1885 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1889 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1893 * Remove the 'dist' target and add a tarball building script. The
1894 'dist' target has fallen out of use, and it shouldn't be
1895 necessary to configure just to create a source distribution.
1899 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018] ###
1901 * Timing vulnerability in DSA signature generation
1903 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1904 timing side channel attack. An attacker could use variations in the signing
1905 algorithm to recover the private key.
1907 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1912 * Timing vulnerability in ECDSA signature generation
1914 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1915 timing side channel attack. An attacker could use variations in the signing
1916 algorithm to recover the private key.
1918 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1923 * Add coordinate blinding for EC_POINT and implement projective
1924 coordinate blinding for generic prime curves as a countermeasure to
1925 chosen point SCA attacks.
1927 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1929 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018] ###
1931 * Client DoS due to large DH parameter
1933 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1934 malicious server can send a very large prime value to the client. This will
1935 cause the client to spend an unreasonably long period of time generating a
1936 key for this prime resulting in a hang until the client has finished. This
1937 could be exploited in a Denial Of Service attack.
1939 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1944 * Cache timing vulnerability in RSA Key Generation
1946 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1947 a cache timing side channel attack. An attacker with sufficient access to
1948 mount cache timing attacks during the RSA key generation process could
1949 recover the private key.
1951 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1952 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1957 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1958 parameter is no longer accepted, as it leads to a corrupt table. NULL
1959 pem_str is reserved for alias entries only.
1963 * Revert blinding in ECDSA sign and instead make problematic addition
1964 length-invariant. Switch even to fixed-length Montgomery multiplication.
1968 * Change generating and checking of primes so that the error rate of not
1969 being prime depends on the intended use based on the size of the input.
1970 For larger primes this will result in more rounds of Miller-Rabin.
1971 The maximal error rate for primes with more than 1080 bits is lowered
1974 *Kurt Roeckx, Annie Yousar*
1976 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1980 * Add blinding to ECDSA and DSA signatures to protect against side channel
1981 attacks discovered by Keegan Ryan (NCC Group).
1985 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1986 now allow empty (zero character) pass phrases.
1990 * Certificate time validation (X509_cmp_time) enforces stricter
1991 compliance with RFC 5280. Fractional seconds and timezone offsets
1992 are no longer allowed.
1996 * Fixed a text canonicalisation bug in CMS
1998 Where a CMS detached signature is used with text content the text goes
1999 through a canonicalisation process first prior to signing or verifying a
2000 signature. This process strips trailing space at the end of lines, converts
2001 line terminators to CRLF and removes additional trailing line terminators
2002 at the end of a file. A bug in the canonicalisation process meant that
2003 some characters, such as form-feed, were incorrectly treated as whitespace
2004 and removed. This is contrary to the specification (RFC5485). This fix
2005 could mean that detached text data signed with an earlier version of
2006 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2007 signed with a fixed OpenSSL may fail to verify with an earlier version of
2008 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2009 and use the "-binary" flag (for the "cms" command line application) or set
2010 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2014 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018] ###
2016 * Constructed ASN.1 types with a recursive definition could exceed the stack
2018 Constructed ASN.1 types with a recursive definition (such as can be found
2019 in PKCS7) could eventually exceed the stack given malicious input with
2020 excessive recursion. This could result in a Denial Of Service attack. There
2021 are no such structures used within SSL/TLS that come from untrusted sources
2022 so this is considered safe.
2024 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2030 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2032 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2033 effectively reduced to only comparing the least significant bit of each
2034 byte. This allows an attacker to forge messages that would be considered as
2035 authenticated in an amount of tries lower than that guaranteed by the
2036 security claims of the scheme. The module can only be compiled by the
2037 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2039 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2045 * Add a build target 'build_all_generated', to build all generated files
2046 and only that. This can be used to prepare everything that requires
2047 things like perl for a system that lacks perl and then move everything
2048 to that system and do the rest of the build there.
2052 * Backport SSL_OP_NO_RENGOTIATION
2054 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2055 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2056 changes this is no longer possible in 1.1.0. Therefore the new
2057 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2058 1.1.0 to provide equivalent functionality.
2060 Note that if an application built against 1.1.0h headers (or above) is run
2061 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2062 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2066 * Removed the OS390-Unix config target. It relied on a script that doesn't
2071 * rsaz_1024_mul_avx2 overflow bug on x86_64
2073 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2074 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2075 Analysis suggests that attacks against RSA and DSA as a result of this
2076 defect would be very difficult to perform and are not believed likely.
2077 Attacks against DH1024 are considered just feasible, because most of the
2078 work necessary to deduce information about a private key may be performed
2079 offline. The amount of resources required for such an attack would be
2080 significant. However, for an attack on TLS to be meaningful, the server
2081 would have to share the DH1024 private key among multiple clients, which is
2082 no longer an option since CVE-2016-0701.
2084 This only affects processors that support the AVX2 but not ADX extensions
2085 like Intel Haswell (4th generation).
2087 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2088 was originally found via the OSS-Fuzz project.
2093 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017] ###
2095 * bn_sqrx8x_internal carry bug on x86_64
2097 There is a carry propagating bug in the x86_64 Montgomery squaring
2098 procedure. No EC algorithms are affected. Analysis suggests that attacks
2099 against RSA and DSA as a result of this defect would be very difficult to
2100 perform and are not believed likely. Attacks against DH are considered just
2101 feasible (although very difficult) because most of the work necessary to
2102 deduce information about a private key may be performed offline. The amount
2103 of resources required for such an attack would be very significant and
2104 likely only accessible to a limited number of attackers. An attacker would
2105 additionally need online access to an unpatched system using the target
2106 private key in a scenario with persistent DH parameters and a private
2107 key that is shared between multiple clients.
2109 This only affects processors that support the BMI1, BMI2 and ADX extensions
2110 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2112 This issue was reported to OpenSSL by the OSS-Fuzz project.
2117 * Malformed X.509 IPAddressFamily could cause OOB read
2119 If an X.509 certificate has a malformed IPAddressFamily extension,
2120 OpenSSL could do a one-byte buffer overread. The most likely result
2121 would be an erroneous display of the certificate in text format.
2123 This issue was reported to OpenSSL by the OSS-Fuzz project.
2128 ### Changes between 1.1.0e and 1.1.0f [25 May 2017] ###
2130 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2131 platform rather than 'mingw'.
2135 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2136 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2137 which is the minimum version we support.
2141 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017] ###
2143 * Encrypt-Then-Mac renegotiation crash
2145 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2146 negotiated where it was not in the original handshake (or vice-versa) then
2147 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2148 and servers are affected.
2150 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2155 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017] ###
2157 * Truncated packet could crash via OOB read
2159 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2160 cipher is being used, then a truncated packet can cause that host to
2161 perform an out-of-bounds read, usually resulting in a crash.
2163 This issue was reported to OpenSSL by Robert Święcki of Google.
2168 * Bad (EC)DHE parameters cause a client crash
2170 If a malicious server supplies bad parameters for a DHE or ECDHE key
2171 exchange then this can result in the client attempting to dereference a
2172 NULL pointer leading to a client crash. This could be exploited in a Denial
2175 This issue was reported to OpenSSL by Guido Vranken.
2180 * BN_mod_exp may produce incorrect results on x86_64
2182 There is a carry propagating bug in the x86_64 Montgomery squaring
2183 procedure. No EC algorithms are affected. Analysis suggests that attacks
2184 against RSA and DSA as a result of this defect would be very difficult to
2185 perform and are not believed likely. Attacks against DH are considered just
2186 feasible (although very difficult) because most of the work necessary to
2187 deduce information about a private key may be performed offline. The amount
2188 of resources required for such an attack would be very significant and
2189 likely only accessible to a limited number of attackers. An attacker would
2190 additionally need online access to an unpatched system using the target
2191 private key in a scenario with persistent DH parameters and a private
2192 key that is shared between multiple clients. For example this can occur by
2193 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2194 similar to CVE-2015-3193 but must be treated as a separate problem.
2196 This issue was reported to OpenSSL by the OSS-Fuzz project.
2201 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016] ###
2203 * ChaCha20/Poly1305 heap-buffer-overflow
2205 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
2206 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2207 crash. This issue is not considered to be exploitable beyond a DoS.
2209 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2214 * CMS Null dereference
2216 Applications parsing invalid CMS structures can crash with a NULL pointer
2217 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2218 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2219 structure callback if an attempt is made to free certain invalid encodings.
2220 Only CHOICE structures using a callback which do not handle NULL value are
2223 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2228 * Montgomery multiplication may produce incorrect results
2230 There is a carry propagating bug in the Broadwell-specific Montgomery
2231 multiplication procedure that handles input lengths divisible by, but
2232 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2233 and DH private keys are impossible. This is because the subroutine in
2234 question is not used in operations with the private key itself and an input
2235 of the attacker's direct choice. Otherwise the bug can manifest itself as
2236 transient authentication and key negotiation failures or reproducible
2237 erroneous outcome of public-key operations with specially crafted input.
2238 Among EC algorithms only Brainpool P-512 curves are affected and one
2239 presumably can attack ECDH key negotiation. Impact was not analyzed in
2240 detail, because pre-requisites for attack are considered unlikely. Namely
2241 multiple clients have to choose the curve in question and the server has to
2242 share the private key among them, neither of which is default behaviour.
2243 Even then only clients that chose the curve will be affected.
2245 This issue was publicly reported as transient failures and was not
2246 initially recognized as a security issue. Thanks to Richard Morgan for
2247 providing reproducible case.
2252 * Removed automatic addition of RPATH in shared libraries and executables,
2253 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2257 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016] ###
2259 * Fix Use After Free for large message sizes
2261 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2262 message larger than approx 16k is received then the underlying buffer to
2263 store the incoming message is reallocated and moved. Unfortunately a
2264 dangling pointer to the old location is left which results in an attempt to
2265 write to the previously freed location. This is likely to result in a
2266 crash, however it could potentially lead to execution of arbitrary code.
2268 This issue only affects OpenSSL 1.1.0a.
2270 This issue was reported to OpenSSL by Robert Święcki.
2275 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016] ###
2277 * OCSP Status Request extension unbounded memory growth
2279 A malicious client can send an excessively large OCSP Status Request
2280 extension. If that client continually requests renegotiation, sending a
2281 large OCSP Status Request extension each time, then there will be unbounded
2282 memory growth on the server. This will eventually lead to a Denial Of
2283 Service attack through memory exhaustion. Servers with a default
2284 configuration are vulnerable even if they do not support OCSP. Builds using
2285 the "no-ocsp" build time option are not affected.
2287 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2292 * SSL_peek() hang on empty record
2294 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2295 sends an empty record. This could be exploited by a malicious peer in a
2296 Denial Of Service attack.
2298 This issue was reported to OpenSSL by Alex Gaynor.
2303 * Excessive allocation of memory in tls_get_message_header() and
2304 dtls1_preprocess_fragment()
2306 A (D)TLS message includes 3 bytes for its length in the header for the
2307 message. This would allow for messages up to 16Mb in length. Messages of
2308 this length are excessive and OpenSSL includes a check to ensure that a
2309 peer is sending reasonably sized messages in order to avoid too much memory
2310 being consumed to service a connection. A flaw in the logic of version
2311 1.1.0 means that memory for the message is allocated too early, prior to
2312 the excessive message length check. Due to way memory is allocated in
2313 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2314 to service a connection. This could lead to a Denial of Service through
2315 memory exhaustion. However, the excessive message length check still takes
2316 place, and this would cause the connection to immediately fail. Assuming
2317 that the application calls SSL_free() on the failed connection in a timely
2318 manner then the 21Mb of allocated memory will then be immediately freed
2319 again. Therefore the excessive memory allocation will be transitory in
2320 nature. This then means that there is only a security impact if:
2322 1) The application does not call SSL_free() in a timely manner in the event
2323 that the connection fails
2325 2) The application is working in a constrained environment where there is
2326 very little free memory
2328 3) The attacker initiates multiple connection attempts such that there are
2329 multiple connections in a state where memory has been allocated for the
2330 connection; SSL_free() has not yet been called; and there is insufficient
2331 memory to service the multiple requests.
2333 Except in the instance of (1) above any Denial Of Service is likely to be
2334 transitory because as soon as the connection fails the memory is
2335 subsequently freed again in the SSL_free() call. However there is an
2336 increased risk during this period of application crashes due to the lack of
2337 memory - which would then mean a more serious Denial of Service.
2339 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2340 (CVE-2016-6307 and CVE-2016-6308)
2344 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2345 had to be removed. Primary reason is that vendor assembler can't
2346 assemble our modules with -KPIC flag. As result it, assembly
2347 support, was not even available as option. But its lack means
2348 lack of side-channel resistant code, which is incompatible with
2349 security by todays standards. Fortunately gcc is readily available
2350 prepackaged option, which we firmly point at...
2354 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016] ###
2356 * Windows command-line tool supports UTF-8 opt-in option for arguments
2357 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2358 (to any value) allows Windows user to access PKCS#12 file generated
2359 with Windows CryptoAPI and protected with non-ASCII password, as well
2360 as files generated under UTF-8 locale on Linux also protected with
2365 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2366 have been disabled by default and removed from DEFAULT, just like RC4.
2367 See the RC4 item below to re-enable both.
2371 * The method for finding the storage location for the Windows RAND seed file
2372 has changed. First we check %RANDFILE%. If that is not set then we check
2373 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2374 all else fails we fall back to C:\.
2378 * The EVP_EncryptUpdate() function has had its return type changed from void
2379 to int. A return of 0 indicates and error while a return of 1 indicates
2384 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2385 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2386 off the constant time implementation for RSA, DSA and DH have been made
2387 no-ops and deprecated.
2391 * Windows RAND implementation was simplified to only get entropy by
2392 calling CryptGenRandom(). Various other RAND-related tickets
2395 *Joseph Wylie Yandle, Rich Salz*
2397 * The stack and lhash API's were renamed to start with OPENSSL_SK_
2398 and OPENSSL_LH_, respectively. The old names are available
2399 with API compatibility. They new names are now completely documented.
2403 * Unify TYPE_up_ref(obj) methods signature.
2404 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2405 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2406 int (instead of void) like all others TYPE_up_ref() methods.
2407 So now these methods also check the return value of CRYPTO_atomic_add(),
2408 and the validity of object reference counter.
2410 *fdasilvayy@gmail.com*
2412 * With Windows Visual Studio builds, the .pdb files are installed
2413 alongside the installed libraries and executables. For a static
2414 library installation, ossl_static.pdb is the associate compiler
2415 generated .pdb file to be used when linking programs.
2419 * Remove openssl.spec. Packaging files belong with the packagers.
2423 * Automatic Darwin/OSX configuration has had a refresh, it will now
2424 recognise x86_64 architectures automatically. You can still decide
2425 to build for a different bitness with the environment variable
2426 KERNEL_BITS (can be 32 or 64), for example:
2428 KERNEL_BITS=32 ./config
2432 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2433 256 bit AES and HMAC with SHA256.
2437 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2441 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2445 * To enable users to have their own config files and build file templates,
2446 Configure looks in the directory indicated by the environment variable
2447 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2448 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2449 name and is used as is.
2453 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2454 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2455 X509_CERT_FILE_CTX was removed.
2459 * "shared" builds are now the default. To create only static libraries use
2460 the "no-shared" Configure option.
2464 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2465 All of these option have not worked for some while and are fundamental
2470 * Make various cleanup routines no-ops and mark them as deprecated. Most
2471 global cleanup functions are no longer required because they are handled
2472 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2473 Explicitly de-initing can cause problems (e.g. where a library that uses
2474 OpenSSL de-inits, but an application is still using it). The affected
2475 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2476 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2477 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2478 COMP_zlib_cleanup().
2482 * --strict-warnings no longer enables runtime debugging options
2483 such as REF_DEBUG. Instead, debug options are automatically
2484 enabled with '--debug' builds.
2486 *Andy Polyakov, Emilia Käsper*
2488 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2489 have been moved out of the public header files. New functions for managing
2490 these have been added.
2494 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2495 objects have been moved out of the public header files. New
2496 functions for managing these have been added.
2500 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2501 have been moved out of the public header files. New functions for managing
2502 these have been added.
2506 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2507 moved out of the public header files. New functions for managing these
2512 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2516 * Removed the mk1mf build scripts.
2520 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2521 it is always safe to #include a header now.
2525 * Removed the aged BC-32 config and all its supporting scripts
2529 * Removed support for Ultrix, Netware, and OS/2.
2533 * Add support for HKDF.
2535 *Alessandro Ghedini*
2537 * Add support for blake2b and blake2s
2541 * Added support for "pipelining". Ciphers that have the
2542 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2543 encryptions/decryptions simultaneously. There are currently no built-in
2544 ciphers with this property but the expectation is that engines will be able
2545 to offer it to significantly improve throughput. Support has been extended
2546 into libssl so that multiple records for a single connection can be
2547 processed in one go (for >=TLS 1.1).
2551 * Added the AFALG engine. This is an async capable engine which is able to
2552 offload work to the Linux kernel. In this initial version it only supports
2553 AES128-CBC. The kernel must be version 4.1.0 or greater.
2557 * OpenSSL now uses a new threading API. It is no longer necessary to
2558 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2559 are two supported threading models: pthreads and windows threads. It is
2560 also possible to configure OpenSSL at compile time for "no-threads". The
2561 old threading API should no longer be used. The functions have been
2562 replaced with "no-op" compatibility macros.
2564 *Alessandro Ghedini, Matt Caswell*
2566 * Modify behavior of ALPN to invoke callback after SNI/servername
2567 callback, such that updates to the SSL_CTX affect ALPN.
2571 * Add SSL_CIPHER queries for authentication and key-exchange.
2575 * Changes to the DEFAULT cipherlist:
2576 - Prefer (EC)DHE handshakes over plain RSA.
2577 - Prefer AEAD ciphers over legacy ciphers.
2578 - Prefer ECDSA over RSA when both certificates are available.
2579 - Prefer TLSv1.2 ciphers/PRF.
2580 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2585 * Change the ECC default curve list to be this, in order: x25519,
2586 secp256r1, secp521r1, secp384r1.
2590 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2591 disabled by default. They can be re-enabled using the
2592 enable-weak-ssl-ciphers option to Configure.
2596 * If the server has ALPN configured, but supports no protocols that the
2597 client advertises, send a fatal "no_application_protocol" alert.
2598 This behaviour is SHALL in RFC 7301, though it isn't universally
2599 implemented by other servers.
2603 * Add X25519 support.
2604 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2605 for public and private key encoding using the format documented in
2606 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2607 key generation and key derivation.
2609 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2614 * Deprecate SRP_VBASE_get_by_user.
2615 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2616 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2617 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2618 seed, even if the seed is configured.
2620 Users should use SRP_VBASE_get1_by_user instead. Note that in
2621 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2622 also that even though configuring the SRP seed attempts to hide
2623 invalid usernames by continuing the handshake with fake
2624 credentials, this behaviour is not constant time and no strong
2625 guarantees are made that the handshake is indistinguishable from
2626 that of a valid user.
2630 * Configuration change; it's now possible to build dynamic engines
2631 without having to build shared libraries and vice versa. This
2632 only applies to the engines in engines/, those in crypto/engine/
2633 will always be built into libcrypto (i.e. "static").
2635 Building dynamic engines is enabled by default; to disable, use
2636 the configuration option "disable-dynamic-engine".
2638 The only requirements for building dynamic engines are the
2639 presence of the DSO module and building with position independent
2640 code, so they will also automatically be disabled if configuring
2641 with "disable-dso" or "disable-pic".
2643 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2644 are also taken away from openssl/opensslconf.h, as they are
2649 * Configuration change; if there is a known flag to compile
2650 position independent code, it will always be applied on the
2651 libcrypto and libssl object files, and never on the application
2652 object files. This means other libraries that use routines from
2653 libcrypto / libssl can be made into shared libraries regardless
2654 of how OpenSSL was configured.
2656 If this isn't desirable, the configuration options "disable-pic"
2657 or "no-pic" can be used to disable the use of PIC. This will
2658 also disable building shared libraries and dynamic engines.
2662 * Removed JPAKE code. It was experimental and has no wide use.
2666 * The INSTALL_PREFIX Makefile variable has been renamed to
2667 DESTDIR. That makes for less confusion on what this variable
2668 is for. Also, the configuration option --install_prefix is
2673 * Heartbeat for TLS has been removed and is disabled by default
2674 for DTLS; configure with enable-heartbeats. Code that uses the
2675 old #define's might need to be updated.
2677 *Emilia Käsper, Rich Salz*
2679 * Rename REF_CHECK to REF_DEBUG.
2683 * New "unified" build system
2685 The "unified" build system is aimed to be a common system for all
2686 platforms we support. With it comes new support for VMS.
2688 This system builds supports building in a different directory tree
2689 than the source tree. It produces one Makefile (for unix family
2690 or lookalikes), or one descrip.mms (for VMS).
2692 The source of information to make the Makefile / descrip.mms is
2693 small files called 'build.info', holding the necessary
2694 information for each directory with source to compile, and a
2695 template in Configurations, like unix-Makefile.tmpl or
2698 With this change, the library names were also renamed on Windows
2699 and on VMS. They now have names that are closer to the standard
2700 on Unix, and include the major version number, and in certain
2701 cases, the architecture they are built for. See "Notes on shared
2702 libraries" in INSTALL.
2704 We rely heavily on the perl module Text::Template.
2708 * Added support for auto-initialisation and de-initialisation of the library.
2709 OpenSSL no longer requires explicit init or deinit routines to be called,
2710 except in certain circumstances. See the OPENSSL_init_crypto() and
2711 OPENSSL_init_ssl() man pages for further information.
2715 * The arguments to the DTLSv1_listen function have changed. Specifically the
2716 "peer" argument is now expected to be a BIO_ADDR object.
2718 * Rewrite of BIO networking library. The BIO library lacked consistent
2719 support of IPv6, and adding it required some more extensive
2720 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2721 which hold all types of addresses and chains of address information.
2722 It also introduces a new API, with functions like BIO_socket,
2723 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2724 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2725 have been adapted accordingly.
2729 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2734 * CRIME protection: disable compression by default, even if OpenSSL is
2735 compiled with zlib enabled. Applications can still enable compression
2736 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2737 using the SSL_CONF library to configure compression.
2741 * The signature of the session callback configured with
2742 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2743 was explicitly marked as 'const unsigned char*' instead of
2748 * Always DPURIFY. Remove the use of uninitialized memory in the
2749 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2753 * Removed many obsolete configuration items, including
2754 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2755 MD2_CHAR, MD2_INT, MD2_LONG
2757 IDEA_SHORT, IDEA_LONG
2758 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2760 *Rich Salz, with advice from Andy Polyakov*
2762 * Many BN internals have been moved to an internal header file.
2764 *Rich Salz with help from Andy Polyakov*
2766 * Configuration and writing out the results from it has changed.
2767 Files such as Makefile include/openssl/opensslconf.h and are now
2768 produced through general templates, such as Makefile.in and
2769 crypto/opensslconf.h.in and some help from the perl module
2772 Also, the center of configuration information is no longer
2773 Makefile. Instead, Configure produces a perl module in
2774 configdata.pm which holds most of the config data (in the hash
2775 table %config), the target data that comes from the target
2776 configuration in one of the Configurations/*.conf files (in
2781 * To clarify their intended purposes, the Configure options
2782 --prefix and --openssldir change their semantics, and become more
2783 straightforward and less interdependent.
2785 --prefix shall be used exclusively to give the location INSTALLTOP
2786 where programs, scripts, libraries, include files and manuals are
2787 going to be installed. The default is now /usr/local.
2789 --openssldir shall be used exclusively to give the default
2790 location OPENSSLDIR where certificates, private keys, CRLs are
2791 managed. This is also where the default openssl.cnf gets
2793 If the directory given with this option is a relative path, the
2794 values of both the --prefix value and the --openssldir value will
2795 be combined to become OPENSSLDIR.
2796 The default for --openssldir is INSTALLTOP/ssl.
2798 Anyone who uses --openssldir to specify where OpenSSL is to be
2799 installed MUST change to use --prefix instead.
2803 * The GOST engine was out of date and therefore it has been removed. An up
2804 to date GOST engine is now being maintained in an external repository.
2805 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2806 support for GOST ciphersuites (these are only activated if a GOST engine
2811 * EGD is no longer supported by default; use enable-egd when
2814 *Ben Kaduk and Rich Salz*
2816 * The distribution now has Makefile.in files, which are used to
2817 create Makefile's when Configure is run. *Configure must be run
2818 before trying to build now.*
2822 * The return value for SSL_CIPHER_description() for error conditions
2827 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2829 Obtaining and performing DNSSEC validation of TLSA records is
2830 the application's responsibility. The application provides
2831 the TLSA records of its choice to OpenSSL, and these are then
2832 used to authenticate the peer.
2834 The TLSA records need not even come from DNS. They can, for
2835 example, be used to implement local end-entity certificate or
2836 trust-anchor "pinning", where the "pin" data takes the form
2837 of TLSA records, which can augment or replace verification
2838 based on the usual WebPKI public certification authorities.
2842 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2843 continues to support deprecated interfaces in default builds.
2844 However, applications are strongly advised to compile their
2845 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2846 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2847 or the 1.1.0 releases.
2849 In environments in which all applications have been ported to
2850 not use any deprecated interfaces OpenSSL's Configure script
2851 should be used with the --api=1.1.0 option to entirely remove
2852 support for the deprecated features from the library and
2853 unconditionally disable them in the installed headers.
2854 Essentially the same effect can be achieved with the "no-deprecated"
2855 argument to Configure, except that this will always restrict
2856 the build to just the latest API, rather than a fixed API
2859 As applications are ported to future revisions of the API,
2860 they should update their compile-time OPENSSL_API_COMPAT define
2861 accordingly, but in most cases should be able to continue to
2862 compile with later releases.
2864 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2865 0x10000000L and 0x00908000L, respectively. However those
2866 versions did not support the OPENSSL_API_COMPAT feature, and
2867 so applications are not typically tested for explicit support
2868 of just the undeprecated features of either release.
2872 * Add support for setting the minimum and maximum supported protocol.
2873 It can bet set via the SSL_set_min_proto_version() and
2874 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2875 MaxProtocol. It's recommended to use the new APIs to disable
2876 protocols instead of disabling individual protocols using
2877 SSL_set_options() or SSL_CONF's Protocol. This change also
2878 removes support for disabling TLS 1.2 in the OpenSSL TLS
2879 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2883 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2887 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2888 and integrates ECDSA and ECDH functionality into EC. Implementations can
2889 now redirect key generation and no longer need to convert to or from
2892 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2893 include the ec.h header file instead.
2897 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2898 ciphers who are no longer supported and drops support the ephemeral RSA key
2899 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2903 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2904 opaque. For HMAC_CTX, the following constructors and destructors
2907 HMAC_CTX *HMAC_CTX_new(void);
2908 void HMAC_CTX_free(HMAC_CTX *ctx);
2910 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2911 destroy such methods has been added. See EVP_MD_meth_new(3) and
2912 EVP_CIPHER_meth_new(3) for documentation.
2915 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2916 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2917 EVP_MD_CTX_reset() should be called instead to reinitialise
2918 an already created structure.
2919 2) For consistency with the majority of our object creators and
2920 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2921 EVP_MD_CTX_(new|free). The old names are retained as macros
2922 for deprecated builds.
2926 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2927 cryptographic operations to be performed asynchronously as long as an
2928 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2929 further details. Libssl has also had this capability integrated with the
2930 introduction of the new mode SSL_MODE_ASYNC and associated error
2931 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2932 pages. This work was developed in partnership with Intel Corp.
2936 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2937 always enabled now. If you want to disable the support you should
2938 exclude it using the list of supported ciphers. This also means that the
2939 "-no_ecdhe" option has been removed from s_server.
2943 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2944 SSL_{CTX_}set1_curves() which can set a list.
2948 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2949 curve you want to support using SSL_{CTX_}set1_curves().
2953 * State machine rewrite. The state machine code has been significantly
2954 refactored in order to remove much duplication of code and solve issues
2955 with the old code (see ssl/statem/README for further details). This change
2956 does have some associated API changes. Notably the SSL_state() function
2957 has been removed and replaced by SSL_get_state which now returns an
2958 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2959 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2964 * All instances of the string "ssleay" in the public API were replaced
2965 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2966 Some error codes related to internal RSA_eay API's were renamed.
2970 * The demo files in crypto/threads were moved to demo/threads.
2974 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2977 *Matt Caswell, Rich Salz*
2979 * New ASN.1 embed macro.
2981 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2982 structure is not allocated: it is part of the parent. That is instead of
2990 This reduces memory fragmentation and make it impossible to accidentally
2991 set a mandatory field to NULL.
2993 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2994 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2995 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3000 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3004 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3005 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3006 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3007 DES and RC4 ciphersuites.
3011 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3012 This changes the decoding behaviour for some invalid messages,
3013 though the change is mostly in the more lenient direction, and
3014 legacy behaviour is preserved as much as possible.
3018 * Fix no-stdio build.
3019 * David Woodhouse <David.Woodhouse@intel.com> and also
3020 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3022 * New testing framework
3023 The testing framework has been largely rewritten and is now using
3024 perl and the perl modules Test::Harness and an extended variant of
3025 Test::More called OpenSSL::Test to do its work. All test scripts in
3026 test/ have been rewritten into test recipes, and all direct calls to
3027 executables in test/Makefile have become individual recipes using the
3028 simplified testing OpenSSL::Test::Simple.
3030 For documentation on our testing modules, do:
3032 perldoc test/testlib/OpenSSL/Test/Simple.pm
3033 perldoc test/testlib/OpenSSL/Test.pm
3037 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3038 are used; the latter aborts on memory leaks (usually checked on exit).
3039 Some undocumented "set malloc, etc., hooks" functions were removed
3040 and others were changed. All are now documented.
3044 * In DSA_generate_parameters_ex, if the provided seed is too short,
3047 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3049 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3050 from RFC4279, RFC4785, RFC5487, RFC5489.
3052 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3053 original RSA_PSK patch.
3057 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3058 era flag was never set throughout the codebase (only read). Also removed
3059 SSL3_FLAGS_POP_BUFFER which was only used if
3060 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3064 * Changed the default name options in the "ca", "crl", "req" and "x509"
3065 to be "oneline" instead of "compat".
3069 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3070 not aware of clients that still exhibit this bug, and the workaround
3071 hasn't been working properly for a while.
3075 * The return type of BIO_number_read() and BIO_number_written() as well as
3076 the corresponding num_read and num_write members in the BIO structure has
3077 changed from unsigned long to uint64_t. On platforms where an unsigned
3078 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3083 * Given the pervasive nature of TLS extensions it is inadvisable to run
3084 OpenSSL without support for them. It also means that maintaining
3085 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3086 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3090 * Removed support for the two export grade static DH ciphersuites
3091 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3092 were newly added (along with a number of other static DH ciphersuites) to
3093 1.0.2. However the two export ones have *never* worked since they were
3094 introduced. It seems strange in any case to be adding new export
3095 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3099 * Version negotiation has been rewritten. In particular SSLv23_method(),
3100 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3101 and turned into macros which simply call the new preferred function names
3102 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3103 should use the new names instead. Also as part of this change the ssl23.h
3104 header file has been removed.
3108 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3109 code and the associated standard is no longer considered fit-for-purpose.
3113 * RT2547 was closed. When generating a private key, try to make the
3114 output file readable only by the owner. This behavior change might
3115 be noticeable when interacting with other software.
3117 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3122 * Added HTTP GET support to the ocsp command.
3126 * Changed default digest for the dgst and enc commands from MD5 to
3131 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3135 * Added support for TLS extended master secret from
3136 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3137 initial patch which was a great help during development.
3141 * All libssl internal structures have been removed from the public header
3142 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3143 now redundant). Users should not attempt to access internal structures
3144 directly. Instead they should use the provided API functions.
3148 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3149 Access to deprecated functions can be re-enabled by running config with
3150 "enable-deprecated". In addition applications wishing to use deprecated
3151 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3152 will, by default, disable some transitive includes that previously existed
3153 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3157 * Added support for OCB mode. OpenSSL has been granted a patent license
3158 compatible with the OpenSSL license for use of OCB. Details are available
3159 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
3160 for OCB can be removed by calling config with no-ocb.
3164 * SSLv2 support has been removed. It still supports receiving a SSLv2
3165 compatible client hello.
3169 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3170 done while fixing the error code for the key-too-small case.
3172 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3174 * CA.sh has been removed; use CA.pl instead.
3178 * Removed old DES API.
3182 * Remove various unsupported platforms:
3188 Sinix/ReliantUNIX RM400
3193 16-bit platforms such as WIN16
3197 * Clean up OPENSSL_NO_xxx #define's
3198 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3199 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3200 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3201 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3202 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3203 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3204 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3205 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3206 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3207 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3211 * Cleaned up dead code
3212 Remove all but one '#ifdef undef' which is to be looked at.
3216 * Clean up calling of xxx_free routines.
3217 Just like free(), fix most of the xxx_free routines to accept
3218 NULL. Remove the non-null checks from callers. Save much code.
3222 * Add secure heap for storage of private keys (when possible).
3223 Add BIO_s_secmem(), CBIGNUM, etc.
3224 Contributed by Akamai Technologies under our Corporate CLA.
3228 * Experimental support for a new, fast, unbiased prime candidate generator,
3229 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3231 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3233 * New output format NSS in the sess_id command line tool. This allows
3234 exporting the session id and the master key in NSS keylog format.
3236 *Martin Kaiser <martin@kaiser.cx>*
3238 * Harmonize version and its documentation. -f flag is used to display
3241 *mancha <mancha1@zoho.com>*
3243 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3244 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3246 *mancha <mancha1@zoho.com>*
3248 * Fix some double frees. These are not thought to be exploitable.
3250 *mancha <mancha1@zoho.com>*
3252 * A missing bounds check in the handling of the TLS heartbeat extension
3253 can be used to reveal up to 64k of memory to a connected client or
3256 Thanks for Neel Mehta of Google Security for discovering this bug and to
3257 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3258 preparing the fix [CVE-2014-0160][]
3260 *Adam Langley, Bodo Moeller*
3262 * Fix for the attack described in the paper "Recovering OpenSSL
3263 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3264 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3265 http://eprint.iacr.org/2014/140
3267 Thanks to Yuval Yarom and Naomi Benger for discovering this
3268 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3270 *Yuval Yarom and Naomi Benger*
3272 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3273 this fixes a limitation in previous versions of OpenSSL.
3277 * Experimental encrypt-then-mac support.
3279 Experimental support for encrypt then mac from
3280 draft-gutmann-tls-encrypt-then-mac-02.txt
3282 To enable it set the appropriate extension number (0x42 for the test
3283 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3285 For non-compliant peers (i.e. just about everything) this should have no
3288 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3293 * Add EVP support for key wrapping algorithms, to avoid problems with
3294 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3295 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3296 algorithms and include tests cases.
3300 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3305 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3306 MGF1 digest and OAEP label.
3310 * Make openssl verify return errors.
3312 *Chris Palmer <palmer@google.com> and Ben Laurie*
3314 * New function ASN1_TIME_diff to calculate the difference between two
3315 ASN1_TIME structures or one structure and the current time.
3319 * Update fips_test_suite to support multiple command line options. New
3320 test to induce all self test errors in sequence and check expected
3325 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3326 sign or verify all in one operation.
3330 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3331 test programs and fips_test_suite. Includes functionality to parse
3332 the minimal script output of fipsalgest.pl directly.
3336 * Add authorisation parameter to FIPS_module_mode_set().
3340 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3344 * Use separate DRBG fields for internal and external flags. New function
3345 FIPS_drbg_health_check() to perform on demand health checking. Add
3346 generation tests to fips_test_suite with reduced health check interval to
3347 demonstrate periodic health checking. Add "nodh" option to
3348 fips_test_suite to skip very slow DH test.
3352 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3357 * More extensive health check for DRBG checking many more failure modes.
3358 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3359 combination: call this in fips_test_suite.
3363 * Add support for canonical generation of DSA parameter 'g'. See
3366 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3367 POST to handle HMAC cases.
3371 * Add functions FIPS_module_version() and FIPS_module_version_text()
3372 to return numerical and string versions of the FIPS module number.
3376 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3377 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3378 outside the validated module in the FIPS capable OpenSSL.
3382 * Minor change to DRBG entropy callback semantics. In some cases
3383 there is no multiple of the block length between min_len and
3384 max_len. Allow the callback to return more than max_len bytes
3385 of entropy but discard any extra: it is the callback's responsibility
3386 to ensure that the extra data discarded does not impact the
3387 requested amount of entropy.
3391 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3392 information in FIPS186-3, SP800-57 and SP800-131A.
3396 * CCM support via EVP. Interface is very similar to GCM case except we
3397 must supply all data in one chunk (i.e. no update, final) and the
3398 message length must be supplied if AAD is used. Add algorithm test
3403 * Initial version of POST overhaul. Add POST callback to allow the status
3404 of POST to be monitored and/or failures induced. Modify fips_test_suite
3405 to use callback. Always run all selftests even if one fails.
3409 * XTS support including algorithm test driver in the fips_gcmtest program.
3410 Note: this does increase the maximum key length from 32 to 64 bytes but
3411 there should be no binary compatibility issues as existing applications
3412 will never use XTS mode.
3416 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3417 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3418 performs algorithm blocking for unapproved PRNG types. Also do not
3419 set PRNG type in FIPS_mode_set(): leave this to the application.
3420 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3421 the standard OpenSSL PRNG: set additional data to a date time vector.
3425 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3426 This shouldn't present any incompatibility problems because applications
3427 shouldn't be using these directly and any that are will need to rethink
3428 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3432 * Extensive self tests and health checking required by SP800-90 DRBG.
3433 Remove strength parameter from FIPS_drbg_instantiate and always
3434 instantiate at maximum supported strength.
3438 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3442 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3446 * New function DH_compute_key_padded() to compute a DH key and pad with
3447 leading zeroes if needed: this complies with SP800-56A et al.
3451 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3452 anything, incomplete, subject to change and largely untested at present.
3456 * Modify fipscanisteronly build option to only build the necessary object
3457 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3461 * Add experimental option FIPSSYMS to give all symbols in
3462 fipscanister.o and FIPS or fips prefix. This will avoid
3463 conflicts with future versions of OpenSSL. Add perl script
3464 util/fipsas.pl to preprocess assembly language source files
3465 and rename any affected symbols.
3469 * Add selftest checks and algorithm block of non-fips algorithms in
3470 FIPS mode. Remove DES2 from selftests.
3474 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3475 return internal method without any ENGINE dependencies. Add new
3476 tiny fips sign and verify functions.
3480 * New build option no-ec2m to disable characteristic 2 code.
3484 * New build option "fipscanisteronly". This only builds fipscanister.o
3485 and (currently) associated fips utilities. Uses the file Makefile.fips
3486 instead of Makefile.org as the prototype.
3490 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3491 Update fips_gcmtest to use IV generator.
3495 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3496 setting output buffer to NULL. The *Final function must be
3497 called although it will not retrieve any additional data. The tag
3498 can be set or retrieved with a ctrl. The IV length is by default 12
3499 bytes (96 bits) but can be set to an alternative value. If the IV
3500 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3505 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3506 underlying do_cipher function handles all cipher semantics itself
3507 including padding and finalisation. This is useful if (for example)
3508 an ENGINE cipher handles block padding itself. The behaviour of
3509 do_cipher is subtly changed if this flag is set: the return value
3510 is the number of characters written to the output buffer (zero is
3511 no longer an error code) or a negative error code. Also if the
3512 input buffer is NULL and length 0 finalisation should be performed.
3516 * If a candidate issuer certificate is already part of the constructed
3517 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3521 * Improve forward-security support: add functions
3523 void SSL_CTX_set_not_resumable_session_callback(
3524 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3525 void SSL_set_not_resumable_session_callback(
3526 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3528 for use by SSL/TLS servers; the callback function will be called whenever a
3529 new session is created, and gets to decide whether the session may be
3530 cached to make it resumable (return 0) or not (return 1). (As by the
3531 SSL/TLS protocol specifications, the session_id sent by the server will be
3532 empty to indicate that the session is not resumable; also, the server will
3533 not generate RFC 4507 (RFC 5077) session tickets.)
3535 A simple reasonable callback implementation is to return is_forward_secure.
3536 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3537 by the SSL/TLS server library, indicating whether it can provide forward
3540 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3542 * New -verify_name option in command line utilities to set verification
3547 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3548 Add CMAC pkey methods.
3552 * Experimental renegotiation in s_server -www mode. If the client
3553 browses /reneg connection is renegotiated. If /renegcert it is
3554 renegotiated requesting a certificate.
3558 * Add an "external" session cache for debugging purposes to s_server. This
3559 should help trace issues which normally are only apparent in deployed
3560 multi-process servers.
3564 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3565 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3566 BIO_set_cipher() and some obscure PEM functions were changed so they
3567 can now return an error. The RAND changes required a change to the
3568 RAND_METHOD structure.
3572 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3573 a gcc attribute to warn if the result of a function is ignored. This
3574 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3575 whose return value is often ignored.
3579 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3580 These allow SCTs (signed certificate timestamps) to be requested and
3581 validated when establishing a connection.
3583 *Rob Percival <robpercival@google.com>*
3588 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019] ###
3590 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3591 used even when parsing explicit parameters, when loading a serialized key
3592 or calling `EC_GROUP_new_from_ecpkparameters()`/
3593 `EC_GROUP_new_from_ecparameters()`.
3594 This prevents bypass of security hardening and performance gains,
3595 especially for curves with specialized EC_METHODs.
3596 By default, if a key encoded with explicit parameters is loaded and later
3597 serialized, the output is still encoded with explicit parameters, even if
3598 internally a "named" EC_GROUP is used for computation.
3602 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3603 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3604 NULL. After this change, only the cofactor parameter can be NULL. It also
3605 does some minimal sanity checks on the passed order.
3610 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3611 An attack is simple, if the first CMS_recipientInfo is valid but the
3612 second CMS_recipientInfo is chosen ciphertext. If the second
3613 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3614 encryption key will be replaced by garbage, and the message cannot be
3615 decoded, but if the RSA decryption fails, the correct encryption key is
3616 used and the recipient will not notice the attack.
3617 As a work around for this potential attack the length of the decrypted
3618 key must be equal to the cipher default key length, in case the
3619 certifiate is not given and all recipientInfo are tried out.
3620 The old behaviour can be re-enabled in the CMS code by setting the
3621 CMS_DEBUG_DECRYPT flag.
3626 * Document issue with installation paths in diverse Windows builds
3628 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3629 binaries and run-time config file.
3634 ### Changes between 1.0.2r and 1.0.2s [28 May 2019] ###
3636 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3637 This changes the size when using the genpkey app when no size is given. It
3638 fixes an omission in earlier changes that changed all RSA, DSA and DH
3639 generation apps to use 2048 bits by default.
3643 * Add FIPS support for Android Arm 64-bit
3645 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3646 Module in Version 2.0.10. For some reason, the corresponding target
3647 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3648 built with FIPS support on Android Arm 64-bit. This omission has been
3651 *Matthias St. Pierre*
3653 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019] ###
3655 * 0-byte record padding oracle
3657 If an application encounters a fatal protocol error and then calls
3658 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3659 then OpenSSL can respond differently to the calling application if a 0 byte
3660 record is received with invalid padding compared to if a 0 byte record is
3661 received with an invalid MAC. If the application then behaves differently
3662 based on that in a way that is detectable to the remote peer, then this
3663 amounts to a padding oracle that could be used to decrypt data.
3665 In order for this to be exploitable "non-stitched" ciphersuites must be in
3666 use. Stitched ciphersuites are optimised implementations of certain
3667 commonly used ciphersuites. Also the application must call SSL_shutdown()
3668 twice even if a protocol error has occurred (applications should not do
3669 this but some do anyway).
3671 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3672 Aviram, with additional investigation by Steven Collison and Andrew
3673 Hourselt. It was reported to OpenSSL on 10th December 2018.
3678 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3682 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018] ###
3684 * Microarchitecture timing vulnerability in ECC scalar multiplication
3686 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3687 shown to be vulnerable to a microarchitecture timing side channel attack.
3688 An attacker with sufficient access to mount local timing attacks during
3689 ECDSA signature generation could recover the private key.
3691 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3692 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3698 * Timing vulnerability in DSA signature generation
3700 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3701 timing side channel attack. An attacker could use variations in the signing
3702 algorithm to recover the private key.
3704 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3709 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3710 Module, accidentally introduced while backporting security fixes from the
3711 development branch and hindering the use of ECC in FIPS mode.
3715 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018] ###
3717 * Client DoS due to large DH parameter
3719 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3720 malicious server can send a very large prime value to the client. This will
3721 cause the client to spend an unreasonably long period of time generating a
3722 key for this prime resulting in a hang until the client has finished. This
3723 could be exploited in a Denial Of Service attack.
3725 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3730 * Cache timing vulnerability in RSA Key Generation
3732 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3733 a cache timing side channel attack. An attacker with sufficient access to
3734 mount cache timing attacks during the RSA key generation process could
3735 recover the private key.
3737 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3738 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3743 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3744 parameter is no longer accepted, as it leads to a corrupt table. NULL
3745 pem_str is reserved for alias entries only.
3749 * Revert blinding in ECDSA sign and instead make problematic addition
3750 length-invariant. Switch even to fixed-length Montgomery multiplication.
3754 * Change generating and checking of primes so that the error rate of not
3755 being prime depends on the intended use based on the size of the input.
3756 For larger primes this will result in more rounds of Miller-Rabin.
3757 The maximal error rate for primes with more than 1080 bits is lowered
3760 *Kurt Roeckx, Annie Yousar*
3762 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3766 * Add blinding to ECDSA and DSA signatures to protect against side channel
3767 attacks discovered by Keegan Ryan (NCC Group).
3771 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3772 now allow empty (zero character) pass phrases.
3776 * Certificate time validation (X509_cmp_time) enforces stricter
3777 compliance with RFC 5280. Fractional seconds and timezone offsets
3778 are no longer allowed.
3782 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018] ###
3784 * Constructed ASN.1 types with a recursive definition could exceed the stack
3786 Constructed ASN.1 types with a recursive definition (such as can be found
3787 in PKCS7) could eventually exceed the stack given malicious input with
3788 excessive recursion. This could result in a Denial Of Service attack. There
3789 are no such structures used within SSL/TLS that come from untrusted sources
3790 so this is considered safe.
3792 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3798 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017] ###
3800 * Read/write after SSL object in error state
3802 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3803 mechanism. The intent was that if a fatal error occurred during a handshake
3804 then OpenSSL would move into the error state and would immediately fail if
3805 you attempted to continue the handshake. This works as designed for the
3806 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3807 SSL_connect()), however due to a bug it does not work correctly if
3808 SSL_read() or SSL_write() is called directly. In that scenario, if the
3809 handshake fails then a fatal error will be returned in the initial function
3810 call. If SSL_read()/SSL_write() is subsequently called by the application
3811 for the same SSL object then it will succeed and the data is passed without
3812 being decrypted/encrypted directly from the SSL/TLS record layer.
3814 In order to exploit this issue an application bug would have to be present
3815 that resulted in a call to SSL_read()/SSL_write() being issued after having
3816 already received a fatal error.
3818 This issue was reported to OpenSSL by David Benjamin (Google).
3823 * rsaz_1024_mul_avx2 overflow bug on x86_64
3825 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3826 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3827 Analysis suggests that attacks against RSA and DSA as a result of this
3828 defect would be very difficult to perform and are not believed likely.
3829 Attacks against DH1024 are considered just feasible, because most of the
3830 work necessary to deduce information about a private key may be performed
3831 offline. The amount of resources required for such an attack would be
3832 significant. However, for an attack on TLS to be meaningful, the server
3833 would have to share the DH1024 private key among multiple clients, which is
3834 no longer an option since CVE-2016-0701.
3836 This only affects processors that support the AVX2 but not ADX extensions
3837 like Intel Haswell (4th generation).
3839 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3840 was originally found via the OSS-Fuzz project.
3845 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017] ###
3847 * bn_sqrx8x_internal carry bug on x86_64
3849 There is a carry propagating bug in the x86_64 Montgomery squaring
3850 procedure. No EC algorithms are affected. Analysis suggests that attacks
3851 against RSA and DSA as a result of this defect would be very difficult to
3852 perform and are not believed likely. Attacks against DH are considered just
3853 feasible (although very difficult) because most of the work necessary to
3854 deduce information about a private key may be performed offline. The amount
3855 of resources required for such an attack would be very significant and
3856 likely only accessible to a limited number of attackers. An attacker would
3857 additionally need online access to an unpatched system using the target
3858 private key in a scenario with persistent DH parameters and a private
3859 key that is shared between multiple clients.
3861 This only affects processors that support the BMI1, BMI2 and ADX extensions
3862 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3864 This issue was reported to OpenSSL by the OSS-Fuzz project.
3869 * Malformed X.509 IPAddressFamily could cause OOB read
3871 If an X.509 certificate has a malformed IPAddressFamily extension,
3872 OpenSSL could do a one-byte buffer overread. The most likely result
3873 would be an erroneous display of the certificate in text format.
3875 This issue was reported to OpenSSL by the OSS-Fuzz project.
3880 ### Changes between 1.0.2k and 1.0.2l [25 May 2017] ###
3882 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3883 platform rather than 'mingw'.
3887 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017] ###
3889 * Truncated packet could crash via OOB read
3891 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3892 cipher is being used, then a truncated packet can cause that host to
3893 perform an out-of-bounds read, usually resulting in a crash.
3895 This issue was reported to OpenSSL by Robert Święcki of Google.
3900 * BN_mod_exp may produce incorrect results on x86_64
3902 There is a carry propagating bug in the x86_64 Montgomery squaring
3903 procedure. No EC algorithms are affected. Analysis suggests that attacks
3904 against RSA and DSA as a result of this defect would be very difficult to
3905 perform and are not believed likely. Attacks against DH are considered just
3906 feasible (although very difficult) because most of the work necessary to
3907 deduce information about a private key may be performed offline. The amount
3908 of resources required for such an attack would be very significant and
3909 likely only accessible to a limited number of attackers. An attacker would
3910 additionally need online access to an unpatched system using the target
3911 private key in a scenario with persistent DH parameters and a private
3912 key that is shared between multiple clients. For example this can occur by
3913 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3914 similar to CVE-2015-3193 but must be treated as a separate problem.
3916 This issue was reported to OpenSSL by the OSS-Fuzz project.
3921 * Montgomery multiplication may produce incorrect results
3923 There is a carry propagating bug in the Broadwell-specific Montgomery
3924 multiplication procedure that handles input lengths divisible by, but
3925 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3926 and DH private keys are impossible. This is because the subroutine in
3927 question is not used in operations with the private key itself and an input
3928 of the attacker's direct choice. Otherwise the bug can manifest itself as
3929 transient authentication and key negotiation failures or reproducible
3930 erroneous outcome of public-key operations with specially crafted input.
3931 Among EC algorithms only Brainpool P-512 curves are affected and one
3932 presumably can attack ECDH key negotiation. Impact was not analyzed in
3933 detail, because pre-requisites for attack are considered unlikely. Namely
3934 multiple clients have to choose the curve in question and the server has to
3935 share the private key among them, neither of which is default behaviour.
3936 Even then only clients that chose the curve will be affected.
3938 This issue was publicly reported as transient failures and was not
3939 initially recognized as a security issue. Thanks to Richard Morgan for
3940 providing reproducible case.
3945 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3946 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3947 prevent issues where no progress is being made and the peer continually
3948 sends unrecognised record types, using up resources processing them.
3952 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016] ###
3954 * Missing CRL sanity check
3956 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
3957 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
3958 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
3960 This issue only affects the OpenSSL 1.0.2i
3965 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016] ###
3967 * OCSP Status Request extension unbounded memory growth
3969 A malicious client can send an excessively large OCSP Status Request
3970 extension. If that client continually requests renegotiation, sending a
3971 large OCSP Status Request extension each time, then there will be unbounded
3972 memory growth on the server. This will eventually lead to a Denial Of
3973 Service attack through memory exhaustion. Servers with a default
3974 configuration are vulnerable even if they do not support OCSP. Builds using
3975 the "no-ocsp" build time option are not affected.
3977 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3982 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
3985 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
3991 * OOB write in MDC2_Update()
3993 An overflow can occur in MDC2_Update() either if called directly or
3994 through the EVP_DigestUpdate() function using MDC2. If an attacker
3995 is able to supply very large amounts of input data after a previous
3996 call to EVP_EncryptUpdate() with a partial block then a length check
3997 can overflow resulting in a heap corruption.
3999 The amount of data needed is comparable to SIZE_MAX which is impractical
4002 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4007 * Malformed SHA512 ticket DoS
4009 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4010 DoS attack where a malformed ticket will result in an OOB read which will
4013 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4014 a custom server callback and ticket lookup mechanism.
4016 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4021 * OOB write in BN_bn2dec()
4023 The function BN_bn2dec() does not check the return value of BN_div_word().
4024 This can cause an OOB write if an application uses this function with an
4025 overly large BIGNUM. This could be a problem if an overly large certificate
4026 or CRL is printed out from an untrusted source. TLS is not affected because
4027 record limits will reject an oversized certificate before it is parsed.
4029 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4034 * OOB read in TS_OBJ_print_bio()
4036 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4037 the total length the OID text representation would use and not the amount
4038 of data written. This will result in OOB reads when large OIDs are
4041 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4046 * Pointer arithmetic undefined behaviour
4048 Avoid some undefined pointer arithmetic
4050 A common idiom in the codebase is to check limits in the following manner:
4053 Where "p" points to some malloc'd data of SIZE bytes and
4056 "len" here could be from some externally supplied data (e.g. from a TLS
4059 The rules of C pointer arithmetic are such that "p + len" is only well
4060 defined where len <= SIZE. Therefore the above idiom is actually
4061 undefined behaviour.
4063 For example this could cause problems if some malloc implementation
4064 provides an address for "p" such that "p + len" actually overflows for
4065 values of len that are too big and therefore p + len < limit.
4067 This issue was reported to OpenSSL by Guido Vranken
4072 * Constant time flag not preserved in DSA signing
4074 Operations in the DSA signing algorithm should run in constant time in
4075 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4076 implementation means that a non-constant time codepath is followed for
4077 certain operations. This has been demonstrated through a cache-timing
4078 attack to be sufficient for an attacker to recover the private DSA key.
4080 This issue was reported by César Pereida (Aalto University), Billy Brumley
4081 (Tampere University of Technology), and Yuval Yarom (The University of
4082 Adelaide and NICTA).
4087 * DTLS buffered message DoS
4089 In a DTLS connection where handshake messages are delivered out-of-order
4090 those messages that OpenSSL is not yet ready to process will be buffered
4091 for later use. Under certain circumstances, a flaw in the logic means that
4092 those messages do not get removed from the buffer even though the handshake
4093 has been completed. An attacker could force up to approx. 15 messages to
4094 remain in the buffer when they are no longer required. These messages will
4095 be cleared when the DTLS connection is closed. The default maximum size for
4096 a message is 100k. Therefore the attacker could force an additional 1500k
4097 to be consumed per connection. By opening many simulataneous connections an
4098 attacker could cause a DoS attack through memory exhaustion.
4100 This issue was reported to OpenSSL by Quan Luo.
4105 * DTLS replay protection DoS
4107 A flaw in the DTLS replay attack protection mechanism means that records
4108 that arrive for future epochs update the replay protection "window" before
4109 the MAC for the record has been validated. This could be exploited by an
4110 attacker by sending a record for the next epoch (which does not have to
4111 decrypt or have a valid MAC), with a very large sequence number. This means
4112 that all subsequent legitimate packets are dropped causing a denial of
4113 service for a specific DTLS connection.
4115 This issue was reported to OpenSSL by the OCAP audit team.
4120 * Certificate message OOB reads
4122 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4123 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4124 theoretical DoS risk but this has not been observed in practice on common
4127 The messages affected are client certificate, client certificate request
4128 and server certificate. As a result the attack can only be performed
4129 against a client or a server which enables client authentication.
4131 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4136 ### Changes between 1.0.2g and 1.0.2h [3 May 2016] ###
4138 * Prevent padding oracle in AES-NI CBC MAC check
4140 A MITM attacker can use a padding oracle attack to decrypt traffic
4141 when the connection uses an AES CBC cipher and the server support
4144 This issue was introduced as part of the fix for Lucky 13 padding
4145 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4146 constant time by making sure that always the same bytes are read and
4147 compared against either the MAC or padding bytes. But it no longer
4148 checked that there was enough data to have both the MAC and padding
4151 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4156 * Fix EVP_EncodeUpdate overflow
4158 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4159 Base64 encoding of binary data. If an attacker is able to supply very large
4160 amounts of input data then a length check can overflow resulting in a heap
4163 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4164 the PEM_write_bio* family of functions. These are mainly used within the
4165 OpenSSL command line applications, so any application which processes data
4166 from an untrusted source and outputs it as a PEM file should be considered
4167 vulnerable to this issue. User applications that call these APIs directly
4168 with large amounts of untrusted data may also be vulnerable.
4170 This issue was reported by Guido Vranken.
4175 * Fix EVP_EncryptUpdate overflow
4177 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4178 is able to supply very large amounts of input data after a previous call to
4179 EVP_EncryptUpdate() with a partial block then a length check can overflow
4180 resulting in a heap corruption. Following an analysis of all OpenSSL
4181 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4182 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4183 the first called function after an EVP_EncryptInit(), and therefore that
4184 specific call must be safe. The second form is where the length passed to
4185 EVP_EncryptUpdate() can be seen from the code to be some small value and
4186 therefore there is no possibility of an overflow. Since all instances are
4187 one of these two forms, it is believed that there can be no overflows in
4188 internal code due to this problem. It should be noted that
4189 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4190 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4191 of these calls have also been analysed too and it is believed there are no
4192 instances in internal usage where an overflow could occur.
4194 This issue was reported by Guido Vranken.
4199 * Prevent ASN.1 BIO excessive memory allocation
4201 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4202 a short invalid encoding can cause allocation of large amounts of memory
4203 potentially consuming excessive resources or exhausting memory.
4205 Any application parsing untrusted data through d2i BIO functions is
4206 affected. The memory based functions such as d2i_X509() are *not* affected.
4207 Since the memory based functions are used by the TLS library, TLS
4208 applications are not affected.
4210 This issue was reported by Brian Carpenter.
4217 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4218 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4219 in arbitrary stack data being returned in the buffer.
4221 This issue was reported by Guido Vranken.
4226 * Modify behavior of ALPN to invoke callback after SNI/servername
4227 callback, such that updates to the SSL_CTX affect ALPN.
4231 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4236 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4237 methods are enabled and ssl2 is disabled the methods return NULL.
4241 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016] ###
4243 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4244 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4245 provide any "EXPORT" or "LOW" strength ciphers.
4249 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4250 is by default disabled at build-time. Builds that are not configured with
4251 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4252 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4253 will need to explicitly call either of:
4255 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4257 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4259 as appropriate. Even if either of those is used, or the application
4260 explicitly uses the version-specific SSLv2_method() or its client and
4261 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4262 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4263 ciphers, and SSLv2 56-bit DES are no longer available.
4268 * Fix a double-free in DSA code
4270 A double free bug was discovered when OpenSSL parses malformed DSA private
4271 keys and could lead to a DoS attack or memory corruption for applications
4272 that receive DSA private keys from untrusted sources. This scenario is
4275 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4281 * Disable SRP fake user seed to address a server memory leak.
4283 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4285 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4286 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4287 was changed to ignore the "fake user" SRP seed, even if the seed
4290 Users should use SRP_VBASE_get1_by_user instead. Note that in
4291 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4292 also that even though configuring the SRP seed attempts to hide
4293 invalid usernames by continuing the handshake with fake
4294 credentials, this behaviour is not constant time and no strong
4295 guarantees are made that the handshake is indistinguishable from
4296 that of a valid user.
4301 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4303 In the BN_hex2bn function the number of hex digits is calculated using an
4304 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4305 large values of |i| this can result in |bn_expand| not allocating any
4306 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4307 field as NULL leading to a subsequent NULL ptr deref. For very large values
4308 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4309 In this case memory is allocated to the internal BIGNUM data field, but it
4310 is insufficiently sized leading to heap corruption. A similar issue exists
4311 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4312 is ever called by user applications with very large untrusted hex/dec data.
4313 This is anticipated to be a rare occurrence.
4315 All OpenSSL internal usage of these functions use data that is not expected
4316 to be untrusted, e.g. config file data or application command line
4317 arguments. If user developed applications generate config file data based
4318 on untrusted data then it is possible that this could also lead to security
4319 consequences. This is also anticipated to be rare.
4321 This issue was reported to OpenSSL by Guido Vranken.
4326 * Fix memory issues in BIO_*printf functions
4328 The internal |fmtstr| function used in processing a "%s" format string in
4329 the BIO_*printf functions could overflow while calculating the length of a
4330 string and cause an OOB read when printing very long strings.
4332 Additionally the internal |doapr_outch| function can attempt to write to an
4333 OOB memory location (at an offset from the NULL pointer) in the event of a
4334 memory allocation failure. In 1.0.2 and below this could be caused where
4335 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4336 could be in processing a very long "%s" format string. Memory leaks can
4339 The first issue may mask the second issue dependent on compiler behaviour.
4340 These problems could enable attacks where large amounts of untrusted data
4341 is passed to the BIO_*printf functions. If applications use these functions
4342 in this way then they could be vulnerable. OpenSSL itself uses these
4343 functions when printing out human-readable dumps of ASN.1 data. Therefore
4344 applications that print this data could be vulnerable if the data is from
4345 untrusted sources. OpenSSL command line applications could also be
4346 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4347 as command line arguments.
4349 Libssl is not considered directly vulnerable. Additionally certificates etc
4350 received via remote connections via libssl are also unlikely to be able to
4351 trigger these issues because of message size limits enforced within libssl.
4353 This issue was reported to OpenSSL Guido Vranken.
4358 * Side channel attack on modular exponentiation
4360 A side-channel attack was found which makes use of cache-bank conflicts on
4361 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4362 of RSA keys. The ability to exploit this issue is limited as it relies on
4363 an attacker who has control of code in a thread running on the same
4364 hyper-threaded core as the victim thread which is performing decryptions.
4366 This issue was reported to OpenSSL by Yuval Yarom, The University of
4367 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4368 Nadia Heninger, University of Pennsylvania with more information at
4369 http://cachebleed.info.
4374 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4375 if no keysize is specified with default_bits. This fixes an
4376 omission in an earlier change that changed all RSA/DSA key generation
4377 apps to use 2048 bits by default.
4381 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016] ###
4382 * DH small subgroups
4384 Historically OpenSSL only ever generated DH parameters based on "safe"
4385 primes. More recently (in version 1.0.2) support was provided for
4386 generating X9.42 style parameter files such as those required for RFC 5114
4387 support. The primes used in such files may not be "safe". Where an
4388 application is using DH configured with parameters based on primes that are
4389 not "safe" then an attacker could use this fact to find a peer's private
4390 DH exponent. This attack requires that the attacker complete multiple
4391 handshakes in which the peer uses the same private DH exponent. For example
4392 this could be used to discover a TLS server's private DH exponent if it's
4393 reusing the private DH exponent or it's using a static DH ciphersuite.
4395 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4396 TLS. It is not on by default. If the option is not set then the server
4397 reuses the same private DH exponent for the life of the server process and
4398 would be vulnerable to this attack. It is believed that many popular
4399 applications do set this option and would therefore not be at risk.
4401 The fix for this issue adds an additional check where a "q" parameter is
4402 available (as is the case in X9.42 based parameters). This detects the
4403 only known attack, and is the only possible defense for static DH
4404 ciphersuites. This could have some performance impact.
4406 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4407 default and cannot be disabled. This could have some performance impact.
4409 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4414 * SSLv2 doesn't block disabled ciphers
4416 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4417 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4418 been disabled, provided that the SSLv2 protocol was not also disabled via
4421 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4422 and Sebastian Schinzel.
4427 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015] ###
4429 * BN_mod_exp may produce incorrect results on x86_64
4431 There is a carry propagating bug in the x86_64 Montgomery squaring
4432 procedure. No EC algorithms are affected. Analysis suggests that attacks
4433 against RSA and DSA as a result of this defect would be very difficult to
4434 perform and are not believed likely. Attacks against DH are considered just
4435 feasible (although very difficult) because most of the work necessary to
4436 deduce information about a private key may be performed offline. The amount
4437 of resources required for such an attack would be very significant and
4438 likely only accessible to a limited number of attackers. An attacker would
4439 additionally need online access to an unpatched system using the target
4440 private key in a scenario with persistent DH parameters and a private
4441 key that is shared between multiple clients. For example this can occur by
4442 default in OpenSSL DHE based SSL/TLS ciphersuites.
4444 This issue was reported to OpenSSL by Hanno Böck.
4449 * Certificate verify crash with missing PSS parameter
4451 The signature verification routines will crash with a NULL pointer
4452 dereference if presented with an ASN.1 signature using the RSA PSS
4453 algorithm and absent mask generation function parameter. Since these
4454 routines are used to verify certificate signature algorithms this can be
4455 used to crash any certificate verification operation and exploited in a
4456 DoS attack. Any application which performs certificate verification is
4457 vulnerable including OpenSSL clients and servers which enable client
4460 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4465 * X509_ATTRIBUTE memory leak
4467 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4468 memory. This structure is used by the PKCS#7 and CMS routines so any
4469 application which reads PKCS#7 or CMS data from untrusted sources is
4470 affected. SSL/TLS is not affected.
4472 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4478 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4479 This changes the decoding behaviour for some invalid messages,
4480 though the change is mostly in the more lenient direction, and
4481 legacy behaviour is preserved as much as possible.
4485 * In DSA_generate_parameters_ex, if the provided seed is too short,
4488 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4490 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015] ###
4492 * Alternate chains certificate forgery
4494 During certificate verification, OpenSSL will attempt to find an
4495 alternative certificate chain if the first attempt to build such a chain
4496 fails. An error in the implementation of this logic can mean that an
4497 attacker could cause certain checks on untrusted certificates to be
4498 bypassed, such as the CA flag, enabling them to use a valid leaf
4499 certificate to act as a CA and "issue" an invalid certificate.
4501 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4506 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015] ###
4508 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4509 incompatibility in the handling of HMAC. The previous ABI has now been
4514 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015] ###
4516 * Malformed ECParameters causes infinite loop
4518 When processing an ECParameters structure OpenSSL enters an infinite loop
4519 if the curve specified is over a specially malformed binary polynomial
4522 This can be used to perform denial of service against any
4523 system which processes public keys, certificate requests or
4524 certificates. This includes TLS clients and TLS servers with
4525 client authentication enabled.
4527 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4532 * Exploitable out-of-bounds read in X509_cmp_time
4534 X509_cmp_time does not properly check the length of the ASN1_TIME
4535 string and can read a few bytes out of bounds. In addition,
4536 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4539 An attacker can use this to craft malformed certificates and CRLs of
4540 various sizes and potentially cause a segmentation fault, resulting in
4541 a DoS on applications that verify certificates or CRLs. TLS clients
4542 that verify CRLs are affected. TLS clients and servers with client
4543 authentication enabled may be affected if they use custom verification
4546 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4547 independently by Hanno Böck.
4552 * PKCS7 crash with missing EnvelopedContent
4554 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4555 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4556 with missing content and trigger a NULL pointer dereference on parsing.
4558 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4559 structures from untrusted sources are affected. OpenSSL clients and
4560 servers are not affected.
4562 This issue was reported to OpenSSL by Michal Zalewski (Google).
4567 * CMS verify infinite loop with unknown hash function
4569 When verifying a signedData message the CMS code can enter an infinite loop
4570 if presented with an unknown hash function OID. This can be used to perform
4571 denial of service against any system which verifies signedData messages using
4573 This issue was reported to OpenSSL by Johannes Bauer.
4578 * Race condition handling NewSessionTicket
4580 If a NewSessionTicket is received by a multi-threaded client when attempting to
4581 reuse a previous ticket then a race condition can occur potentially leading to
4582 a double free of the ticket data.
4587 * Only support 256-bit or stronger elliptic curves with the
4588 'ecdh_auto' setting (server) or by default (client). Of supported
4589 curves, prefer P-256 (both).
4593 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015] ###
4595 * ClientHello sigalgs DoS fix
4597 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4598 invalid signature algorithms extension a NULL pointer dereference will
4599 occur. This can be exploited in a DoS attack against the server.
4601 This issue was was reported to OpenSSL by David Ramos of Stanford
4605 *Stephen Henson and Matt Caswell*
4607 * Multiblock corrupted pointer fix
4609 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4610 feature only applies on 64 bit x86 architecture platforms that support AES
4611 NI instructions. A defect in the implementation of "multiblock" can cause
4612 OpenSSL's internal write buffer to become incorrectly set to NULL when
4613 using non-blocking IO. Typically, when the user application is using a
4614 socket BIO for writing, this will only result in a failed connection.
4615 However if some other BIO is used then it is likely that a segmentation
4616 fault will be triggered, thus enabling a potential DoS attack.
4618 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4623 * Segmentation fault in DTLSv1_listen fix
4625 The DTLSv1_listen function is intended to be stateless and processes the
4626 initial ClientHello from many peers. It is common for user code to loop
4627 over the call to DTLSv1_listen until a valid ClientHello is received with
4628 an associated cookie. A defect in the implementation of DTLSv1_listen means
4629 that state is preserved in the SSL object from one invocation to the next
4630 that can lead to a segmentation fault. Errors processing the initial
4631 ClientHello can trigger this scenario. An example of such an error could be
4632 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4635 This issue was reported to OpenSSL by Per Allansson.
4640 * Segmentation fault in ASN1_TYPE_cmp fix
4642 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4643 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4644 certificate signature algorithm consistency this can be used to crash any
4645 certificate verification operation and exploited in a DoS attack. Any
4646 application which performs certificate verification is vulnerable including
4647 OpenSSL clients and servers which enable client authentication.
4652 * Segmentation fault for invalid PSS parameters fix
4654 The signature verification routines will crash with a NULL pointer
4655 dereference if presented with an ASN.1 signature using the RSA PSS
4656 algorithm and invalid parameters. Since these routines are used to verify
4657 certificate signature algorithms this can be used to crash any
4658 certificate verification operation and exploited in a DoS attack. Any
4659 application which performs certificate verification is vulnerable including
4660 OpenSSL clients and servers which enable client authentication.
4662 This issue was was reported to OpenSSL by Brian Carpenter.
4667 * ASN.1 structure reuse memory corruption fix
4669 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4670 memory corruption via an invalid write. Such reuse is and has been
4671 strongly discouraged and is believed to be rare.
4673 Applications that parse structures containing CHOICE or ANY DEFINED BY
4674 components may be affected. Certificate parsing (d2i_X509 and related
4675 functions) are however not affected. OpenSSL clients and servers are
4681 * PKCS7 NULL pointer dereferences fix
4683 The PKCS#7 parsing code does not handle missing outer ContentInfo
4684 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4685 missing content and trigger a NULL pointer dereference on parsing.
4687 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4688 otherwise parse PKCS#7 structures from untrusted sources are
4689 affected. OpenSSL clients and servers are not affected.
4691 This issue was reported to OpenSSL by Michal Zalewski (Google).
4696 * DoS via reachable assert in SSLv2 servers fix
4698 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4699 servers that both support SSLv2 and enable export cipher suites by sending
4700 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4702 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4703 (OpenSSL development team).
4708 * Empty CKE with client auth and DHE fix
4710 If client auth is used then a server can seg fault in the event of a DHE
4711 ciphersuite being selected and a zero length ClientKeyExchange message
4712 being sent by the client. This could be exploited in a DoS attack.
4717 * Handshake with unseeded PRNG fix
4719 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4720 with an unseeded PRNG. The conditions are:
4721 - The client is on a platform where the PRNG has not been seeded
4722 automatically, and the user has not seeded manually
4723 - A protocol specific client method version has been used (i.e. not
4724 SSL_client_methodv23)
4725 - A ciphersuite is used that does not require additional random data from
4726 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4728 If the handshake succeeds then the client random that has been used will
4729 have been generated from a PRNG with insufficient entropy and therefore the
4730 output may be predictable.
4732 For example using the following command with an unseeded openssl will
4733 succeed on an unpatched platform:
4735 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4740 * Use After Free following d2i_ECPrivatekey error fix
4742 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4743 could cause a use after free condition. This, in turn, could cause a double
4744 free in several private key parsing functions (such as d2i_PrivateKey
4745 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4746 for applications that receive EC private keys from untrusted
4747 sources. This scenario is considered rare.
4749 This issue was discovered by the BoringSSL project and fixed in their
4755 * X509_to_X509_REQ NULL pointer deref fix
4757 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4758 the certificate key is invalid. This function is rarely used in practice.
4760 This issue was discovered by Brian Carpenter.
4765 * Removed the export ciphers from the DEFAULT ciphers
4769 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015] ###
4771 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4772 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4773 So far those who have to target multiple platforms would compromise
4774 and argue that binary targeting say ARMv5 would still execute on
4775 ARMv8. "Universal" build resolves this compromise by providing
4776 near-optimal performance even on newer platforms.
4780 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4781 (other platforms pending).
4783 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4785 * Add support for the SignedCertificateTimestampList certificate and
4786 OCSP response extensions from RFC6962.
4790 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4791 for corner cases. (Certain input points at infinity could lead to
4792 bogus results, with non-infinity inputs mapped to infinity too.)
4796 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4797 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4798 common cases are optimized and there still is room for further
4799 improvements. Vector Permutation AES for Altivec is also added.
4803 * Add support for little-endian ppc64 Linux target.
4805 *Marcelo Cerri (IBM)*
4807 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4808 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4809 are optimized and there still is room for further improvements.
4810 Both 32- and 64-bit modes are supported.
4812 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4814 * Improved ARMv7 NEON support.
4818 * Support for SPARC Architecture 2011 crypto extensions, first
4819 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4820 SHA256/512, MD5, GHASH and modular exponentiation.
4822 *Andy Polyakov, David Miller*
4824 * Accelerated modular exponentiation for Intel processors, a.k.a.
4827 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4829 * Support for new and upcoming Intel processors, including AVX2,
4830 BMI and SHA ISA extensions. This includes additional "stitched"
4831 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4834 This work was sponsored by Intel Corp.
4838 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4839 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4840 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4844 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4845 this fixes a limitation in previous versions of OpenSSL.
4849 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4850 MGF1 digest and OAEP label.
4854 * Add EVP support for key wrapping algorithms, to avoid problems with
4855 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4856 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4857 algorithms and include tests cases.
4861 * Add functions to allocate and set the fields of an ECDSA_METHOD
4864 *Douglas E. Engert, Steve Henson*
4866 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4867 difference in days and seconds between two tm or ASN1_TIME structures.
4871 * Add -rev test option to s_server to just reverse order of characters
4872 received by client and send back to server. Also prints an abbreviated
4873 summary of the connection parameters.
4877 * New option -brief for s_client and s_server to print out a brief summary
4878 of connection parameters.
4882 * Add callbacks for arbitrary TLS extensions.
4884 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4886 * New option -crl_download in several openssl utilities to download CRLs
4887 from CRLDP extension in certificates.
4891 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4895 * New function X509_CRL_diff to generate a delta CRL from the difference
4896 of two full CRLs. Add support to "crl" utility.
4900 * New functions to set lookup_crls function and to retrieve
4901 X509_STORE from X509_STORE_CTX.
4905 * Print out deprecated issuer and subject unique ID fields in
4910 * Extend OCSP I/O functions so they can be used for simple general purpose
4911 HTTP as well as OCSP. New wrapper function which can be used to download
4912 CRLs using the OCSP API.
4916 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
4920 * SSL_CONF* functions. These provide a common framework for application
4921 configuration using configuration files or command lines.
4925 * SSL/TLS tracing code. This parses out SSL/TLS records using the
4926 message callback and prints the results. Needs compile time option
4927 "enable-ssl-trace". New options to s_client and s_server to enable
4932 * New ctrl and macro to retrieve supported points extensions.
4933 Print out extension in s_server and s_client.
4937 * New functions to retrieve certificate signature and signature
4942 * Add functions to retrieve and manipulate the raw cipherlist sent by a
4947 * New Suite B modes for TLS code. These use and enforce the requirements
4948 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
4949 only use Suite B curves. The Suite B modes can be set by using the
4950 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
4954 * New chain verification flags for Suite B levels of security. Check
4955 algorithms are acceptable when flags are set in X509_verify_cert.
4959 * Make tls1_check_chain return a set of flags indicating checks passed
4960 by a certificate chain. Add additional tests to handle client
4961 certificates: checks for matching certificate type and issuer name
4966 * If an attempt is made to use a signature algorithm not in the peer
4967 preference list abort the handshake. If client has no suitable
4968 signature algorithms in response to a certificate request do not
4969 use the certificate.
4973 * If server EC tmp key is not in client preference list abort handshake.
4977 * Add support for certificate stores in CERT structure. This makes it
4978 possible to have different stores per SSL structure or one store in
4979 the parent SSL_CTX. Include distinct stores for certificate chain
4980 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
4981 to build and store a certificate chain in CERT structure: returning
4982 an error if the chain cannot be built: this will allow applications
4983 to test if a chain is correctly configured.
4985 Note: if the CERT based stores are not set then the parent SSL_CTX
4986 store is used to retain compatibility with existing behaviour.
4991 * New function ssl_set_client_disabled to set a ciphersuite disabled
4992 mask based on the current session, check mask when sending client
4993 hello and checking the requested ciphersuite.
4997 * New ctrls to retrieve and set certificate types in a certificate
4998 request message. Print out received values in s_client. If certificate
4999 types is not set with custom values set sensible values based on
5000 supported signature algorithms.
5004 * Support for distinct client and server supported signature algorithms.
5008 * Add certificate callback. If set this is called whenever a certificate
5009 is required by client or server. An application can decide which
5010 certificate chain to present based on arbitrary criteria: for example
5011 supported signature algorithms. Add very simple example to s_server.
5012 This fixes many of the problems and restrictions of the existing client
5013 certificate callback: for example you can now clear an existing
5014 certificate and specify the whole chain.
5018 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5019 the certificate can be used for (if anything). Set valid_flags field
5020 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5021 to have similar checks in it.
5023 Add new "cert_flags" field to CERT structure and include a "strict mode".
5024 This enforces some TLS certificate requirements (such as only permitting
5025 certificate signature algorithms contained in the supported algorithms
5026 extension) which some implementations ignore: this option should be used
5027 with caution as it could cause interoperability issues.
5031 * Update and tidy signature algorithm extension processing. Work out
5032 shared signature algorithms based on preferences and peer algorithms
5033 and print them out in s_client and s_server. Abort handshake if no
5034 shared signature algorithms.
5038 * Add new functions to allow customised supported signature algorithms
5039 for SSL and SSL_CTX structures. Add options to s_client and s_server
5044 * New function SSL_certs_clear() to delete all references to certificates
5045 from an SSL structure. Before this once a certificate had been added
5046 it couldn't be removed.
5050 * Integrate hostname, email address and IP address checking with certificate
5051 verification. New verify options supporting checking in openssl utility.
5055 * Fixes and wildcard matching support to hostname and email checking
5056 functions. Add manual page.
5058 *Florian Weimer (Red Hat Product Security Team)*
5060 * New functions to check a hostname email or IP address against a
5061 certificate. Add options x509 utility to print results of checks against
5066 * Fix OCSP checking.
5068 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5070 * Initial experimental support for explicitly trusted non-root CAs.
5071 OpenSSL still tries to build a complete chain to a root but if an
5072 intermediate CA has a trust setting included that is used. The first
5073 setting is used: whether to trust (e.g., -addtrust option to the x509
5078 * Add -trusted_first option which attempts to find certificates in the
5079 trusted store even if an untrusted chain is also supplied.
5083 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5084 platform support for Linux and Android.
5088 * Support for linux-x32, ILP32 environment in x86_64 framework.
5092 * Experimental multi-implementation support for FIPS capable OpenSSL.
5093 When in FIPS mode the approved implementations are used as normal,
5094 when not in FIPS mode the internal unapproved versions are used instead.
5095 This means that the FIPS capable OpenSSL isn't forced to use the
5096 (often lower performance) FIPS implementations outside FIPS mode.
5100 * Transparently support X9.42 DH parameters when calling
5101 PEM_read_bio_DHparameters. This means existing applications can handle
5102 the new parameter format automatically.
5106 * Initial experimental support for X9.42 DH parameter format: mainly
5107 to support use of 'q' parameter for RFC5114 parameters.
5111 * Add DH parameters from RFC5114 including test data to dhtest.
5115 * Support for automatic EC temporary key parameter selection. If enabled
5116 the most preferred EC parameters are automatically used instead of
5117 hardcoded fixed parameters. Now a server just has to call:
5118 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5119 support ECDH and use the most appropriate parameters.
5123 * Enhance and tidy EC curve and point format TLS extension code. Use
5124 static structures instead of allocation if default values are used.
5125 New ctrls to set curves we wish to support and to retrieve shared curves.
5126 Print out shared curves in s_server. New options to s_server and s_client
5127 to set list of supported curves.
5131 * New ctrls to retrieve supported signature algorithms and
5132 supported curve values as an array of NIDs. Extend openssl utility
5133 to print out received values.
5137 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5138 between NIDs and the more common NIST names such as "P-256". Enhance
5139 ecparam utility and ECC method to recognise the NIST names for curves.
5143 * Enhance SSL/TLS certificate chain handling to support different
5144 chains for each certificate instead of one chain in the parent SSL_CTX.
5148 * Support for fixed DH ciphersuite client authentication: where both
5149 server and client use DH certificates with common parameters.
5153 * Support for fixed DH ciphersuites: those requiring DH server
5158 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5160 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5161 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5162 X509_CINF_get_signature were reverted post internal team review.
5168 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016] ###
5170 * OCSP Status Request extension unbounded memory growth
5172 A malicious client can send an excessively large OCSP Status Request
5173 extension. If that client continually requests renegotiation, sending a
5174 large OCSP Status Request extension each time, then there will be unbounded
5175 memory growth on the server. This will eventually lead to a Denial Of
5176 Service attack through memory exhaustion. Servers with a default
5177 configuration are vulnerable even if they do not support OCSP. Builds using
5178 the "no-ocsp" build time option are not affected.
5180 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5185 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5188 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5194 * OOB write in MDC2_Update()
5196 An overflow can occur in MDC2_Update() either if called directly or
5197 through the EVP_DigestUpdate() function using MDC2. If an attacker
5198 is able to supply very large amounts of input data after a previous
5199 call to EVP_EncryptUpdate() with a partial block then a length check
5200 can overflow resulting in a heap corruption.
5202 The amount of data needed is comparable to SIZE_MAX which is impractical
5205 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5210 * Malformed SHA512 ticket DoS
5212 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5213 DoS attack where a malformed ticket will result in an OOB read which will
5216 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5217 a custom server callback and ticket lookup mechanism.
5219 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5224 * OOB write in BN_bn2dec()
5226 The function BN_bn2dec() does not check the return value of BN_div_word().
5227 This can cause an OOB write if an application uses this function with an
5228 overly large BIGNUM. This could be a problem if an overly large certificate
5229 or CRL is printed out from an untrusted source. TLS is not affected because
5230 record limits will reject an oversized certificate before it is parsed.
5232 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5237 * OOB read in TS_OBJ_print_bio()
5239 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5240 the total length the OID text representation would use and not the amount
5241 of data written. This will result in OOB reads when large OIDs are
5244 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5249 * Pointer arithmetic undefined behaviour
5251 Avoid some undefined pointer arithmetic
5253 A common idiom in the codebase is to check limits in the following manner:
5256 Where "p" points to some malloc'd data of SIZE bytes and
5259 "len" here could be from some externally supplied data (e.g. from a TLS
5262 The rules of C pointer arithmetic are such that "p + len" is only well
5263 defined where len <= SIZE. Therefore the above idiom is actually
5264 undefined behaviour.
5266 For example this could cause problems if some malloc implementation
5267 provides an address for "p" such that "p + len" actually overflows for
5268 values of len that are too big and therefore p + len < limit.
5270 This issue was reported to OpenSSL by Guido Vranken
5275 * Constant time flag not preserved in DSA signing
5277 Operations in the DSA signing algorithm should run in constant time in
5278 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5279 implementation means that a non-constant time codepath is followed for
5280 certain operations. This has been demonstrated through a cache-timing
5281 attack to be sufficient for an attacker to recover the private DSA key.
5283 This issue was reported by César Pereida (Aalto University), Billy Brumley
5284 (Tampere University of Technology), and Yuval Yarom (The University of
5285 Adelaide and NICTA).
5290 * DTLS buffered message DoS
5292 In a DTLS connection where handshake messages are delivered out-of-order
5293 those messages that OpenSSL is not yet ready to process will be buffered
5294 for later use. Under certain circumstances, a flaw in the logic means that
5295 those messages do not get removed from the buffer even though the handshake
5296 has been completed. An attacker could force up to approx. 15 messages to
5297 remain in the buffer when they are no longer required. These messages will
5298 be cleared when the DTLS connection is closed. The default maximum size for
5299 a message is 100k. Therefore the attacker could force an additional 1500k
5300 to be consumed per connection. By opening many simulataneous connections an
5301 attacker could cause a DoS attack through memory exhaustion.
5303 This issue was reported to OpenSSL by Quan Luo.
5308 * DTLS replay protection DoS
5310 A flaw in the DTLS replay attack protection mechanism means that records
5311 that arrive for future epochs update the replay protection "window" before
5312 the MAC for the record has been validated. This could be exploited by an
5313 attacker by sending a record for the next epoch (which does not have to
5314 decrypt or have a valid MAC), with a very large sequence number. This means
5315 that all subsequent legitimate packets are dropped causing a denial of
5316 service for a specific DTLS connection.
5318 This issue was reported to OpenSSL by the OCAP audit team.
5323 * Certificate message OOB reads
5325 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5326 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5327 theoretical DoS risk but this has not been observed in practice on common
5330 The messages affected are client certificate, client certificate request
5331 and server certificate. As a result the attack can only be performed
5332 against a client or a server which enables client authentication.
5334 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5339 ### Changes between 1.0.1s and 1.0.1t [3 May 2016] ###
5341 * Prevent padding oracle in AES-NI CBC MAC check
5343 A MITM attacker can use a padding oracle attack to decrypt traffic
5344 when the connection uses an AES CBC cipher and the server support
5347 This issue was introduced as part of the fix for Lucky 13 padding
5348 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5349 constant time by making sure that always the same bytes are read and
5350 compared against either the MAC or padding bytes. But it no longer
5351 checked that there was enough data to have both the MAC and padding
5354 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5359 * Fix EVP_EncodeUpdate overflow
5361 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5362 Base64 encoding of binary data. If an attacker is able to supply very large
5363 amounts of input data then a length check can overflow resulting in a heap
5366 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5367 the PEM_write_bio* family of functions. These are mainly used within the
5368 OpenSSL command line applications, so any application which processes data
5369 from an untrusted source and outputs it as a PEM file should be considered
5370 vulnerable to this issue. User applications that call these APIs directly
5371 with large amounts of untrusted data may also be vulnerable.
5373 This issue was reported by Guido Vranken.
5378 * Fix EVP_EncryptUpdate overflow
5380 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5381 is able to supply very large amounts of input data after a previous call to
5382 EVP_EncryptUpdate() with a partial block then a length check can overflow
5383 resulting in a heap corruption. Following an analysis of all OpenSSL
5384 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5385 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5386 the first called function after an EVP_EncryptInit(), and therefore that
5387 specific call must be safe. The second form is where the length passed to
5388 EVP_EncryptUpdate() can be seen from the code to be some small value and
5389 therefore there is no possibility of an overflow. Since all instances are
5390 one of these two forms, it is believed that there can be no overflows in
5391 internal code due to this problem. It should be noted that
5392 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5393 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5394 of these calls have also been analysed too and it is believed there are no
5395 instances in internal usage where an overflow could occur.
5397 This issue was reported by Guido Vranken.
5402 * Prevent ASN.1 BIO excessive memory allocation
5404 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5405 a short invalid encoding can casuse allocation of large amounts of memory
5406 potentially consuming excessive resources or exhausting memory.
5408 Any application parsing untrusted data through d2i BIO functions is
5409 affected. The memory based functions such as d2i_X509() are *not* affected.
5410 Since the memory based functions are used by the TLS library, TLS
5411 applications are not affected.
5413 This issue was reported by Brian Carpenter.
5420 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5421 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5422 in arbitrary stack data being returned in the buffer.
5424 This issue was reported by Guido Vranken.
5429 * Modify behavior of ALPN to invoke callback after SNI/servername
5430 callback, such that updates to the SSL_CTX affect ALPN.
5434 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5439 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5440 methods are enabled and ssl2 is disabled the methods return NULL.
5444 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016] ###
5446 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5447 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5448 provide any "EXPORT" or "LOW" strength ciphers.
5452 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5453 is by default disabled at build-time. Builds that are not configured with
5454 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5455 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5456 will need to explicitly call either of:
5458 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5460 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5462 as appropriate. Even if either of those is used, or the application
5463 explicitly uses the version-specific SSLv2_method() or its client and
5464 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5465 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5466 ciphers, and SSLv2 56-bit DES are no longer available.
5471 * Fix a double-free in DSA code
5473 A double free bug was discovered when OpenSSL parses malformed DSA private
5474 keys and could lead to a DoS attack or memory corruption for applications
5475 that receive DSA private keys from untrusted sources. This scenario is
5478 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5484 * Disable SRP fake user seed to address a server memory leak.
5486 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5488 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5489 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5490 was changed to ignore the "fake user" SRP seed, even if the seed
5493 Users should use SRP_VBASE_get1_by_user instead. Note that in
5494 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5495 also that even though configuring the SRP seed attempts to hide
5496 invalid usernames by continuing the handshake with fake
5497 credentials, this behaviour is not constant time and no strong
5498 guarantees are made that the handshake is indistinguishable from
5499 that of a valid user.
5504 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5506 In the BN_hex2bn function the number of hex digits is calculated using an
5507 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5508 large values of |i| this can result in |bn_expand| not allocating any
5509 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5510 field as NULL leading to a subsequent NULL ptr deref. For very large values
5511 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5512 In this case memory is allocated to the internal BIGNUM data field, but it
5513 is insufficiently sized leading to heap corruption. A similar issue exists
5514 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5515 is ever called by user applications with very large untrusted hex/dec data.
5516 This is anticipated to be a rare occurrence.
5518 All OpenSSL internal usage of these functions use data that is not expected
5519 to be untrusted, e.g. config file data or application command line
5520 arguments. If user developed applications generate config file data based
5521 on untrusted data then it is possible that this could also lead to security
5522 consequences. This is also anticipated to be rare.
5524 This issue was reported to OpenSSL by Guido Vranken.
5529 * Fix memory issues in BIO_*printf functions
5531 The internal |fmtstr| function used in processing a "%s" format string in
5532 the BIO_*printf functions could overflow while calculating the length of a
5533 string and cause an OOB read when printing very long strings.
5535 Additionally the internal |doapr_outch| function can attempt to write to an
5536 OOB memory location (at an offset from the NULL pointer) in the event of a
5537 memory allocation failure. In 1.0.2 and below this could be caused where
5538 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5539 could be in processing a very long "%s" format string. Memory leaks can
5542 The first issue may mask the second issue dependent on compiler behaviour.
5543 These problems could enable attacks where large amounts of untrusted data
5544 is passed to the BIO_*printf functions. If applications use these functions
5545 in this way then they could be vulnerable. OpenSSL itself uses these
5546 functions when printing out human-readable dumps of ASN.1 data. Therefore
5547 applications that print this data could be vulnerable if the data is from
5548 untrusted sources. OpenSSL command line applications could also be
5549 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5550 as command line arguments.
5552 Libssl is not considered directly vulnerable. Additionally certificates etc
5553 received via remote connections via libssl are also unlikely to be able to
5554 trigger these issues because of message size limits enforced within libssl.
5556 This issue was reported to OpenSSL Guido Vranken.
5561 * Side channel attack on modular exponentiation
5563 A side-channel attack was found which makes use of cache-bank conflicts on
5564 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5565 of RSA keys. The ability to exploit this issue is limited as it relies on
5566 an attacker who has control of code in a thread running on the same
5567 hyper-threaded core as the victim thread which is performing decryptions.
5569 This issue was reported to OpenSSL by Yuval Yarom, The University of
5570 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5571 Nadia Heninger, University of Pennsylvania with more information at
5572 http://cachebleed.info.
5577 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5578 if no keysize is specified with default_bits. This fixes an
5579 omission in an earlier change that changed all RSA/DSA key generation
5580 apps to use 2048 bits by default.
5584 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016] ###
5586 * Protection for DH small subgroup attacks
5588 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5589 switched on by default and cannot be disabled. This could have some
5594 * SSLv2 doesn't block disabled ciphers
5596 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5597 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5598 been disabled, provided that the SSLv2 protocol was not also disabled via
5601 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5602 and Sebastian Schinzel.
5607 * Reject DH handshakes with parameters shorter than 1024 bits.
5611 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015] ###
5613 * Certificate verify crash with missing PSS parameter
5615 The signature verification routines will crash with a NULL pointer
5616 dereference if presented with an ASN.1 signature using the RSA PSS
5617 algorithm and absent mask generation function parameter. Since these
5618 routines are used to verify certificate signature algorithms this can be
5619 used to crash any certificate verification operation and exploited in a
5620 DoS attack. Any application which performs certificate verification is
5621 vulnerable including OpenSSL clients and servers which enable client
5624 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5629 * X509_ATTRIBUTE memory leak
5631 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5632 memory. This structure is used by the PKCS#7 and CMS routines so any
5633 application which reads PKCS#7 or CMS data from untrusted sources is
5634 affected. SSL/TLS is not affected.
5636 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5642 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5643 This changes the decoding behaviour for some invalid messages,
5644 though the change is mostly in the more lenient direction, and
5645 legacy behaviour is preserved as much as possible.
5649 * In DSA_generate_parameters_ex, if the provided seed is too short,
5650 use a random seed, as already documented.
5652 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5654 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015] ###
5656 * Alternate chains certificate forgery
5658 During certificate verfification, OpenSSL will attempt to find an
5659 alternative certificate chain if the first attempt to build such a chain
5660 fails. An error in the implementation of this logic can mean that an
5661 attacker could cause certain checks on untrusted certificates to be
5662 bypassed, such as the CA flag, enabling them to use a valid leaf
5663 certificate to act as a CA and "issue" an invalid certificate.
5665 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5671 * Race condition handling PSK identify hint
5673 If PSK identity hints are received by a multi-threaded client then
5674 the values are wrongly updated in the parent SSL_CTX structure. This can
5675 result in a race condition potentially leading to a double free of the
5681 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015] ###
5682 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5683 incompatibility in the handling of HMAC. The previous ABI has now been
5686 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015] ###
5688 * Malformed ECParameters causes infinite loop
5690 When processing an ECParameters structure OpenSSL enters an infinite loop
5691 if the curve specified is over a specially malformed binary polynomial
5694 This can be used to perform denial of service against any
5695 system which processes public keys, certificate requests or
5696 certificates. This includes TLS clients and TLS servers with
5697 client authentication enabled.
5699 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5704 * Exploitable out-of-bounds read in X509_cmp_time
5706 X509_cmp_time does not properly check the length of the ASN1_TIME
5707 string and can read a few bytes out of bounds. In addition,
5708 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5711 An attacker can use this to craft malformed certificates and CRLs of
5712 various sizes and potentially cause a segmentation fault, resulting in
5713 a DoS on applications that verify certificates or CRLs. TLS clients
5714 that verify CRLs are affected. TLS clients and servers with client
5715 authentication enabled may be affected if they use custom verification
5718 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5719 independently by Hanno Böck.
5724 * PKCS7 crash with missing EnvelopedContent
5726 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5727 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5728 with missing content and trigger a NULL pointer dereference on parsing.
5730 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5731 structures from untrusted sources are affected. OpenSSL clients and
5732 servers are not affected.
5734 This issue was reported to OpenSSL by Michal Zalewski (Google).
5739 * CMS verify infinite loop with unknown hash function
5741 When verifying a signedData message the CMS code can enter an infinite loop
5742 if presented with an unknown hash function OID. This can be used to perform
5743 denial of service against any system which verifies signedData messages using
5745 This issue was reported to OpenSSL by Johannes Bauer.
5750 * Race condition handling NewSessionTicket
5752 If a NewSessionTicket is received by a multi-threaded client when attempting to
5753 reuse a previous ticket then a race condition can occur potentially leading to
5754 a double free of the ticket data.
5759 * Reject DH handshakes with parameters shorter than 768 bits.
5761 *Kurt Roeckx and Emilia Kasper*
5763 * dhparam: generate 2048-bit parameters by default.
5765 *Kurt Roeckx and Emilia Kasper*
5767 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015] ###
5769 * Segmentation fault in ASN1_TYPE_cmp fix
5771 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5772 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5773 certificate signature algorithm consistency this can be used to crash any
5774 certificate verification operation and exploited in a DoS attack. Any
5775 application which performs certificate verification is vulnerable including
5776 OpenSSL clients and servers which enable client authentication.
5781 * ASN.1 structure reuse memory corruption fix
5783 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5784 memory corruption via an invalid write. Such reuse is and has been
5785 strongly discouraged and is believed to be rare.
5787 Applications that parse structures containing CHOICE or ANY DEFINED BY
5788 components may be affected. Certificate parsing (d2i_X509 and related
5789 functions) are however not affected. OpenSSL clients and servers are
5795 * PKCS7 NULL pointer dereferences fix
5797 The PKCS#7 parsing code does not handle missing outer ContentInfo
5798 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5799 missing content and trigger a NULL pointer dereference on parsing.
5801 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5802 otherwise parse PKCS#7 structures from untrusted sources are
5803 affected. OpenSSL clients and servers are not affected.
5805 This issue was reported to OpenSSL by Michal Zalewski (Google).
5810 * DoS via reachable assert in SSLv2 servers fix
5812 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5813 servers that both support SSLv2 and enable export cipher suites by sending
5814 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5816 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5817 (OpenSSL development team).
5822 * Use After Free following d2i_ECPrivatekey error fix
5824 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5825 could cause a use after free condition. This, in turn, could cause a double
5826 free in several private key parsing functions (such as d2i_PrivateKey
5827 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5828 for applications that receive EC private keys from untrusted
5829 sources. This scenario is considered rare.
5831 This issue was discovered by the BoringSSL project and fixed in their
5837 * X509_to_X509_REQ NULL pointer deref fix
5839 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5840 the certificate key is invalid. This function is rarely used in practice.
5842 This issue was discovered by Brian Carpenter.
5847 * Removed the export ciphers from the DEFAULT ciphers
5851 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015] ###
5853 * Build fixes for the Windows and OpenVMS platforms
5855 *Matt Caswell and Richard Levitte*
5857 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015] ###
5859 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5860 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5861 dereference. This could lead to a Denial Of Service attack. Thanks to
5862 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5867 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5868 dtls1_buffer_record function under certain conditions. In particular this
5869 could occur if an attacker sent repeated DTLS records with the same
5870 sequence number but for the next epoch. The memory leak could be exploited
5871 by an attacker in a Denial of Service attack through memory exhaustion.
5872 Thanks to Chris Mueller for reporting this issue.
5877 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5878 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5879 method would be set to NULL which could later result in a NULL pointer
5880 dereference. Thanks to Frank Schmirler for reporting this issue.
5885 * Abort handshake if server key exchange message is omitted for ephemeral
5888 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5889 reporting this issue.
5894 * Remove non-export ephemeral RSA code on client and server. This code
5895 violated the TLS standard by allowing the use of temporary RSA keys in
5896 non-export ciphersuites and could be used by a server to effectively
5897 downgrade the RSA key length used to a value smaller than the server
5898 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
5899 INRIA or reporting this issue.
5904 * Fixed issue where DH client certificates are accepted without verification.
5905 An OpenSSL server will accept a DH certificate for client authentication
5906 without the certificate verify message. This effectively allows a client to
5907 authenticate without the use of a private key. This only affects servers
5908 which trust a client certificate authority which issues certificates
5909 containing DH keys: these are extremely rare and hardly ever encountered.
5910 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
5916 * Ensure that the session ID context of an SSL is updated when its
5917 SSL_CTX is updated via SSL_set_SSL_CTX.
5919 The session ID context is typically set from the parent SSL_CTX,
5920 and can vary with the CTX.
5924 * Fix various certificate fingerprint issues.
5926 By using non-DER or invalid encodings outside the signed portion of a
5927 certificate the fingerprint can be changed without breaking the signature.
5928 Although no details of the signed portion of the certificate can be changed
5929 this can cause problems with some applications: e.g. those using the
5930 certificate fingerprint for blacklists.
5932 1. Reject signatures with non zero unused bits.
5934 If the BIT STRING containing the signature has non zero unused bits reject
5935 the signature. All current signature algorithms require zero unused bits.
5937 2. Check certificate algorithm consistency.
5939 Check the AlgorithmIdentifier inside TBS matches the one in the
5940 certificate signature. NB: this will result in signature failure
5941 errors for some broken certificates.
5943 Thanks to Konrad Kraszewski from Google for reporting this issue.
5945 3. Check DSA/ECDSA signatures use DER.
5947 Re-encode DSA/ECDSA signatures and compare with the original received
5948 signature. Return an error if there is a mismatch.
5950 This will reject various cases including garbage after signature
5951 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
5952 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
5953 (negative or with leading zeroes).
5955 Further analysis was conducted and fixes were developed by Stephen Henson
5956 of the OpenSSL core team.
5962 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
5963 results on some platforms, including x86_64. This bug occurs at random
5964 with a very low probability, and is not known to be exploitable in any
5965 way, though its exact impact is difficult to determine. Thanks to Pieter
5966 Wuille (Blockstream) who reported this issue and also suggested an initial
5967 fix. Further analysis was conducted by the OpenSSL development team and
5968 Adam Langley of Google. The final fix was developed by Andy Polyakov of
5969 the OpenSSL core team.
5974 *) Do not resume sessions on the server if the negotiated protocol
5975 version does not match the session's version. Resuming with a different
5976 version, while not strictly forbidden by the RFC, is of questionable
5977 sanity and breaks all known clients.
5979 *David Benjamin, Emilia Käsper*
5981 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
5982 early CCS messages during renegotiation. (Note that because
5983 renegotiation is encrypted, this early CCS was not exploitable.)
5987 *) Tighten client-side session ticket handling during renegotiation:
5988 ensure that the client only accepts a session ticket if the server sends
5989 the extension anew in the ServerHello. Previously, a TLS client would
5990 reuse the old extension state and thus accept a session ticket if one was
5991 announced in the initial ServerHello.
5993 Similarly, ensure that the client requires a session ticket if one
5994 was advertised in the ServerHello. Previously, a TLS client would
5995 ignore a missing NewSessionTicket message.
5999 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014] ###
6003 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6004 sends a carefully crafted handshake message, to cause OpenSSL to fail
6005 to free up to 64k of memory causing a memory leak. This could be
6006 exploited in a Denial Of Service attack. This issue affects OpenSSL
6007 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6008 whether SRTP is used or configured. Implementations of OpenSSL that
6009 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6011 The fix was developed by the OpenSSL team.
6016 * Session Ticket Memory Leak.
6018 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6019 integrity of that ticket is first verified. In the event of a session
6020 ticket integrity check failing, OpenSSL will fail to free memory
6021 causing a memory leak. By sending a large number of invalid session
6022 tickets an attacker could exploit this issue in a Denial Of Service
6028 * Build option no-ssl3 is incomplete.
6030 When OpenSSL is configured with "no-ssl3" as a build option, servers
6031 could accept and complete a SSL 3.0 handshake, and clients could be
6032 configured to send them.
6035 *Akamai and the OpenSSL team*
6037 * Add support for TLS_FALLBACK_SCSV.
6038 Client applications doing fallback retries should call
6039 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6042 *Adam Langley, Bodo Moeller*
6044 * Add additional DigestInfo checks.
6046 Re-encode DigestInto in DER and check against the original when
6047 verifying RSA signature: this will reject any improperly encoded
6048 DigestInfo structures.
6050 Note: this is a precautionary measure and no attacks are currently known.
6055 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014] ###
6057 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6058 SRP code can be overrun an internal buffer. Add sanity check that
6059 g, A, B < N to SRP code.
6061 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6062 Group for discovering this issue.
6067 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6068 TLS 1.0 instead of higher protocol versions when the ClientHello message
6069 is badly fragmented. This allows a man-in-the-middle attacker to force a
6070 downgrade to TLS 1.0 even if both the server and the client support a
6071 higher protocol version, by modifying the client's TLS records.
6073 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6074 researching this issue.
6079 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6080 to a denial of service attack. A malicious server can crash the client
6081 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6082 ciphersuite and sending carefully crafted handshake messages.
6084 Thanks to Felix Gröbert (Google) for discovering and researching this
6090 * By sending carefully crafted DTLS packets an attacker could cause openssl
6091 to leak memory. This can be exploited through a Denial of Service attack.
6092 Thanks to Adam Langley for discovering and researching this issue.
6097 * An attacker can force openssl to consume large amounts of memory whilst
6098 processing DTLS handshake messages. This can be exploited through a
6099 Denial of Service attack.
6100 Thanks to Adam Langley for discovering and researching this issue.
6105 * An attacker can force an error condition which causes openssl to crash
6106 whilst processing DTLS packets due to memory being freed twice. This
6107 can be exploited through a Denial of Service attack.
6108 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6114 * If a multithreaded client connects to a malicious server using a resumed
6115 session and the server sends an ec point format extension it could write
6116 up to 255 bytes to freed memory.
6118 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6124 * A malicious server can crash an OpenSSL client with a null pointer
6125 dereference (read) by specifying an SRP ciphersuite even though it was not
6126 properly negotiated with the client. This can be exploited through a
6127 Denial of Service attack.
6129 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6130 discovering and researching this issue.
6135 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6136 X509_name_oneline, X509_name_print_ex et al. to leak some information
6137 from the stack. Applications may be affected if they echo pretty printing
6138 output to the attacker.
6140 Thanks to Ivan Fratric (Google) for discovering this issue.
6143 *Emilia Käsper, and Steve Henson*
6145 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6146 for corner cases. (Certain input points at infinity could lead to
6147 bogus results, with non-infinity inputs mapped to infinity too.)
6151 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014] ###
6153 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6154 handshake can force the use of weak keying material in OpenSSL
6155 SSL/TLS clients and servers.
6157 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6158 researching this issue. [CVE-2014-0224][]
6160 *KIKUCHI Masashi, Steve Henson*
6162 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6163 OpenSSL DTLS client the code can be made to recurse eventually crashing
6166 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6169 *Imre Rad, Steve Henson*
6171 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6172 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6173 client or server. This is potentially exploitable to run arbitrary
6174 code on a vulnerable client or server.
6176 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6178 *Jüri Aedla, Steve Henson*
6180 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6181 are subject to a denial of service attack.
6183 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6184 this issue. [CVE-2014-3470][]
6186 *Felix Gröbert, Ivan Fratric, Steve Henson*
6188 * Harmonize version and its documentation. -f flag is used to display
6191 *mancha <mancha1@zoho.com>*
6193 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6194 in i2d_ECPrivateKey.
6196 *mancha <mancha1@zoho.com>*
6198 * Fix some double frees. These are not thought to be exploitable.
6200 *mancha <mancha1@zoho.com>*
6202 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014] ###
6204 * A missing bounds check in the handling of the TLS heartbeat extension
6205 can be used to reveal up to 64k of memory to a connected client or
6208 Thanks for Neel Mehta of Google Security for discovering this bug and to
6209 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6210 preparing the fix [CVE-2014-0160][]
6212 *Adam Langley, Bodo Moeller*
6214 * Fix for the attack described in the paper "Recovering OpenSSL
6215 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6216 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6217 http://eprint.iacr.org/2014/140
6219 Thanks to Yuval Yarom and Naomi Benger for discovering this
6220 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6222 *Yuval Yarom and Naomi Benger*
6224 * TLS pad extension: draft-agl-tls-padding-03
6226 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6227 TLS client Hello record length value would otherwise be > 255 and
6228 less that 512 pad with a dummy extension containing zeroes so it
6229 is at least 512 bytes long.
6232 *Adam Langley, Steve Henson*
6234 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014] ###
6236 * Fix for TLS record tampering bug. A carefully crafted invalid
6237 handshake could crash OpenSSL with a NULL pointer exception.
6238 Thanks to Anton Johansson for reporting this issues.
6241 * Keep original DTLS digest and encryption contexts in retransmission
6242 structures so we can use the previous session parameters if they need
6243 to be resent. [CVE-2013-6450][]
6247 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6248 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6249 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6250 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6251 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6252 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6254 *Rob Stradling, Adam Langley*
6256 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013] ###
6258 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6259 supporting platforms or when small records were transferred.
6261 *Andy Polyakov, Steve Henson*
6263 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013] ###
6265 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6267 This addresses the flaw in CBC record processing discovered by
6268 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6269 at: http://www.isg.rhul.ac.uk/tls/
6271 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6272 Security Group at Royal Holloway, University of London
6273 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6274 Emilia Käsper for the initial patch.
6277 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6279 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6280 ciphersuites which can be exploited in a denial of service attack.
6281 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6282 and detecting this bug and to Wolfgang Ettlinger
6283 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6288 * Return an error when checking OCSP signatures when key is NULL.
6289 This fixes a DoS attack. [CVE-2013-0166][]
6293 * Make openssl verify return errors.
6295 *Chris Palmer <palmer@google.com> and Ben Laurie*
6297 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6298 the right response is stapled. Also change SSL_get_certificate()
6299 so it returns the certificate actually sent.
6300 See http://rt.openssl.org/Ticket/Display.html?id=2836.
6302 *Rob Stradling <rob.stradling@comodo.com>*
6304 * Fix possible deadlock when decoding public keys.
6308 * Don't use TLS 1.0 record version number in initial client hello
6313 ### Changes between 1.0.1b and 1.0.1c [10 May 2012] ###
6315 * Sanity check record length before skipping explicit IV in TLS
6316 1.2, 1.1 and DTLS to fix DoS attack.
6318 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6319 fuzzing as a service testing platform.
6324 * Initialise tkeylen properly when encrypting CMS messages.
6325 Thanks to Solar Designer of Openwall for reporting this issue.
6329 * In FIPS mode don't try to use composite ciphers as they are not
6334 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012] ###
6336 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6337 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6338 mean any application compiled against OpenSSL 1.0.0 headers setting
6339 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6340 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6341 0x10000000L Any application which was previously compiled against
6342 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6343 will need to be recompiled as a result. Letting be results in
6344 inability to disable specifically TLS 1.1 and in client context,
6345 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6349 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6350 disable just protocol X, but all protocols above X *if* there are
6351 protocols *below* X still enabled. In more practical terms it means
6352 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6353 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6354 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6359 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012] ###
6361 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6362 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6363 in CRYPTO_realloc_clean.
6365 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6366 issue and to Adam Langley <agl@chromium.org> for fixing it.
6369 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6371 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6375 * Workarounds for some broken servers that "hang" if a client hello
6376 record length exceeds 255 bytes.
6378 1. Do not use record version number > TLS 1.0 in initial client
6379 hello: some (but not all) hanging servers will now work.
6380 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6381 the number of ciphers sent in the client hello. This should be
6382 set to an even number, such as 50, for example by passing:
6383 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6384 Most broken servers should now work.
6385 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6386 TLS 1.2 client support entirely.
6390 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6394 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012] ###
6396 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6397 STRING form instead of a DigestInfo.
6401 * The format used for MDC2 RSA signatures is inconsistent between EVP
6402 and the RSA_sign/RSA_verify functions. This was made more apparent when
6403 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6404 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6405 the correct format in RSA_verify so both forms transparently work.
6409 * Some servers which support TLS 1.0 can choke if we initially indicate
6410 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6411 encrypted premaster secret. As a workaround use the maximum permitted
6412 client version in client hello, this should keep such servers happy
6413 and still work with previous versions of OpenSSL.
6417 * Add support for TLS/DTLS heartbeats.
6419 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6421 * Add support for SCTP.
6423 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6425 * Improved PRNG seeding for VOS.
6427 *Paul Green <Paul.Green@stratus.com>*
6429 * Extensive assembler packs updates, most notably:
6431 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6432 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6433 - x86_64: bit-sliced AES implementation;
6434 - ARM: NEON support, contemporary platforms optimizations;
6435 - s390x: z196 support;
6436 - *: GHASH and GF(2^m) multiplication implementations;
6440 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6441 (removal of unnecessary code)
6443 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6445 * Add TLS key material exporter from RFC 5705.
6449 * Add DTLS-SRTP negotiation from RFC 5764.
6453 * Add Next Protocol Negotiation,
6454 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
6455 disabled with a no-npn flag to config or Configure. Code donated
6458 *Adam Langley <agl@google.com> and Ben Laurie*
6460 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6461 NIST-P256, NIST-P521, with constant-time single point multiplication on
6462 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6463 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6464 Code made available under Apache License version 2.0.
6466 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6467 line to include this in your build of OpenSSL, and run "make depend" (or
6468 "make update"). This enables the following EC_METHODs:
6470 EC_GFp_nistp224_method()
6471 EC_GFp_nistp256_method()
6472 EC_GFp_nistp521_method()
6474 EC_GROUP_new_by_curve_name() will automatically use these (while
6475 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6478 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6480 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6481 all platforms. Move ssize_t definition from e_os.h to the public
6482 header file e_os2.h as it now appears in public header file cms.h
6486 * New -sigopt option to the ca, req and x509 utilities. Additional
6487 signature parameters can be passed using this option and in
6492 * Add RSA PSS signing function. This will generate and set the
6493 appropriate AlgorithmIdentifiers for PSS based on those in the
6494 corresponding EVP_MD_CTX structure. No application support yet.
6498 * Support for companion algorithm specific ASN1 signing routines.
6499 New function ASN1_item_sign_ctx() signs a pre-initialised
6500 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6501 the appropriate parameters.
6505 * Add new algorithm specific ASN1 verification initialisation function
6506 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6507 handling will be the same no matter what EVP_PKEY_METHOD is used.
6508 Add a PSS handler to support verification of PSS signatures: checked
6509 against a number of sample certificates.
6513 * Add signature printing for PSS. Add PSS OIDs.
6515 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6517 * Add algorithm specific signature printing. An individual ASN1 method
6518 can now print out signatures instead of the standard hex dump.
6520 More complex signatures (e.g. PSS) can print out more meaningful
6521 information. Include DSA version that prints out the signature
6526 * Password based recipient info support for CMS library: implementing
6531 * Split password based encryption into PBES2 and PBKDF2 functions. This
6532 neatly separates the code into cipher and PBE sections and is required
6533 for some algorithms that split PBES2 into separate pieces (such as
6534 password based CMS).
6538 * Session-handling fixes:
6539 - Fix handling of connections that are resuming with a session ID,
6540 but also support Session Tickets.
6541 - Fix a bug that suppressed issuing of a new ticket if the client
6542 presented a ticket with an expired session.
6543 - Try to set the ticket lifetime hint to something reasonable.
6544 - Make tickets shorter by excluding irrelevant information.
6545 - On the client side, don't ignore renewed tickets.
6547 *Adam Langley, Bodo Moeller (Google)*
6549 * Fix PSK session representation.
6553 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6555 This work was sponsored by Intel.
6559 * Add GCM support to TLS library. Some custom code is needed to split
6560 the IV between the fixed (from PRF) and explicit (from TLS record)
6561 portions. This adds all GCM ciphersuites supported by RFC5288 and
6562 RFC5289. Generalise some AES* cipherstrings to include GCM and
6563 add a special AESGCM string for GCM only.
6567 * Expand range of ctrls for AES GCM. Permit setting invocation
6568 field on decrypt and retrieval of invocation field only on encrypt.
6572 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6573 As required by RFC5289 these ciphersuites cannot be used if for
6574 versions of TLS earlier than 1.2.
6578 * For FIPS capable OpenSSL interpret a NULL default public key method
6579 as unset and return the appropriate default but do *not* set the default.
6580 This means we can return the appropriate method in applications that
6581 switch between FIPS and non-FIPS modes.
6585 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6586 ENGINE is used then we cannot handle that in the FIPS module so we
6587 keep original code iff non-FIPS operations are allowed.
6591 * Add -attime option to openssl utilities.
6593 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6595 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6599 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6600 FIPS EC methods unconditionally for now.
6604 * New build option no-ec2m to disable characteristic 2 code.
6608 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6609 all cases can be covered as some introduce binary incompatibilities.
6613 * Redirect RSA operations to FIPS module including keygen,
6614 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6618 * Add similar low level API blocking to ciphers.
6622 * Low level digest APIs are not approved in FIPS mode: any attempt
6623 to use these will cause a fatal error. Applications that *really* want
6624 to use them can use the private_* version instead.
6628 * Redirect cipher operations to FIPS module for FIPS builds.
6632 * Redirect digest operations to FIPS module for FIPS builds.
6636 * Update build system to add "fips" flag which will link in fipscanister.o
6637 for static and shared library builds embedding a signature if needed.
6641 * Output TLS supported curves in preference order instead of numerical
6642 order. This is currently hardcoded for the highest order curves first.
6643 This should be configurable so applications can judge speed vs strength.
6647 * Add TLS v1.2 server support for client authentication.
6651 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6656 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6657 FIPS modules versions.
6661 * Add TLS v1.2 client side support for client authentication. Keep cache
6662 of handshake records longer as we don't know the hash algorithm to use
6663 until after the certificate request message is received.
6667 * Initial TLS v1.2 client support. Add a default signature algorithms
6668 extension including all the algorithms we support. Parse new signature
6669 format in client key exchange. Relax some ECC signing restrictions for
6670 TLS v1.2 as indicated in RFC5246.
6674 * Add server support for TLS v1.2 signature algorithms extension. Switch
6675 to new signature format when needed using client digest preference.
6676 All server ciphersuites should now work correctly in TLS v1.2. No client
6677 support yet and no support for client certificates.
6681 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6682 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6683 ciphersuites. At present only RSA key exchange ciphersuites work with
6684 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6685 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6686 and version checking.
6690 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6691 with this defined it will not be affected by any changes to ssl internal
6692 structures. Add several utility functions to allow openssl application
6693 to work with OPENSSL_NO_SSL_INTERN defined.
6697 * A long standing patch to add support for SRP from EdelWeb (Peter
6698 Sylvester and Christophe Renou) was integrated.
6699 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6700 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6703 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6707 * Permit abbreviated handshakes when renegotiating using the function
6708 SSL_renegotiate_abbreviated().
6710 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6712 * Add call to ENGINE_register_all_complete() to
6713 ENGINE_load_builtin_engines(), so some implementations get used
6714 automatically instead of needing explicit application support.
6718 * Add support for TLS key exporter as described in RFC5705.
6720 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6722 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6723 a few changes are required:
6725 Add SSL_OP_NO_TLSv1_1 flag.
6726 Add TLSv1_1 methods.
6727 Update version checking logic to handle version 1.1.
6728 Add explicit IV handling (ported from DTLS code).
6729 Add command line options to s_client/s_server.
6736 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015] ###
6738 * X509_ATTRIBUTE memory leak
6740 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6741 memory. This structure is used by the PKCS#7 and CMS routines so any
6742 application which reads PKCS#7 or CMS data from untrusted sources is
6743 affected. SSL/TLS is not affected.
6745 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6751 * Race condition handling PSK identify hint
6753 If PSK identity hints are received by a multi-threaded client then
6754 the values are wrongly updated in the parent SSL_CTX structure. This can
6755 result in a race condition potentially leading to a double free of the
6761 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015] ###
6763 * Malformed ECParameters causes infinite loop
6765 When processing an ECParameters structure OpenSSL enters an infinite loop
6766 if the curve specified is over a specially malformed binary polynomial
6769 This can be used to perform denial of service against any
6770 system which processes public keys, certificate requests or
6771 certificates. This includes TLS clients and TLS servers with
6772 client authentication enabled.
6774 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6779 * Exploitable out-of-bounds read in X509_cmp_time
6781 X509_cmp_time does not properly check the length of the ASN1_TIME
6782 string and can read a few bytes out of bounds. In addition,
6783 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6786 An attacker can use this to craft malformed certificates and CRLs of
6787 various sizes and potentially cause a segmentation fault, resulting in
6788 a DoS on applications that verify certificates or CRLs. TLS clients
6789 that verify CRLs are affected. TLS clients and servers with client
6790 authentication enabled may be affected if they use custom verification
6793 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6794 independently by Hanno Böck.
6799 * PKCS7 crash with missing EnvelopedContent
6801 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6802 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6803 with missing content and trigger a NULL pointer dereference on parsing.
6805 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6806 structures from untrusted sources are affected. OpenSSL clients and
6807 servers are not affected.
6809 This issue was reported to OpenSSL by Michal Zalewski (Google).
6814 * CMS verify infinite loop with unknown hash function
6816 When verifying a signedData message the CMS code can enter an infinite loop
6817 if presented with an unknown hash function OID. This can be used to perform
6818 denial of service against any system which verifies signedData messages using
6820 This issue was reported to OpenSSL by Johannes Bauer.
6825 * Race condition handling NewSessionTicket
6827 If a NewSessionTicket is received by a multi-threaded client when attempting to
6828 reuse a previous ticket then a race condition can occur potentially leading to
6829 a double free of the ticket data.
6834 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015] ###
6836 * Segmentation fault in ASN1_TYPE_cmp fix
6838 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6839 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6840 certificate signature algorithm consistency this can be used to crash any
6841 certificate verification operation and exploited in a DoS attack. Any
6842 application which performs certificate verification is vulnerable including
6843 OpenSSL clients and servers which enable client authentication.
6848 * ASN.1 structure reuse memory corruption fix
6850 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6851 memory corruption via an invalid write. Such reuse is and has been
6852 strongly discouraged and is believed to be rare.
6854 Applications that parse structures containing CHOICE or ANY DEFINED BY
6855 components may be affected. Certificate parsing (d2i_X509 and related
6856 functions) are however not affected. OpenSSL clients and servers are
6862 * PKCS7 NULL pointer dereferences fix
6864 The PKCS#7 parsing code does not handle missing outer ContentInfo
6865 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6866 missing content and trigger a NULL pointer dereference on parsing.
6868 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6869 otherwise parse PKCS#7 structures from untrusted sources are
6870 affected. OpenSSL clients and servers are not affected.
6872 This issue was reported to OpenSSL by Michal Zalewski (Google).
6877 * DoS via reachable assert in SSLv2 servers fix
6879 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6880 servers that both support SSLv2 and enable export cipher suites by sending
6881 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6883 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6884 (OpenSSL development team).
6889 * Use After Free following d2i_ECPrivatekey error fix
6891 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6892 could cause a use after free condition. This, in turn, could cause a double
6893 free in several private key parsing functions (such as d2i_PrivateKey
6894 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6895 for applications that receive EC private keys from untrusted
6896 sources. This scenario is considered rare.
6898 This issue was discovered by the BoringSSL project and fixed in their
6904 * X509_to_X509_REQ NULL pointer deref fix
6906 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6907 the certificate key is invalid. This function is rarely used in practice.
6909 This issue was discovered by Brian Carpenter.
6914 * Removed the export ciphers from the DEFAULT ciphers
6918 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015] ###
6920 * Build fixes for the Windows and OpenVMS platforms
6922 *Matt Caswell and Richard Levitte*
6924 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015] ###
6926 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6927 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6928 dereference. This could lead to a Denial Of Service attack. Thanks to
6929 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6934 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6935 dtls1_buffer_record function under certain conditions. In particular this
6936 could occur if an attacker sent repeated DTLS records with the same
6937 sequence number but for the next epoch. The memory leak could be exploited
6938 by an attacker in a Denial of Service attack through memory exhaustion.
6939 Thanks to Chris Mueller for reporting this issue.
6944 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6945 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6946 method would be set to NULL which could later result in a NULL pointer
6947 dereference. Thanks to Frank Schmirler for reporting this issue.
6952 * Abort handshake if server key exchange message is omitted for ephemeral
6955 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6956 reporting this issue.
6961 * Remove non-export ephemeral RSA code on client and server. This code
6962 violated the TLS standard by allowing the use of temporary RSA keys in
6963 non-export ciphersuites and could be used by a server to effectively
6964 downgrade the RSA key length used to a value smaller than the server
6965 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6966 INRIA or reporting this issue.
6971 * Fixed issue where DH client certificates are accepted without verification.
6972 An OpenSSL server will accept a DH certificate for client authentication
6973 without the certificate verify message. This effectively allows a client to
6974 authenticate without the use of a private key. This only affects servers
6975 which trust a client certificate authority which issues certificates
6976 containing DH keys: these are extremely rare and hardly ever encountered.
6977 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6983 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6984 results on some platforms, including x86_64. This bug occurs at random
6985 with a very low probability, and is not known to be exploitable in any
6986 way, though its exact impact is difficult to determine. Thanks to Pieter
6987 Wuille (Blockstream) who reported this issue and also suggested an initial
6988 fix. Further analysis was conducted by the OpenSSL development team and
6989 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6990 the OpenSSL core team.
6995 * Fix various certificate fingerprint issues.
6997 By using non-DER or invalid encodings outside the signed portion of a
6998 certificate the fingerprint can be changed without breaking the signature.
6999 Although no details of the signed portion of the certificate can be changed
7000 this can cause problems with some applications: e.g. those using the
7001 certificate fingerprint for blacklists.
7003 1. Reject signatures with non zero unused bits.
7005 If the BIT STRING containing the signature has non zero unused bits reject
7006 the signature. All current signature algorithms require zero unused bits.
7008 2. Check certificate algorithm consistency.
7010 Check the AlgorithmIdentifier inside TBS matches the one in the
7011 certificate signature. NB: this will result in signature failure
7012 errors for some broken certificates.
7014 Thanks to Konrad Kraszewski from Google for reporting this issue.
7016 3. Check DSA/ECDSA signatures use DER.
7018 Reencode DSA/ECDSA signatures and compare with the original received
7019 signature. Return an error if there is a mismatch.
7021 This will reject various cases including garbage after signature
7022 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7023 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7024 (negative or with leading zeroes).
7026 Further analysis was conducted and fixes were developed by Stephen Henson
7027 of the OpenSSL core team.
7033 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014] ###
7035 * Session Ticket Memory Leak.
7037 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7038 integrity of that ticket is first verified. In the event of a session
7039 ticket integrity check failing, OpenSSL will fail to free memory
7040 causing a memory leak. By sending a large number of invalid session
7041 tickets an attacker could exploit this issue in a Denial Of Service
7047 * Build option no-ssl3 is incomplete.
7049 When OpenSSL is configured with "no-ssl3" as a build option, servers
7050 could accept and complete a SSL 3.0 handshake, and clients could be
7051 configured to send them.
7054 *Akamai and the OpenSSL team*
7056 * Add support for TLS_FALLBACK_SCSV.
7057 Client applications doing fallback retries should call
7058 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7061 *Adam Langley, Bodo Moeller*
7063 * Add additional DigestInfo checks.
7065 Reencode DigestInto in DER and check against the original when
7066 verifying RSA signature: this will reject any improperly encoded
7067 DigestInfo structures.
7069 Note: this is a precautionary measure and no attacks are currently known.
7074 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014] ###
7076 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7077 to a denial of service attack. A malicious server can crash the client
7078 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7079 ciphersuite and sending carefully crafted handshake messages.
7081 Thanks to Felix Gröbert (Google) for discovering and researching this
7087 * By sending carefully crafted DTLS packets an attacker could cause openssl
7088 to leak memory. This can be exploited through a Denial of Service attack.
7089 Thanks to Adam Langley for discovering and researching this issue.
7094 * An attacker can force openssl to consume large amounts of memory whilst
7095 processing DTLS handshake messages. This can be exploited through a
7096 Denial of Service attack.
7097 Thanks to Adam Langley for discovering and researching this issue.
7102 * An attacker can force an error condition which causes openssl to crash
7103 whilst processing DTLS packets due to memory being freed twice. This
7104 can be exploited through a Denial of Service attack.
7105 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7111 * If a multithreaded client connects to a malicious server using a resumed
7112 session and the server sends an ec point format extension it could write
7113 up to 255 bytes to freed memory.
7115 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7121 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7122 X509_name_oneline, X509_name_print_ex et al. to leak some information
7123 from the stack. Applications may be affected if they echo pretty printing
7124 output to the attacker.
7126 Thanks to Ivan Fratric (Google) for discovering this issue.
7129 *Emilia Käsper, and Steve Henson*
7131 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7132 for corner cases. (Certain input points at infinity could lead to
7133 bogus results, with non-infinity inputs mapped to infinity too.)
7137 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014] ###
7139 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7140 handshake can force the use of weak keying material in OpenSSL
7141 SSL/TLS clients and servers.
7143 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7144 researching this issue. [CVE-2014-0224][]
7146 *KIKUCHI Masashi, Steve Henson*
7148 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7149 OpenSSL DTLS client the code can be made to recurse eventually crashing
7152 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7155 *Imre Rad, Steve Henson*
7157 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7158 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7159 client or server. This is potentially exploitable to run arbitrary
7160 code on a vulnerable client or server.
7162 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7164 *Jüri Aedla, Steve Henson*
7166 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7167 are subject to a denial of service attack.
7169 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7170 this issue. [CVE-2014-3470][]
7172 *Felix Gröbert, Ivan Fratric, Steve Henson*
7174 * Harmonize version and its documentation. -f flag is used to display
7177 *mancha <mancha1@zoho.com>*
7179 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7180 in i2d_ECPrivateKey.
7182 *mancha <mancha1@zoho.com>*
7184 * Fix some double frees. These are not thought to be exploitable.
7186 *mancha <mancha1@zoho.com>*
7188 * Fix for the attack described in the paper "Recovering OpenSSL
7189 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7190 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7191 http://eprint.iacr.org/2014/140
7193 Thanks to Yuval Yarom and Naomi Benger for discovering this
7194 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7196 *Yuval Yarom and Naomi Benger*
7198 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014] ###
7200 * Keep original DTLS digest and encryption contexts in retransmission
7201 structures so we can use the previous session parameters if they need
7202 to be resent. [CVE-2013-6450][]
7206 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7207 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7208 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7209 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7210 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7211 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7213 *Rob Stradling, Adam Langley*
7215 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013] ###
7217 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7219 This addresses the flaw in CBC record processing discovered by
7220 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7221 at: http://www.isg.rhul.ac.uk/tls/
7223 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7224 Security Group at Royal Holloway, University of London
7225 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7226 Emilia Käsper for the initial patch.
7229 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7231 * Return an error when checking OCSP signatures when key is NULL.
7232 This fixes a DoS attack. [CVE-2013-0166][]
7236 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7237 the right response is stapled. Also change SSL_get_certificate()
7238 so it returns the certificate actually sent.
7239 See http://rt.openssl.org/Ticket/Display.html?id=2836.
7240 (This is a backport)
7242 *Rob Stradling <rob.stradling@comodo.com>*
7244 * Fix possible deadlock when decoding public keys.
7248 ### Changes between 1.0.0i and 1.0.0j [10 May 2012] ###
7250 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7253 * Sanity check record length before skipping explicit IV in DTLS
7256 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7257 fuzzing as a service testing platform.
7262 * Initialise tkeylen properly when encrypting CMS messages.
7263 Thanks to Solar Designer of Openwall for reporting this issue.
7267 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012] ###
7269 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7270 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7271 in CRYPTO_realloc_clean.
7273 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7274 issue and to Adam Langley <agl@chromium.org> for fixing it.
7277 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7279 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012] ###
7281 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7282 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7283 content decryption and always return the same error. Note: this attack
7284 needs on average 2^20 messages so it only affects automated senders. The
7285 old behaviour can be re-enabled in the CMS code by setting the
7286 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7287 an MMA defence is not necessary.
7288 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7289 this issue. [CVE-2012-0884][]
7293 * Fix CVE-2011-4619: make sure we really are receiving a
7294 client hello before rejecting multiple SGC restarts. Thanks to
7295 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7299 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012] ###
7301 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7302 Thanks to Antonio Martin, Enterprise Secure Access Research and
7303 Development, Cisco Systems, Inc. for discovering this bug and
7304 preparing a fix. [CVE-2012-0050][]
7308 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012] ###
7310 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7311 of the Vaudenay padding oracle attack on CBC mode encryption
7312 which enables an efficient plaintext recovery attack against
7313 the OpenSSL implementation of DTLS. Their attack exploits timing
7314 differences arising during decryption processing. A research
7315 paper describing this attack can be found at:
7316 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
7317 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7318 Security Group at Royal Holloway, University of London
7319 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7320 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7321 for preparing the fix. [CVE-2011-4108][]
7323 *Robin Seggelmann, Michael Tuexen*
7325 * Clear bytes used for block padding of SSL 3.0 records.
7328 *Adam Langley (Google)*
7330 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7331 Kadianakis <desnacked@gmail.com> for discovering this issue and
7332 Adam Langley for preparing the fix. [CVE-2011-4619][]
7334 *Adam Langley (Google)*
7336 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7338 *Andrey Kulikov <amdeich@gmail.com>*
7340 * Prevent malformed RFC3779 data triggering an assertion failure.
7341 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7342 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7344 *Rob Austein <sra@hactrn.net>*
7346 * Improved PRNG seeding for VOS.
7348 *Paul Green <Paul.Green@stratus.com>*
7350 * Fix ssl_ciph.c set-up race.
7352 *Adam Langley (Google)*
7354 * Fix spurious failures in ecdsatest.c.
7356 *Emilia Käsper (Google)*
7358 * Fix the BIO_f_buffer() implementation (which was mixing different
7359 interpretations of the '..._len' fields).
7361 *Adam Langley (Google)*
7363 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7364 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7365 threads won't reuse the same blinding coefficients.
7367 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7368 lock to call BN_BLINDING_invert_ex, and avoids one use of
7369 BN_BLINDING_update for each BN_BLINDING structure (previously,
7370 the last update always remained unused).
7372 *Emilia Käsper (Google)*
7374 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7376 *Bob Buckholz (Google)*
7378 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011] ###
7380 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7381 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7383 *Kaspar Brand <ossl@velox.ch>*
7385 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7386 for multi-threaded use of ECDH. [CVE-2011-3210][]
7388 *Adam Langley (Google)*
7390 * Fix x509_name_ex_d2i memory leak on bad inputs.
7394 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7395 signature public key algorithm by using OID xref utilities instead.
7396 Before this you could only use some ECC ciphersuites with SHA1 only.
7400 * Add protection against ECDSA timing attacks as mentioned in the paper
7401 by Billy Bob Brumley and Nicola Tuveri, see:
7403 http://eprint.iacr.org/2011/232.pdf
7406 *Billy Bob Brumley and Nicola Tuveri*
7408 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011] ###
7410 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7412 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7414 * Fix bug in string printing code: if *any* escaping is enabled we must
7415 escape the escape character (backslash) or the resulting string is
7420 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010] ###
7422 * Disable code workaround for ancient and obsolete Netscape browsers
7423 and servers: an attacker can use it in a ciphersuite downgrade attack.
7424 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7428 * Fixed J-PAKE implementation error, originally discovered by
7429 Sebastien Martini, further info and confirmation from Stefan
7430 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7434 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010] ###
7436 * Fix extension code to avoid race conditions which can result in a buffer
7437 overrun vulnerability: resumed sessions must not be modified as they can
7438 be shared by multiple threads. CVE-2010-3864
7442 * Fix WIN32 build system to correctly link an ENGINE directory into
7447 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010] ###
7449 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7452 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7454 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010] ###
7456 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7457 context. The operation can be customised via the ctrl mechanism in
7458 case ENGINEs want to include additional functionality.
7462 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7466 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7467 output hashes compatible with older versions of OpenSSL.
7469 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7471 * Fix compression algorithm handling: if resuming a session use the
7472 compression algorithm of the resumed session instead of determining
7473 it from client hello again. Don't allow server to change algorithm.
7477 * Add load_crls() function to apps tidying load_certs() too. Add option
7478 to verify utility to allow additional CRLs to be included.
7482 * Update OCSP request code to permit adding custom headers to the request:
7483 some responders need this.
7487 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7490 *Julia Lawall <julia@diku.dk>*
7492 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7493 needlessly dereferenced structures, used obsolete functions and
7494 didn't handle all updated verify codes correctly.
7498 * Disable MD2 in the default configuration.
7502 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7503 indicate the initial BIO being pushed or popped. This makes it possible
7504 to determine whether the BIO is the one explicitly called or as a result
7505 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7506 it handles reference counts correctly and doesn't zero out the I/O bio
7507 when it is not being explicitly popped. WARNING: applications which
7508 included workarounds for the old buggy behaviour will need to be modified
7509 or they could free up already freed BIOs.
7513 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7514 renaming to all platforms (within the 0.9.8 branch, this was
7515 done conditionally on Netware platforms to avoid a name clash).
7517 *Guenter <lists@gknw.net>*
7519 * Add ECDHE and PSK support to DTLS.
7521 *Michael Tuexen <tuexen@fh-muenster.de>*
7523 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7528 * Add "missing" function EVP_MD_flags() (without this the only way to
7529 retrieve a digest flags is by accessing the structure directly. Update
7530 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
7531 or cipher is registered as in the "from" argument. Print out all
7532 registered digests in the dgst usage message instead of manually
7533 attempting to work them out.
7537 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7538 this allows the use of compression and extensions. Change default cipher
7539 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7540 by default unless an application cipher string requests it.
7544 * Alter match criteria in PKCS12_parse(). It used to try to use local
7545 key ids to find matching certificates and keys but some PKCS#12 files
7546 don't follow the (somewhat unwritten) rules and this strategy fails.
7547 Now just gather all certificates together and the first private key
7548 then look for the first certificate that matches the key.
7552 * Support use of registered digest and cipher names for dgst and cipher
7553 commands instead of having to add each one as a special case. So now
7560 openssl dgst -sha256 foo
7562 and this works for ENGINE based algorithms too.
7567 * Update Gost ENGINE to support parameter files.
7569 *Victor B. Wagner <vitus@cryptocom.ru>*
7571 * Support GeneralizedTime in ca utility.
7573 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7575 * Enhance the hash format used for certificate directory links. The new
7576 form uses the canonical encoding (meaning equivalent names will work
7577 even if they aren't identical) and uses SHA1 instead of MD5. This form
7578 is incompatible with the older format and as a result c_rehash should
7579 be used to rebuild symbolic links.
7583 * Make PKCS#8 the default write format for private keys, replacing the
7584 traditional format. This form is standardised, more secure and doesn't
7585 include an implicit MD5 dependency.
7589 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7590 committed to OpenSSL should pass this lot as a minimum.
7594 * Add session ticket override functionality for use by EAP-FAST.
7596 *Jouni Malinen <j@w1.fi>*
7598 * Modify HMAC functions to return a value. Since these can be implemented
7599 in an ENGINE errors can occur.
7603 * Type-checked OBJ_bsearch_ex.
7607 * Type-checked OBJ_bsearch. Also some constification necessitated
7608 by type-checking. Still to come: TXT_DB, bsearch(?),
7609 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7614 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7615 seconds to a tm structure directly, instead of going through OS
7616 specific date routines. This avoids any issues with OS routines such
7617 as the year 2038 bug. New *_adj() functions for ASN1 time structures
7618 and X509_time_adj_ex() to cover the extended range. The existing
7619 X509_time_adj() is still usable and will no longer have any date issues.
7623 * Delta CRL support. New use deltas option which will attempt to locate
7624 and search any appropriate delta CRLs available.
7626 This work was sponsored by Google.
7630 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7631 code and add additional score elements. Validate alternate CRL paths
7632 as part of the CRL checking and indicate a new error "CRL path validation
7633 error" in this case. Applications wanting additional details can use
7634 the verify callback and check the new "parent" field. If this is not
7635 NULL CRL path validation is taking place. Existing applications won't
7636 see this because it requires extended CRL support which is off by
7639 This work was sponsored by Google.
7643 * Support for freshest CRL extension.
7645 This work was sponsored by Google.
7649 * Initial indirect CRL support. Currently only supported in the CRLs
7650 passed directly and not via lookup. Process certificate issuer
7651 CRL entry extension and lookup CRL entries by bother issuer name
7652 and serial number. Check and process CRL issuer entry in IDP extension.
7654 This work was sponsored by Google.
7658 * Add support for distinct certificate and CRL paths. The CRL issuer
7659 certificate is validated separately in this case. Only enabled if
7660 an extended CRL support flag is set: this flag will enable additional
7661 CRL functionality in future.
7663 This work was sponsored by Google.
7667 * Add support for policy mappings extension.
7669 This work was sponsored by Google.
7673 * Fixes to pathlength constraint, self issued certificate handling,
7674 policy processing to align with RFC3280 and PKITS tests.
7676 This work was sponsored by Google.
7680 * Support for name constraints certificate extension. DN, email, DNS
7681 and URI types are currently supported.
7683 This work was sponsored by Google.
7687 * To cater for systems that provide a pointer-based thread ID rather
7688 than numeric, deprecate the current numeric thread ID mechanism and
7689 replace it with a structure and associated callback type. This
7690 mechanism allows a numeric "hash" to be extracted from a thread ID in
7691 either case, and on platforms where pointers are larger than 'long',
7692 mixing is done to help ensure the numeric 'hash' is usable even if it
7693 can't be guaranteed unique. The default mechanism is to use "&errno"
7694 as a pointer-based thread ID to distinguish between threads.
7696 Applications that want to provide their own thread IDs should now use
7697 CRYPTO_THREADID_set_callback() to register a callback that will call
7698 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7700 Note that ERR_remove_state() is now deprecated, because it is tied
7701 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7702 to free the current thread's error state should be replaced by
7703 ERR_remove_thread_state(NULL).
7705 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7706 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7707 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7708 application was previously providing a numeric thread callback that
7709 was inappropriate for distinguishing threads, then uniqueness might
7710 have been obtained with &errno that happened immediately in the
7711 intermediate development versions of OpenSSL; this is no longer the
7712 case, the numeric thread callback will now override the automatic use
7715 *Geoff Thorpe, with help from Bodo Moeller*
7717 * Initial support for different CRL issuing certificates. This covers a
7718 simple case where the self issued certificates in the chain exist and
7719 the real CRL issuer is higher in the existing chain.
7721 This work was sponsored by Google.
7725 * Removed effectively defunct crypto/store from the build.
7729 * Revamp of STACK to provide stronger type-checking. Still to come:
7730 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7731 ASN1_STRING, CONF_VALUE.
7735 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7736 RAM on SSL connections. This option can save about 34k per idle SSL.
7740 * Revamp of LHASH to provide stronger type-checking. Still to come:
7741 STACK, TXT_DB, bsearch, qsort.
7745 * Initial support for Cryptographic Message Syntax (aka CMS) based
7746 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7747 support for data, signedData, compressedData, digestedData and
7748 encryptedData, envelopedData types included. Scripts to check against
7749 RFC4134 examples draft and interop and consistency checks of many
7750 content types and variants.
7754 * Add options to enc utility to support use of zlib compression BIO.
7758 * Extend mk1mf to support importing of options and assembly language
7759 files from Configure script, currently only included in VC-WIN32.
7760 The assembly language rules can now optionally generate the source
7761 files from the associated perl scripts.
7765 * Implement remaining functionality needed to support GOST ciphersuites.
7766 Interop testing has been performed using CryptoPro implementations.
7768 *Victor B. Wagner <vitus@cryptocom.ru>*
7770 * s390x assembler pack.
7774 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7779 * Implement Opaque PRF Input TLS extension as specified in
7780 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7781 official specification yet and no extension type assignment by
7782 IANA exists, this extension (for now) will have to be explicitly
7783 enabled when building OpenSSL by providing the extension number
7784 to use. For example, specify an option
7786 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7788 to the "config" or "Configure" script to enable the extension,
7789 assuming extension number 0x9527 (which is a completely arbitrary
7790 and unofficial assignment based on the MD5 hash of the Internet
7791 Draft). Note that by doing so, you potentially lose
7792 interoperability with other TLS implementations since these might
7793 be using the same extension number for other purposes.
7795 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7796 opaque PRF input value to use in the handshake. This will create
7797 an internal copy of the length-'len' string at 'src', and will
7798 return non-zero for success.
7800 To get more control and flexibility, provide a callback function
7803 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7804 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7808 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7811 Callback function 'cb' will be called in handshakes, and is
7812 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7813 Argument 'arg' is for application purposes (the value as given to
7814 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7815 be provided to the callback function). The callback function
7816 has to return non-zero to report success: usually 1 to use opaque
7817 PRF input just if possible, or 2 to enforce use of the opaque PRF
7818 input. In the latter case, the library will abort the handshake
7819 if opaque PRF input is not successfully negotiated.
7821 Arguments 'peerinput' and 'len' given to the callback function
7822 will always be NULL and 0 in the case of a client. A server will
7823 see the client's opaque PRF input through these variables if
7824 available (NULL and 0 otherwise). Note that if the server
7825 provides an opaque PRF input, the length must be the same as the
7826 length of the client's opaque PRF input.
7828 Note that the callback function will only be called when creating
7829 a new session (session resumption can resume whatever was
7830 previously negotiated), and will not be called in SSL 2.0
7831 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7832 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7833 for applications that need to enforce opaque PRF input.
7838 * Update ssl code to support digests other than SHA1+MD5 for handshake
7842 *Victor B. Wagner <vitus@cryptocom.ru>*
7844 * Add RFC4507 support to OpenSSL. This includes the corrections in
7845 RFC4507bis. The encrypted ticket format is an encrypted encoded
7846 SSL_SESSION structure, that way new session features are automatically
7849 If a client application caches session in an SSL_SESSION structure
7850 support is transparent because tickets are now stored in the encoded
7853 The SSL_CTX structure automatically generates keys for ticket
7854 protection in servers so again support should be possible
7855 with no application modification.
7857 If a client or server wishes to disable RFC4507 support then the option
7858 SSL_OP_NO_TICKET can be set.
7860 Add a TLS extension debugging callback to allow the contents of any client
7861 or server extensions to be examined.
7863 This work was sponsored by Google.
7867 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7868 OpenSSL should now compile cleanly on gcc 4.2
7870 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7872 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7873 support including streaming MAC support: this is required for GOST
7874 ciphersuite support.
7876 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7878 * Add option -stream to use PKCS#7 streaming in smime utility. New
7879 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7880 to output in BER and PEM format.
7884 * Experimental support for use of HMAC via EVP_PKEY interface. This
7885 allows HMAC to be handled via the EVP_DigestSign*() interface. The
7886 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7887 ENGINE support for HMAC keys which are unextractable. New -mac and
7888 -macopt options to dgst utility.
7892 * New option -sigopt to dgst utility. Update dgst to use
7893 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
7894 alternative signing parameters such as X9.31 or PSS in the dgst
7899 * Change ssl_cipher_apply_rule(), the internal function that does
7900 the work each time a ciphersuite string requests enabling
7901 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7902 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7903 the order of disabled ciphersuites such that those ciphersuites
7904 that most recently went from enabled to disabled not only stay
7905 in order with respect to each other, but also have higher priority
7906 than other disabled ciphersuites the next time ciphersuites are
7909 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
7910 the same ciphersuites as with "HIGH" alone, but in a specific
7911 order where the PSK ciphersuites come first (since they are the
7912 most recently disabled ciphersuites when "HIGH" is parsed).
7914 Also, change ssl_create_cipher_list() (using this new
7915 functionality) such that between otherwise identical
7916 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
7921 * Change ssl_create_cipher_list() so that it automatically
7922 arranges the ciphersuites in reasonable order before starting
7923 to process the rule string. Thus, the definition for "DEFAULT"
7924 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
7925 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
7926 This makes it much easier to arrive at a reasonable default order
7927 in applications for which anonymous ciphers are OK (meaning
7928 that you can't actually use DEFAULT).
7930 *Bodo Moeller; suggested by Victor Duchovni*
7932 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
7933 processing) into multiple integers instead of setting
7934 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
7935 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
7936 (These masks as well as the individual bit definitions are hidden
7937 away into the non-exported interface ssl/ssl_locl.h, so this
7938 change to the definition of the SSL_CIPHER structure shouldn't
7939 affect applications.) This give us more bits for each of these
7940 categories, so there is no longer a need to coagulate AES128 and
7941 AES256 into a single algorithm bit, and to coagulate Camellia128
7942 and Camellia256 into a single algorithm bit, which has led to all
7945 Thus, among other things, the kludge introduced in 0.9.7m and
7946 0.9.8e for masking out AES256 independently of AES128 or masking
7947 out Camellia256 independently of AES256 is not needed here in 0.9.9.
7949 With the change, we also introduce new ciphersuite aliases that
7950 so far were missing: "AES128", "AES256", "CAMELLIA128", and
7955 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
7956 Use the leftmost N bytes of the signature input if the input is
7957 larger than the prime q (with N being the size in bytes of q).
7961 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
7962 it yet and it is largely untested.
7966 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
7970 * Initial incomplete changes to avoid need for function casts in OpenSSL
7971 some compilers (gcc 4.2 and later) reject their use. Safestack is
7972 reimplemented. Update ASN1 to avoid use of legacy functions.
7976 * Win32/64 targets are linked with Winsock2.
7980 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
7981 to external functions. This can be used to increase CRL handling
7982 efficiency especially when CRLs are very large by (for example) storing
7983 the CRL revoked certificates in a database.
7987 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
7988 new CRLs added to a directory can be used. New command line option
7989 -verify_return_error to s_client and s_server. This causes real errors
7990 to be returned by the verify callback instead of carrying on no matter
7991 what. This reflects the way a "real world" verify callback would behave.
7995 * GOST engine, supporting several GOST algorithms and public key formats.
7996 Kindly donated by Cryptocom.
8000 * Partial support for Issuing Distribution Point CRL extension. CRLs
8001 partitioned by DP are handled but no indirect CRL or reason partitioning
8002 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8003 selected via a scoring technique which handles IDP and AKID in CRLs.
8007 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8008 will ultimately be used for all verify operations: this will remove the
8009 X509_STORE dependency on certificate verification and allow alternative
8010 lookup methods. X509_STORE based implementations of these two callbacks.
8014 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8015 Modify get_crl() to find a valid (unexpired) CRL if possible.
8019 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8020 this would be called X509_CRL_cmp() but that name is already used by
8021 a function that just compares CRL issuer names. Cache several CRL
8022 extensions in X509_CRL structure and cache CRLDP in X509.
8026 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8027 this maps equivalent X509_NAME structures into a consistent structure.
8028 Name comparison can then be performed rapidly using memcmp().
8032 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8037 * Allow digests to supply their own micalg string for S/MIME type using
8038 the ctrl EVP_MD_CTRL_MICALG.
8042 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8043 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8044 ctrl. It can then customise the structure before and/or after signing
8049 * New function OBJ_add_sigid() to allow application defined signature OIDs
8050 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8051 to free up any added signature OIDs.
8055 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8056 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8057 digest and cipher tables. New options added to openssl utility:
8058 list-message-digest-algorithms and list-cipher-algorithms.
8062 * Change the array representation of binary polynomials: the list
8063 of degrees of non-zero coefficients is now terminated with -1.
8064 Previously it was terminated with 0, which was also part of the
8065 value; thus, the array representation was not applicable to
8066 polynomials where t^0 has coefficient zero. This change makes
8067 the array representation useful in a more general context.
8071 * Various modifications and fixes to SSL/TLS cipher string
8072 handling. For ECC, the code now distinguishes between fixed ECDH
8073 with RSA certificates on the one hand and with ECDSA certificates
8074 on the other hand, since these are separate ciphersuites. The
8075 unused code for Fortezza ciphersuites has been removed.
8077 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8078 (not "ECDHE"). For consistency with the code for DH
8079 certificates, use of ECDH certificates is now considered ECDH
8080 authentication, not RSA or ECDSA authentication (the latter is
8081 merely the CA's signing algorithm and not actively used in the
8084 The temporary ciphersuite alias "ECCdraft" is no longer
8085 available, and ECC ciphersuites are no longer excluded from "ALL"
8086 and "DEFAULT". The following aliases now exist for RFC 4492
8087 ciphersuites, most of these by analogy with the DH case:
8089 kECDHr - ECDH cert, signed with RSA
8090 kECDHe - ECDH cert, signed with ECDSA
8091 kECDH - ECDH cert (signed with either RSA or ECDSA)
8092 kEECDH - ephemeral ECDH
8093 ECDH - ECDH cert or ephemeral ECDH
8099 AECDH - anonymous ECDH
8100 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8105 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8106 Use correct micalg parameters depending on digest(s) in signed message.
8110 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8111 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8115 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8116 an engine to register a method. Add ENGINE lookups for methods and
8117 functional reference processing.
8121 * New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
8122 EVP_{Sign,Verify}* which allow an application to customise the signature
8127 * New -resign option to smime utility. This adds one or more signers
8128 to an existing PKCS#7 signedData structure. Also -md option to use an
8129 alternative message digest algorithm for signing.
8133 * Tidy up PKCS#7 routines and add new functions to make it easier to
8134 create PKCS7 structures containing multiple signers. Update smime
8135 application to support multiple signers.
8139 * New -macalg option to pkcs12 utility to allow setting of an alternative
8144 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8145 Reorganize PBE internals to lookup from a static table using NIDs,
8146 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8147 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8148 PRF which will be automatically used with PBES2.
8152 * Replace the algorithm specific calls to generate keys in "req" with the
8157 * Update PKCS#7 enveloped data routines to use new API. This is now
8158 supported by any public key method supporting the encrypt operation. A
8159 ctrl is added to allow the public key algorithm to examine or modify
8160 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8165 * Add a ctrl to asn1 method to allow a public key algorithm to express
8166 a default digest type to use. In most cases this will be SHA1 but some
8167 algorithms (such as GOST) need to specify an alternative digest. The
8168 return value indicates how strong the preference is 1 means optional and
8169 2 is mandatory (that is it is the only supported type). Modify
8170 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8171 use the default md. Update openssl utilities to use the default digest
8172 type for signing if it is not explicitly indicated.
8176 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8177 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8178 signing method from the key type. This effectively removes the link
8179 between digests and public key types.
8183 * Add an OID cross reference table and utility functions. Its purpose is to
8184 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8185 rsaEncryption. This will allow some of the algorithm specific hackery
8186 needed to use the correct OID to be removed.
8190 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8191 structures for PKCS7_sign(). They are now set up by the relevant public
8196 * Add provisional EC pkey method with support for ECDSA and ECDH.
8200 * Add support for key derivation (agreement) in the API, DH method and
8205 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8206 public and private key formats. As a side effect these add additional
8207 command line functionality not previously available: DSA signatures can be
8208 generated and verified using pkeyutl and DH key support and generation in
8215 *Oliver Tappe <zooey@hirschkaefer.de>*
8217 * New make target "install_html_docs" installs HTML renditions of the
8220 *Oliver Tappe <zooey@hirschkaefer.de>*
8222 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8223 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8224 support key and parameter generation and add initial key generation
8225 functionality for RSA.
8229 * Add functions for main EVP_PKEY_method operations. The undocumented
8230 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
8231 EVP_PKEY_{encrypt,decrypt}_old.
8235 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8236 key API, doesn't do much yet.
8240 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8241 public key algorithms. New option to openssl utility:
8242 "list-public-key-algorithms" to print out info.
8246 * Implement the Supported Elliptic Curves Extension for
8247 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8251 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8252 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8256 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8257 utilities such as rsa, dsa, dsaparam etc except they process any key
8262 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8263 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8264 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8269 * Initial support for pluggable public key ASN1.
8270 De-spaghettify the public key ASN1 handling. Move public and private
8271 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8272 algorithm specific handling to a single module within the relevant
8273 algorithm directory. Add functions to allow (near) opaque processing
8274 of public and private key structures.
8278 * Implement the Supported Point Formats Extension for
8279 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8283 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8284 for the psk identity [hint] and the psk callback functions to the
8285 SSL_SESSION, SSL and SSL_CTX structure.
8288 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8292 SSL_CTX_use_psk_identity_hint
8293 SSL_get_psk_identity_hint
8294 SSL_get_psk_identity
8295 SSL_use_psk_identity_hint
8298 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8300 * Add RFC 3161 compliant time stamp request creation, response generation
8301 and response verification functionality.
8303 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8305 * Add initial support for TLS extensions, specifically for the server_name
8306 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8307 have new members for a host name. The SSL data structure has an
8308 additional member SSL_CTX *initial_ctx so that new sessions can be
8309 stored in that context to allow for session resumption, even after the
8310 SSL has been switched to a new SSL_CTX in reaction to a client's
8311 server_name extension.
8313 New functions (subject to change):
8315 SSL_get_servername()
8316 SSL_get_servername_type()
8319 New CTRL codes and macros (subject to change):
8321 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8322 - SSL_CTX_set_tlsext_servername_callback()
8323 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8324 - SSL_CTX_set_tlsext_servername_arg()
8325 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8327 openssl s_client has a new '-servername ...' option.
8329 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8330 '-key2 ...', '-servername_fatal' (subject to change). This allows
8331 testing the HostName extension for a specific single host name ('-cert'
8332 and '-key' remain fallbacks for handshakes without HostName
8333 negotiation). If the unrecognized_name alert has to be sent, this by
8334 default is a warning; it becomes fatal with the '-servername_fatal'
8338 *Peter Sylvester, Remy Allais, Christophe Renou*
8340 * Whirlpool hash implementation is added.
8344 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8345 bn(64,32). Because of instruction set limitations it doesn't have
8346 any negative impact on performance. This was done mostly in order
8347 to make it possible to share assembler modules, such as bn_mul_mont
8348 implementations, between 32- and 64-bit builds without hassle.
8352 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8353 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8358 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8359 dedicated Montgomery multiplication procedure, is introduced.
8360 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8361 "64-bit" performance on certain 32-bit targets.
8365 * New option SSL_OP_NO_COMP to disable use of compression selectively
8366 in SSL structures. New SSL ctrl to set maximum send fragment size.
8367 Save memory by setting the I/O buffer sizes dynamically instead of
8368 using the maximum available value.
8372 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8373 in addition to the text details.
8377 * Very, very preliminary EXPERIMENTAL support for printing of general
8378 ASN1 structures. This currently produces rather ugly output and doesn't
8379 handle several customised structures at all.
8383 * Integrated support for PVK file format and some related formats such
8384 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8385 these in the 'rsa' and 'dsa' utilities.
8389 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8393 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8394 place for the (very old) "NETSCAPE" format certificates which are now
8395 handled using new ASN1 code equivalents.
8399 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8400 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8401 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8405 * Modify CRL distribution points extension code to print out previously
8406 unsupported fields. Enhance extension setting code to allow setting of
8411 * Add print and set support for Issuing Distribution Point CRL extension.
8415 * Change 'Configure' script to enable Camellia by default.
8422 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010] ###
8424 * When rejecting SSL/TLS records due to an incorrect version number, never
8425 update s->server with a new major version number. As of
8426 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8427 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8428 the previous behavior could result in a read attempt at NULL when
8429 receiving specific incorrect SSL/TLS records once record payload
8430 protection is active. [CVE-2010-0740][]
8432 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8434 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8435 could be crashed if the relevant tables were not present (e.g. chrooted).
8437 *Tomas Hoger <thoger@redhat.com>*
8439 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010] ###
8441 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8443 *Martin Olsson, Neel Mehta*
8445 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8446 accommodate for stack sorting, always a write lock!).
8450 * On some versions of WIN32 Heap32Next is very slow. This can cause
8451 excessive delays in the RAND_poll(): over a minute. As a workaround
8452 include a time check in the inner Heap32Next loop too.
8456 * The code that handled flushing of data in SSL/TLS originally used the
8457 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8458 the problem outlined in PR#1949. The fix suggested there however can
8459 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8460 of Apache). So instead simplify the code to flush unconditionally.
8461 This should be fine since flushing with no data to flush is a no op.
8465 * Handle TLS versions 2.0 and later properly and correctly use the
8466 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8467 off ancient servers have a habit of sticking around for a while...
8471 * Modify compression code so it frees up structures without using the
8472 ex_data callbacks. This works around a problem where some applications
8473 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8474 restarting) then use compression (e.g. SSL with compression) later.
8475 This results in significant per-connection memory leaks and
8476 has caused some security issues including CVE-2008-1678 and
8481 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8482 change when encrypting or decrypting.
8486 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8487 connect and renegotiate with servers which do not support RI.
8488 Until RI is more widely deployed this option is enabled by default.
8492 * Add "missing" ssl ctrls to clear options and mode.
8496 * If client attempts to renegotiate and doesn't support RI respond with
8497 a no_renegotiation alert as required by RFC5746. Some renegotiating
8498 TLS clients will continue a connection gracefully when they receive
8499 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8500 waiting for a server hello which it will never receive. Now we treat a
8501 received no_renegotiation alert as a fatal error. This is because
8502 applications requesting a renegotiation might well expect it to succeed
8503 and would have no code in place to handle the server denying it so the
8504 only safe thing to do is to terminate the connection.
8508 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8509 peer supports secure renegotiation and 0 otherwise. Print out peer
8510 renegotiation support in s_client/s_server.
8514 * Replace the highly broken and deprecated SPKAC certification method with
8515 the updated NID creation version. This should correctly handle UTF8.
8519 * Implement RFC5746. Re-enable renegotiation but require the extension
8520 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8521 turns out to be a bad idea. It has been replaced by
8522 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8523 SSL_CTX_set_options(). This is really not recommended unless you
8524 know what you are doing.
8526 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8528 * Fixes to stateless session resumption handling. Use initial_ctx when
8529 issuing and attempting to decrypt tickets in case it has changed during
8530 servername handling. Use a non-zero length session ID when attempting
8531 stateless session resumption: this makes it possible to determine if
8532 a resumption has occurred immediately after receiving server hello
8533 (several places in OpenSSL subtly assume this) instead of later in
8538 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8539 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8540 fixes for a few places where the return code is not checked
8543 *Julia Lawall <julia@diku.dk>*
8545 * Add --strict-warnings option to Configure script to include devteam
8546 warnings in other configurations.
8550 * Add support for --libdir option and LIBDIR variable in makefiles. This
8551 makes it possible to install openssl libraries in locations which
8552 have names other than "lib", for example "/usr/lib64" which some
8555 *Steve Henson, based on patch from Jeremy Utley*
8557 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8558 X690 8.9.12 and can produce some misleading textual output of OIDs.
8560 *Steve Henson, reported by Dan Kaminsky*
8562 * Delete MD2 from algorithm tables. This follows the recommendation in
8563 several standards that it is not used in new applications due to
8564 several cryptographic weaknesses. For binary compatibility reasons
8565 the MD2 API is still compiled in by default.
8569 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8574 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8575 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8578 *Guenter <lists@gknw.net>*
8580 * Fix the server certificate chain building code to use X509_verify_cert(),
8581 it used to have an ad-hoc builder which was unable to cope with anything
8582 other than a simple chain.
8584 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8586 * Don't check self signed certificate signatures in X509_verify_cert()
8587 by default (a flag can override this): it just wastes time without
8588 adding any security. As a useful side effect self signed root CAs
8589 with non-FIPS digests are now usable in FIPS mode.
8593 * In dtls1_process_out_of_seq_message() the check if the current message
8594 is already buffered was missing. For every new message was memory
8595 allocated, allowing an attacker to perform an denial of service attack
8596 with sending out of seq handshake messages until there is no memory
8597 left. Additionally every future message was buffered, even if the
8598 sequence number made no sense and would be part of another handshake.
8599 So only messages with sequence numbers less than 10 in advance will be
8600 buffered. [CVE-2009-1378][]
8602 *Robin Seggelmann, discovered by Daniel Mentz*
8604 * Records are buffered if they arrive with a future epoch to be
8605 processed after finishing the corresponding handshake. There is
8606 currently no limitation to this buffer allowing an attacker to perform
8607 a DOS attack with sending records with future epochs until there is no
8608 memory left. This patch adds the pqueue_size() function to determine
8609 the size of a buffer and limits the record buffer to 100 entries.
8612 *Robin Seggelmann, discovered by Daniel Mentz*
8614 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8615 parent structure is freed. [CVE-2009-1379][]
8619 * Handle non-blocking I/O properly in SSL_shutdown() call.
8621 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8625 *Ilya O. <vrghost@gmail.com>*
8627 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009] ###
8629 * Disable renegotiation completely - this fixes a severe security
8630 problem [CVE-2009-3555][] at the cost of breaking all
8631 renegotiation. Renegotiation can be re-enabled by setting
8632 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8633 run-time. This is really not recommended unless you know what
8638 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009] ###
8640 * Don't set val to NULL when freeing up structures, it is freed up by
8641 underlying code. If sizeof(void *) > sizeof(long) this can result in
8642 zeroing past the valid field. [CVE-2009-0789][]
8644 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8646 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8647 checked correctly. This would allow some invalid signed attributes to
8648 appear to verify correctly. [CVE-2009-0591][]
8650 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8652 * Reject UniversalString and BMPString types with invalid lengths. This
8653 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8654 a legal length. [CVE-2009-0590][]
8658 * Set S/MIME signing as the default purpose rather than setting it
8659 unconditionally. This allows applications to override it at the store
8664 * Permit restricted recursion of ASN1 strings. This is needed in practice
8665 to handle some structures.
8669 * Improve efficiency of mem_gets: don't search whole buffer each time
8672 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8674 * New -hex option for openssl rand.
8678 * Print out UTF8String and NumericString when parsing ASN1.
8682 * Support NumericString type for name components.
8686 * Allow CC in the environment to override the automatically chosen
8687 compiler. Note that nothing is done to ensure flags work with the
8692 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009] ###
8694 * Properly check EVP_VerifyFinal() and similar return values
8697 *Ben Laurie, Bodo Moeller, Google Security Team*
8699 * Enable TLS extensions by default.
8703 * Allow the CHIL engine to be loaded, whether the application is
8704 multithreaded or not. (This does not release the developer from the
8705 obligation to set up the dynamic locking callbacks.)
8707 *Sander Temme <sander@temme.net>*
8709 * Use correct exit code if there is an error in dgst command.
8711 *Steve Henson; problem pointed out by Roland Dirlewanger*
8713 * Tweak Configure so that you need to say "experimental-jpake" to enable
8714 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8718 * Add experimental JPAKE support, including demo authentication in
8719 s_client and s_server.
8723 * Set the comparison function in v3_addr_canonize().
8725 *Rob Austein <sra@hactrn.net>*
8727 * Add support for XMPP STARTTLS in s_client.
8729 *Philip Paeps <philip@freebsd.org>*
8731 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8732 to ensure that even with this option, only ciphersuites in the
8733 server's preference list will be accepted. (Note that the option
8734 applies only when resuming a session, so the earlier behavior was
8735 just about the algorithm choice for symmetric cryptography.)
8739 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008] ###
8741 * Fix NULL pointer dereference if a DTLS server received
8742 ChangeCipherSpec as first record [CVE-2009-1386][].
8746 * Fix a state transition in s3_srvr.c and d1_srvr.c
8747 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
8751 * The fix in 0.9.8c that supposedly got rid of unsafe
8752 double-checked locking was incomplete for RSA blinding,
8753 addressing just one layer of what turns out to have been
8754 doubly unsafe triple-checked locking.
8756 So now fix this for real by retiring the MONT_HELPER macro
8757 in crypto/rsa/rsa_eay.c.
8760 *Bodo Moeller; problem pointed out by Marius Schilder*
8762 * Various precautionary measures:
8764 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8766 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8767 (NB: This would require knowledge of the secret session ticket key
8768 to exploit, in which case you'd be SOL either way.)
8770 - Change bn_nist.c so that it will properly handle input BIGNUMs
8771 outside the expected range.
8773 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8777 *Neel Mehta, Bodo Moeller*
8779 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8780 the load fails. Useful for distros.
8782 *Ben Laurie and the FreeBSD team*
8784 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8788 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8792 * Expand ENGINE to support engine supplied SSL client certificate functions.
8794 This work was sponsored by Logica.
8798 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8799 keystores. Support for SSL/TLS client authentication too.
8800 Not compiled unless enable-capieng specified to Configure.
8802 This work was sponsored by Logica.
8806 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8807 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8808 attribute creation routines such as certificate requests and PKCS#12
8813 ### Changes between 0.9.8g and 0.9.8h [28 May 2008] ###
8815 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8816 handshake which could lead to a client crash as found using the
8817 Codenomicon TLS test suite [CVE-2008-1672][]
8819 *Steve Henson, Mark Cox*
8821 * Fix double free in TLS server name extensions which could lead to
8822 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8826 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8828 Clear the error queue to ensure that error entries left from
8829 older function calls do not interfere with the correct operation.
8831 *Lutz Jaenicke, Erik de Castro Lopo*
8833 * Remove root CA certificates of commercial CAs:
8835 The OpenSSL project does not recommend any specific CA and does not
8836 have any policy with respect to including or excluding any CA.
8837 Therefore it does not make any sense to ship an arbitrary selection
8838 of root CA certificates with the OpenSSL software.
8842 * RSA OAEP patches to fix two separate invalid memory reads.
8843 The first one involves inputs when 'lzero' is greater than
8844 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8845 before the beginning of from). The second one involves inputs where
8846 the 'db' section contains nothing but zeroes (there is a one-byte
8847 invalid read after the end of 'db').
8849 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8851 * Partial backport from 0.9.9-dev:
8853 Introduce bn_mul_mont (dedicated Montgomery multiplication
8854 procedure) as a candidate for BIGNUM assembler implementation.
8855 While 0.9.9-dev uses assembler for various architectures, only
8856 x86_64 is available by default here in the 0.9.8 branch, and
8857 32-bit x86 is available through a compile-time setting.
8859 To try the 32-bit x86 assembler implementation, use Configure
8860 option "enable-montasm" (which exists only for this backport).
8862 As "enable-montasm" for 32-bit x86 disclaims code stability
8863 anyway, in this constellation we activate additional code
8864 backported from 0.9.9-dev for further performance improvements,
8865 namely BN_from_montgomery_word. (To enable this otherwise,
8866 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8869 *Andy Polyakov (backport partially by Bodo Moeller)*
8871 * Add TLS session ticket callback. This allows an application to set
8872 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8873 values. This is useful for key rollover for example where several key
8874 sets may exist with different names.
8878 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8879 This was broken until now in 0.9.8 releases, such that the only way
8880 a registered ENGINE could be used (assuming it initialises
8881 successfully on the host) was to explicitly set it as the default
8882 for the relevant algorithms. This is in contradiction with 0.9.7
8883 behaviour and the documentation. With this fix, when an ENGINE is
8884 registered into a given algorithm's table of implementations, the
8885 'uptodate' flag is reset so that auto-discovery will be used next
8886 time a new context for that algorithm attempts to select an
8889 *Ian Lister (tweaked by Geoff Thorpe)*
8891 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8892 implementation in the following ways:
8894 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8897 Lack of BER streaming support means one pass streaming processing is
8898 only supported if data is detached: setting the streaming flag is
8899 ignored for embedded content.
8901 CMS support is disabled by default and must be explicitly enabled
8902 with the enable-cms configuration option.
8906 * Update the GMP engine glue to do direct copies between BIGNUM and
8907 mpz_t when openssl and GMP use the same limb size. Otherwise the
8908 existing "conversion via a text string export" trick is still used.
8910 *Paul Sheer <paulsheer@gmail.com>*
8912 * Zlib compression BIO. This is a filter BIO which compressed and
8913 uncompresses any data passed through it.
8917 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
8918 RFC3394 compatible AES key wrapping.
8922 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
8923 sets string data without copying. X509_ALGOR_set0() and
8924 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
8925 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
8926 from an X509_ATTRIBUTE structure optionally checking it occurs only
8927 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
8932 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
8933 to get the expected BN_FLG_CONSTTIME behavior.
8935 *Bodo Moeller (Google)*
8939 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
8940 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
8941 - added some more tests to do_tests.pl
8942 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
8943 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
8944 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
8945 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
8946 - various changes to netware.pl to enable gcc-cross builds on Win32
8948 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
8949 - various changes to fix missing prototype warnings
8950 - fixed x86nasm.pl to create correct asm files for NASM COFF output
8951 - added AES, WHIRLPOOL and CPUID assembler code to build files
8952 - added missing AES assembler make rules to mk1mf.pl
8953 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
8955 *Guenter Knauf <eflash@gmx.net>*
8957 * Implement certificate status request TLS extension defined in RFC3546.
8958 A client can set the appropriate parameters and receive the encoded
8959 OCSP response via a callback. A server can query the supplied parameters
8960 and set the encoded OCSP response in the callback. Add simplified examples
8961 to s_client and s_server.
8965 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007] ###
8968 + Binary incompatibility of ssl_ctx_st structure
8969 + DTLS interoperation with non-compliant servers
8970 + Don't call get_session_cb() without proposed session
8971 + Fix ia64 assembler code
8973 *Andy Polyakov, Steve Henson*
8975 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007] ###
8977 * DTLS Handshake overhaul. There were longstanding issues with
8978 OpenSSL DTLS implementation, which were making it impossible for
8979 RFC 4347 compliant client to communicate with OpenSSL server.
8980 Unfortunately just fixing these incompatibilities would "cut off"
8981 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
8982 server keeps tolerating non RFC compliant syntax. The opposite is
8983 not true, 0.9.8f client can not communicate with earlier server.
8984 This update even addresses CVE-2007-4995.
8988 * Changes to avoid need for function casts in OpenSSL: some compilers
8989 (gcc 4.2 and later) reject their use.
8990 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
8993 * Add RFC4507 support to OpenSSL. This includes the corrections in
8994 RFC4507bis. The encrypted ticket format is an encrypted encoded
8995 SSL_SESSION structure, that way new session features are automatically
8998 If a client application caches session in an SSL_SESSION structure
8999 support is transparent because tickets are now stored in the encoded
9002 The SSL_CTX structure automatically generates keys for ticket
9003 protection in servers so again support should be possible
9004 with no application modification.
9006 If a client or server wishes to disable RFC4507 support then the option
9007 SSL_OP_NO_TICKET can be set.
9009 Add a TLS extension debugging callback to allow the contents of any client
9010 or server extensions to be examined.
9012 This work was sponsored by Google.
9016 * Add initial support for TLS extensions, specifically for the server_name
9017 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9018 have new members for a host name. The SSL data structure has an
9019 additional member SSL_CTX *initial_ctx so that new sessions can be
9020 stored in that context to allow for session resumption, even after the
9021 SSL has been switched to a new SSL_CTX in reaction to a client's
9022 server_name extension.
9024 New functions (subject to change):
9026 SSL_get_servername()
9027 SSL_get_servername_type()
9030 New CTRL codes and macros (subject to change):
9032 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9033 - SSL_CTX_set_tlsext_servername_callback()
9034 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9035 - SSL_CTX_set_tlsext_servername_arg()
9036 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9038 openssl s_client has a new '-servername ...' option.
9040 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9041 '-key2 ...', '-servername_fatal' (subject to change). This allows
9042 testing the HostName extension for a specific single host name ('-cert'
9043 and '-key' remain fallbacks for handshakes without HostName
9044 negotiation). If the unrecognized_name alert has to be sent, this by
9045 default is a warning; it becomes fatal with the '-servername_fatal'
9049 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9051 * Add AES and SSE2 assembly language support to VC++ build.
9055 * Mitigate attack on final subtraction in Montgomery reduction.
9059 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9060 (which previously caused an internal error).
9064 * Squeeze another 10% out of IGE mode when in != out.
9068 * AES IGE mode speedup.
9070 *Dean Gaudet (Google)*
9072 * Add the Korean symmetric 128-bit cipher SEED (see
9073 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
9074 add SEED ciphersuites from RFC 4162:
9076 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9077 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9078 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9079 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9081 To minimize changes between patchlevels in the OpenSSL 0.9.8
9082 series, SEED remains excluded from compilation unless OpenSSL
9083 is configured with 'enable-seed'.
9085 *KISA, Bodo Moeller*
9087 * Mitigate branch prediction attacks, which can be practical if a
9088 single processor is shared, allowing a spy process to extract
9089 information. For detailed background information, see
9090 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
9091 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9092 and Necessary Software Countermeasures"). The core of the change
9093 are new versions BN_div_no_branch() and
9094 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9095 respectively, which are slower, but avoid the security-relevant
9096 conditional branches. These are automatically called by BN_div()
9097 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9098 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9099 remove a conditional branch.
9101 BN_FLG_CONSTTIME is the new name for the previous
9102 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9103 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9104 in the exponent causes BN_mod_exp_mont() to use the alternative
9105 implementation in BN_mod_exp_mont_consttime().) The old name
9106 remains as a deprecated alias.
9108 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9109 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9110 constant-time implementations for more than just exponentiation.
9111 Here too the old name is kept as a deprecated alias.
9113 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9114 the BN_BLINDING structure gets an independent copy of the
9115 modulus. This means that the previous "BIGNUM *m" argument to
9116 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9117 essentially becomes "const BIGNUM *m", although we can't actually
9118 change this in the header file before 0.9.9. It allows
9119 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9120 enable BN_FLG_CONSTTIME.
9123 *Matthew D Wood (Intel Corp)*
9125 * In the SSL/TLS server implementation, be strict about session ID
9126 context matching (which matters if an application uses a single
9127 external cache for different purposes). Previously,
9128 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9129 set. This did ensure strict client verification, but meant that,
9130 with applications using a single external cache for quite
9131 different requirements, clients could circumvent ciphersuite
9132 restrictions for a given session ID context by starting a session
9133 in a different context.
9137 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9138 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9139 authentication-only ciphersuites.
9143 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9144 not complete and could lead to a possible single byte overflow
9145 [CVE-2007-5135][] [Ben Laurie]
9147 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007] ###
9149 * Since AES128 and AES256 (and similarly Camellia128 and
9150 Camellia256) share a single mask bit in the logic of
9151 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9152 kludge to work properly if AES128 is available and AES256 isn't
9153 (or if Camellia128 is available and Camellia256 isn't).
9157 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9158 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9159 When a point or a seed is encoded in a BIT STRING, we need to
9160 prevent the removal of trailing zero bits to get the proper DER
9161 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9162 of a NamedBitList, for which trailing 0 bits need to be removed.)
9166 * Have SSL/TLS server implementation tolerate "mismatched" record
9167 protocol version while receiving ClientHello even if the
9168 ClientHello is fragmented. (The server can't insist on the
9169 particular protocol version it has chosen before the ServerHello
9170 message has informed the client about his choice.)
9174 * Add RFC 3779 support.
9176 *Rob Austein for ARIN, Ben Laurie*
9178 * Load error codes if they are not already present instead of using a
9179 static variable. This allows them to be cleanly unloaded and reloaded.
9180 Improve header file function name parsing.
9184 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9185 or CAPABILITY handshake as required by RFCs.
9189 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006] ###
9191 * Introduce limits to prevent malicious keys being able to
9192 cause a denial of service. [CVE-2006-2940][]
9194 *Steve Henson, Bodo Moeller*
9196 * Fix ASN.1 parsing of certain invalid structures that can result
9197 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9199 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9200 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9202 * Fix SSL client code which could crash if connecting to a
9203 malicious SSLv2 server. [CVE-2006-4343][]
9205 *Tavis Ormandy and Will Drewry, Google Security Team*
9207 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9208 match only those. Before that, "AES256-SHA" would be interpreted
9209 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9210 the same strength classification in 0.9.7h) as we currently only
9211 have a single AES bit in the ciphersuite description bitmap.
9212 That change, however, also applied to ciphersuite strings such as
9213 "RC4-MD5" that intentionally matched multiple ciphersuites --
9214 namely, SSL 2.0 ciphersuites in addition to the more common ones
9215 from SSL 3.0/TLS 1.0.
9217 So we change the selection algorithm again: Naming an explicit
9218 ciphersuite selects this one ciphersuite, and any other similar
9219 ciphersuite (same bitmap) from *other* protocol versions.
9220 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9221 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9223 Since SSL 2.0 does not have any ciphersuites for which the
9224 128/256 bit distinction would be relevant, this works for now.
9225 The proper fix will be to use different bits for AES128 and
9226 AES256, which would have avoided the problems from the beginning;
9227 however, bits are scarce, so we can only do this in a new release
9228 (not just a patchlevel) when we can change the SSL_CIPHER
9229 definition to split the single 'unsigned long mask' bitmap into
9230 multiple values to extend the available space.
9235 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006] ###
9237 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9238 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9240 * Add AES IGE and biIGE modes.
9244 * Change the Unix randomness entropy gathering to use poll() when
9245 possible instead of select(), since the latter has some
9246 undesirable limitations.
9248 *Darryl Miles via Richard Levitte and Bodo Moeller*
9250 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9251 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9252 cannot be implicitly activated as part of, e.g., the "AES" alias.
9253 However, please upgrade to OpenSSL 0.9.9[-dev] for
9254 non-experimental use of the ECC ciphersuites to get TLS extension
9255 support, which is required for curve and point format negotiation
9256 to avoid potential handshake problems.
9260 * Disable rogue ciphersuites:
9262 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9263 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9264 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9266 The latter two were purportedly from
9267 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9270 Also deactivate the remaining ciphersuites from
9271 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9272 unofficial, and the ID has long expired.
9276 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9277 dual-core machines) and other potential thread-safety issues.
9281 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9282 versions), which is now available for royalty-free use
9283 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
9284 Also, add Camellia TLS ciphersuites from RFC 4132.
9286 To minimize changes between patchlevels in the OpenSSL 0.9.8
9287 series, Camellia remains excluded from compilation unless OpenSSL
9288 is configured with 'enable-camellia'.
9292 * Disable the padding bug check when compression is in use. The padding
9293 bug check assumes the first packet is of even length, this is not
9294 necessarily true if compression is enabled and can result in false
9295 positives causing handshake failure. The actual bug test is ancient
9296 code so it is hoped that implementations will either have fixed it by
9297 now or any which still have the bug do not support compression.
9301 ### Changes between 0.9.8a and 0.9.8b [04 May 2006] ###
9303 * When applying a cipher rule check to see if string match is an explicit
9304 cipher suite and only match that one cipher suite if it is.
9308 * Link in manifests for VC++ if needed.
9310 *Austin Ziegler <halostatue@gmail.com>*
9312 * Update support for ECC-based TLS ciphersuites according to
9313 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9314 TLS extensions, which are supported starting with the 0.9.9
9315 branch, not in the OpenSSL 0.9.8 branch).
9319 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9320 opaque EVP_CIPHER_CTX handling.
9324 * Fixes and enhancements to zlib compression code. We now only use
9325 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9326 to conform with the standards mentioned here:
9327 http://www.zlib.net/DLL_FAQ.txt
9328 Static zlib linking now works on Windows and the new --with-zlib-include
9329 --with-zlib-lib options to Configure can be used to supply the location
9330 of the headers and library. Gracefully handle case where zlib library
9335 * Several fixes and enhancements to the OID generation code. The old code
9336 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9337 handle numbers larger than ULONG_MAX, truncated printing and had a
9338 non standard OBJ_obj2txt() behaviour.
9342 * Add support for building of engines under engine/ as shared libraries
9343 under VC++ build system.
9347 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9348 Hopefully, we will not see any false combination of paths any more.
9352 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005] ###
9354 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9355 (part of SSL_OP_ALL). This option used to disable the
9356 countermeasure against man-in-the-middle protocol-version
9357 rollback in the SSL 2.0 server implementation, which is a bad
9358 idea. [CVE-2005-2969][]
9360 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9361 for Information Security, National Institute of Advanced Industrial
9362 Science and Technology [AIST*, Japan)]
9364 * Add two function to clear and return the verify parameter flags.
9368 * Keep cipherlists sorted in the source instead of sorting them at
9369 runtime, thus removing the need for a lock.
9373 * Avoid some small subgroup attacks in Diffie-Hellman.
9375 *Nick Mathewson and Ben Laurie*
9377 * Add functions for well-known primes.
9381 * Extended Windows CE support.
9383 *Satoshi Nakamura and Andy Polyakov*
9385 * Initialize SSL_METHOD structures at compile time instead of during
9386 runtime, thus removing the need for a lock.
9390 * Make PKCS7_decrypt() work even if no certificate is supplied by
9391 attempting to decrypt each encrypted key in turn. Add support to
9396 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005] ###
9398 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9401 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9405 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9406 key into the same file any more.
9410 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9414 * Add -utf8 command line and config file option to 'ca'.
9416 *Stefan <stf@udoma.org*
9418 * Removed the macro des_crypt(), as it seems to conflict with some
9419 libraries. Use DES_crypt().
9423 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9424 involves renaming the source and generated shared-libs for
9425 both. The engines will accept the corrected or legacy ids
9426 ('ncipher' and '4758_cca' respectively) when binding. NB,
9427 this only applies when building 'shared'.
9429 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9431 * Add attribute functions to EVP_PKEY structure. Modify
9432 PKCS12_create() to recognize a CSP name attribute and
9433 use it. Make -CSP option work again in pkcs12 utility.
9437 * Add new functionality to the bn blinding code:
9438 - automatic re-creation of the BN_BLINDING parameters after
9439 a fixed number of uses (currently 32)
9440 - add new function for parameter creation
9441 - introduce flags to control the update behaviour of the
9442 BN_BLINDING parameters
9443 - hide BN_BLINDING structure
9444 Add a second BN_BLINDING slot to the RSA structure to improve
9445 performance when a single RSA object is shared among several
9450 * Add support for DTLS.
9452 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9454 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9455 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9459 * Remove buggy and incomplete DH cert support from
9460 ssl/ssl_rsa.c and ssl/s3_both.c
9464 * Use SHA-1 instead of MD5 as the default digest algorithm for
9465 the apps/openssl applications.
9469 * Compile clean with "-Wall -Wmissing-prototypes
9470 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9471 DEBUG_SAFESTACK must also be set.
9475 * Change ./Configure so that certain algorithms can be disabled by default.
9476 The new counterpiece to "no-xxx" is "enable-xxx".
9478 The patented RC5 and MDC2 algorithms will now be disabled unless
9479 "enable-rc5" and "enable-mdc2", respectively, are specified.
9481 (IDEA remains enabled despite being patented. This is because IDEA
9482 is frequently required for interoperability, and there is no license
9483 fee for non-commercial use. As before, "no-idea" can be used to
9484 avoid this algorithm.)
9489 * Add processing of proxy certificates (see RFC 3820). This work was
9490 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9491 EGEE (Enabling Grids for E-science in Europe).
9495 * RC4 performance overhaul on modern architectures/implementations, such
9496 as Intel P4, IA-64 and AMD64.
9500 * New utility extract-section.pl. This can be used specify an alternative
9501 section number in a pod file instead of having to treat each file as
9502 a separate case in Makefile. This can be done by adding two lines to the
9505 =for comment openssl_section:XXX
9507 The blank line is mandatory.
9512 * New arguments -certform, -keyform and -pass for s_client and s_server
9513 to allow alternative format key and certificate files and passphrase
9518 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9519 update associated structures and add various utility functions.
9521 Add new policy related verify parameters, include policy checking in
9522 standard verify code. Enhance 'smime' application with extra parameters
9523 to support policy checking and print out.
9527 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9528 Nehemiah processors. These extensions support AES encryption in hardware
9529 as well as RNG (though RNG support is currently disabled).
9531 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9533 * Deprecate BN_[get|set]_params() functions (they were ignored internally).
9537 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9539 *Andy Polyakov and a number of other people*
9541 * Improved PowerPC platform support. Most notably BIGNUM assembler
9542 implementation contributed by IBM.
9544 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9546 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9547 exponent rather than 'unsigned long'. There is a corresponding change to
9548 the new 'rsa_keygen' element of the RSA_METHOD structure.
9550 *Jelte Jansen, Geoff Thorpe*
9552 * Functionality for creating the initial serial number file is now
9553 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9555 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9556 number file to 1, which is bound to cause problems. To avoid
9557 the problems while respecting compatibility between different 0.9.7
9558 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9559 CA.pl for serial number initialization. With the new release 0.9.8,
9560 we can fix the problem directly in the 'ca' utility.)
9564 * Reduced header interdependencies by declaring more opaque objects in
9565 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9566 give fewer recursive includes, which could break lazy source code - so
9567 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9568 developers should define this symbol when building and using openssl to
9569 ensure they track the recommended behaviour, interfaces, [etc], but
9570 backwards-compatible behaviour prevails when this isn't defined.
9574 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9578 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9579 This will generate a random key of the appropriate length based on the
9580 cipher context. The EVP_CIPHER can provide its own random key generation
9581 routine to support keys of a specific form. This is used in the des and
9582 3des routines to generate a key of the correct parity. Update S/MIME
9583 code to use new functions and hence generate correct parity DES keys.
9584 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9585 valid (weak or incorrect parity).
9589 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9590 as looking them up. This is useful when the verified structure may contain
9591 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9592 present unless the new PKCS7_NO_CRL flag is asserted.
9596 * Extend ASN1 oid configuration module. It now additionally accepts the
9599 shortName = some long name, 1.2.3.4
9603 * Reimplemented the BN_CTX implementation. There is now no more static
9604 limitation on the number of variables it can handle nor the depth of the
9605 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9606 information can now expand as required, and rather than having a single
9607 static array of bignums, BN_CTX now uses a linked-list of such arrays
9608 allowing it to expand on demand whilst maintaining the usefulness of
9609 BN_CTX's "bundling".
9613 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9614 to allow all RSA operations to function using a single BN_CTX.
9618 * Preliminary support for certificate policy evaluation and checking. This
9619 is initially intended to pass the tests outlined in "Conformance Testing
9620 of Relying Party Client Certificate Path Processing Logic" v1.07.
9624 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9625 remained unused and not that useful. A variety of other little bignum
9626 tweaks and fixes have also been made continuing on from the audit (see
9631 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9632 associated ASN1, EVP and SSL functions and old ASN1 macros.
9636 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9637 and this should never fail. So the return value from the use of
9638 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9639 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9643 * BN_CTX_get() should return zero-valued bignums, providing the same
9644 initialised value as BN_new().
9646 *Geoff Thorpe, suggested by Ulf Möller*
9648 * Support for inhibitAnyPolicy certificate extension.
9652 * An audit of the BIGNUM code is underway, for which debugging code is
9653 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9654 is considered valid when processing BIGNUMs, and causes execution to
9655 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9656 further steps are taken to deliberately pollute unused data in BIGNUM
9657 structures to try and expose faulty code further on. For now, openssl will
9658 (in its default mode of operation) continue to tolerate the inconsistent
9659 forms that it has tolerated in the past, but authors and packagers should
9660 consider trying openssl and their own applications when compiled with
9661 these debugging symbols defined. It will help highlight potential bugs in
9662 their own code, and will improve the test coverage for OpenSSL itself. At
9663 some point, these tighter rules will become openssl's default to improve
9664 maintainability, though the assert()s and other overheads will remain only
9665 in debugging configurations. See bn.h for more details.
9667 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9669 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9670 that can only be obtained through BN_CTX_new() (which implicitly
9671 initialises it). The presence of this function only made it possible
9672 to overwrite an existing structure (and cause memory leaks).
9676 * Because of the callback-based approach for implementing LHASH as a
9677 template type, lh_insert() adds opaque objects to hash-tables and
9678 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9679 to clean up those corresponding objects before destroying the hash table
9680 (and losing the object pointers). So some over-zealous constifications in
9681 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9682 objects as "const" and the lh_doall[_arg] callback wrappers are not
9683 prototyped to have "const" restrictions on the object pointers they are
9684 given (and so aren't required to cast them away any more).
9688 * The tmdiff.h API was so ugly and minimal that our own timing utility
9689 (speed) prefers to use its own implementation. The two implementations
9690 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9691 its object type properly exposed (MS_TM) instead of casting to/from "char
9692 *". This may still change yet if someone realises MS_TM and "ms_time_***"
9693 aren't necessarily the greatest nomenclatures - but this is what was used
9694 internally to the implementation so I've used that for now.
9698 * Ensure that deprecated functions do not get compiled when
9699 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9700 the self-tests were still using deprecated key-generation functions so
9701 these have been updated also.
9705 * Reorganise PKCS#7 code to separate the digest location functionality
9706 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9707 New function PKCS7_set_digest() to set the digest type for PKCS#7
9708 digestedData type. Add additional code to correctly generate the
9709 digestedData type and add support for this type in PKCS7 initialization
9714 * New function PKCS7_set0_type_other() this initializes a PKCS7
9715 structure of type "other".
9719 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9720 sure the loop does correctly stop and breaking ("division by zero")
9721 modulus operations are not performed. The (pre-generated) prime
9722 table crypto/bn/bn_prime.h was already correct, but it could not be
9723 re-generated on some platforms because of the "division by zero"
9724 situation in the script.
9726 *Ralf S. Engelschall*
9728 * Update support for ECC-based TLS ciphersuites according to
9729 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9730 SHA-1 now is only used for "small" curves (where the
9731 representation of a field element takes up to 24 bytes); for
9732 larger curves, the field element resulting from ECDH is directly
9733 used as premaster secret.
9735 *Douglas Stebila (Sun Microsystems Laboratories)*
9737 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9738 curve secp160r1 to the tests.
9740 *Douglas Stebila (Sun Microsystems Laboratories)*
9742 * Add the possibility to load symbols globally with DSO.
9744 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9746 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9747 control of the error stack.
9751 * Add support for STORE in ENGINE.
9755 * Add the STORE type. The intention is to provide a common interface
9756 to certificate and key stores, be they simple file-based stores, or
9757 HSM-type store, or LDAP stores, or...
9758 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9762 * Add a generic structure called OPENSSL_ITEM. This can be used to
9763 pass a list of arguments to any function as well as provide a way
9764 for a function to pass data back to the caller.
9768 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9769 works like BUF_strdup() but can be used to duplicate a portion of
9770 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9775 * Add the function sk_find_ex() which works like sk_find(), but will
9776 return an index to an element even if an exact match couldn't be
9777 found. The index is guaranteed to point at the element where the
9778 searched-for key would be inserted to preserve sorting order.
9782 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9783 takes an extra flags argument for optional functionality. Currently,
9784 the following flags are defined:
9786 OBJ_BSEARCH_VALUE_ON_NOMATCH
9787 This one gets OBJ_bsearch_ex() to return a pointer to the first
9788 element where the comparing function returns a negative or zero
9791 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9792 This one gets OBJ_bsearch_ex() to return a pointer to the first
9793 element where the comparing function returns zero. This is useful
9794 if there are more than one element where the comparing function
9799 * Make it possible to create self-signed certificates with 'openssl ca'
9800 in such a way that the self-signed certificate becomes part of the
9801 CA database and uses the same mechanisms for serial number generation
9802 as all other certificate signing. The new flag '-selfsign' enables
9803 this functionality. Adapt CA.sh and CA.pl.in.
9807 * Add functionality to check the public key of a certificate request
9808 against a given private. This is useful to check that a certificate
9809 request can be signed by that key (self-signing).
9813 * Make it possible to have multiple active certificates with the same
9814 subject in the CA index file. This is done only if the keyword
9815 'unique_subject' is set to 'no' in the main CA section (default
9816 if 'CA_default') of the configuration file. The value is saved
9817 with the database itself in a separate index attribute file,
9818 named like the index file with '.attr' appended to the name.
9822 * Generate multi-valued AVAs using '+' notation in config files for
9827 * Support for nameConstraints certificate extension.
9831 * Support for policyConstraints certificate extension.
9835 * Support for policyMappings certificate extension.
9839 * Make sure the default DSA_METHOD implementation only uses its
9840 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9841 and change its own handlers to be NULL so as to remove unnecessary
9842 indirection. This lets alternative implementations fallback to the
9843 default implementation more easily.
9847 * Support for directoryName in GeneralName related extensions
9852 * Make it possible to link applications using Makefile.shared.
9853 Make that possible even when linking against static libraries!
9857 * Support for single pass processing for S/MIME signing. This now
9858 means that S/MIME signing can be done from a pipe, in addition
9859 cleartext signing (multipart/signed type) is effectively streaming
9860 and the signed data does not need to be all held in memory.
9862 This is done with a new flag PKCS7_STREAM. When this flag is set
9863 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9864 is done after the data is output (and digests calculated) in
9865 SMIME_write_PKCS7().
9869 * Add full support for -rpath/-R, both in shared libraries and
9870 applications, at least on the platforms where it's known how
9875 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9876 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9877 will now compute a table of multiples of the generator that
9878 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9879 faster (notably in the case of a single point multiplication,
9880 scalar * generator).
9882 *Nils Larsch, Bodo Moeller*
9884 * IPv6 support for certificate extensions. The various extensions
9885 which use the IP:a.b.c.d can now take IPv6 addresses using the
9886 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9891 * Added an ENGINE that implements RSA by performing private key
9892 exponentiations with the GMP library. The conversions to and from
9893 GMP's mpz_t format aren't optimised nor are any montgomery forms
9894 cached, and on x86 it appears OpenSSL's own performance has caught up.
9895 However there are likely to be other architectures where GMP could
9896 provide a boost. This ENGINE is not built in by default, but it can be
9897 specified at Configure time and should be accompanied by the necessary
9898 linker additions, eg;
9899 ./config -DOPENSSL_USE_GMP -lgmp
9903 * "openssl engine" will not display ENGINE/DSO load failure errors when
9904 testing availability of engines with "-t" - the old behaviour is
9905 produced by increasing the feature's verbosity with "-tt".
9909 * ECDSA routines: under certain error conditions uninitialized BN objects
9910 could be freed. Solution: make sure initialization is performed early
9911 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
9916 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
9917 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9918 software implementations. For DSA and DH, parameter generation can
9919 also be overridden by providing the appropriate method callbacks.
9923 * Change the "progress" mechanism used in key-generation and
9924 primality testing to functions that take a new BN_GENCB pointer in
9925 place of callback/argument pairs. The new API functions have "_ex"
9926 postfixes and the older functions are reimplemented as wrappers for
9927 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
9928 declarations of the old functions to help (graceful) attempts to
9929 migrate to the new functions. Also, the new key-generation API
9930 functions operate on a caller-supplied key-structure and return
9931 success/failure rather than returning a key or NULL - this is to
9932 help make "keygen" another member function of RSA_METHOD etc.
9934 Example for using the new callback interface:
9936 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
9940 BN_GENCB_set(&my_cb, my_callback, my_arg);
9942 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
9943 /* For the meaning of a, b in calls to my_callback(), see the
9944 * documentation of the function that calls the callback.
9945 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
9946 * my_callback should return 1 if it wants BN_is_prime_ex()
9947 * to continue, or 0 to stop.
9952 * Change the ZLIB compression method to be stateful, and make it
9953 available to TLS with the number defined in
9954 draft-ietf-tls-compression-04.txt.
9958 * Add the ASN.1 structures and functions for CertificatePair, which
9959 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
9961 CertificatePair ::= SEQUENCE {
9962 forward [0] Certificate OPTIONAL,
9963 reverse [1] Certificate OPTIONAL,
9964 -- at least one of the pair shall be present -- }
9966 Also implement the PEM functions to read and write certificate
9967 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
9969 This needed to be defined, mostly for the sake of the LDAP
9970 attribute crossCertificatePair, but may prove useful elsewhere as
9975 * Make it possible to inhibit symlinking of shared libraries in
9976 Makefile.shared, for Cygwin's sake.
9980 * Extend the BIGNUM API by creating a function
9981 void BN_set_negative(BIGNUM *a, int neg);
9982 and a macro that behave like
9983 int BN_is_negative(const BIGNUM *a);
9985 to avoid the need to access 'a->neg' directly in applications.
9989 * Implement fast modular reduction for pseudo-Mersenne primes
9990 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
9991 EC_GROUP_new_curve_GFp() will now automatically use this
9994 *Nils Larsch <nla@trustcenter.de>*
9996 * Add new lock type (CRYPTO_LOCK_BN).
10000 * Change the ENGINE framework to automatically load engines
10001 dynamically from specific directories unless they could be
10002 found to already be built in or loaded. Move all the
10003 current engines except for the cryptodev one to a new
10004 directory engines/.
10005 The engines in engines/ are built as shared libraries if
10006 the "shared" options was given to ./Configure or ./config.
10007 Otherwise, they are inserted in libcrypto.a.
10008 /usr/local/ssl/engines is the default directory for dynamic
10009 engines, but that can be overridden at configure time through
10010 the usual use of --prefix and/or --openssldir, and at run
10011 time with the environment variable OPENSSL_ENGINES.
10013 *Geoff Thorpe and Richard Levitte*
10015 * Add Makefile.shared, a helper makefile to build shared
10016 libraries. Adapt Makefile.org.
10020 * Add version info to Win32 DLLs.
10022 *Peter 'Luna' Runestig" <peter@runestig.com>*
10024 * Add new 'medium level' PKCS#12 API. Certificates and keys
10025 can be added using this API to created arbitrary PKCS#12
10026 files while avoiding the low level API.
10028 New options to PKCS12_create(), key or cert can be NULL and
10029 will then be omitted from the output file. The encryption
10030 algorithm NIDs can be set to -1 for no encryption, the mac
10031 iteration count can be set to 0 to omit the mac.
10033 Enhance pkcs12 utility by making the -nokeys and -nocerts
10034 options work when creating a PKCS#12 file. New option -nomac
10035 to omit the mac, NONE can be set for an encryption algorithm.
10036 New code is modified to use the enhanced PKCS12_create()
10037 instead of the low level API.
10041 * Extend ASN1 encoder to support indefinite length constructed
10042 encoding. This can output sequences tags and octet strings in
10043 this form. Modify pk7_asn1.c to support indefinite length
10044 encoding. This is experimental and needs additional code to
10045 be useful, such as an ASN1 bio and some enhanced streaming
10048 Extend template encode functionality so that tagging is passed
10049 down to the template encoder.
10053 * Let 'openssl req' fail if an argument to '-newkey' is not
10054 recognized instead of using RSA as a default.
10058 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10059 As these are not official, they are not included in "ALL";
10060 the "ECCdraft" ciphersuite group alias can be used to select them.
10062 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10064 * Add ECDH engine support.
10066 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10068 * Add ECDH in new directory crypto/ecdh/.
10070 *Douglas Stebila (Sun Microsystems Laboratories)*
10072 * Let BN_rand_range() abort with an error after 100 iterations
10073 without success (which indicates a broken PRNG).
10077 * Change BN_mod_sqrt() so that it verifies that the input value
10078 is really the square of the return value. (Previously,
10079 BN_mod_sqrt would show GIGO behaviour.)
10083 * Add named elliptic curves over binary fields from X9.62, SECG,
10084 and WAP/WTLS; add OIDs that were still missing.
10086 *Sheueling Chang Shantz and Douglas Stebila
10087 (Sun Microsystems Laboratories)*
10089 * Extend the EC library for elliptic curves over binary fields
10090 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10093 EC_GF2m_simple_method
10097 EC_GROUP_new_curve_GF2m
10098 EC_GROUP_set_curve_GF2m
10099 EC_GROUP_get_curve_GF2m
10100 EC_POINT_set_affine_coordinates_GF2m
10101 EC_POINT_get_affine_coordinates_GF2m
10102 EC_POINT_set_compressed_coordinates_GF2m
10104 Point compression for binary fields is disabled by default for
10105 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10108 As binary polynomials are represented as BIGNUMs, various members
10109 of the EC_GROUP and EC_POINT data structures can be shared
10110 between the implementations for prime fields and binary fields;
10111 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
10112 are essentially identical to their ..._GFp counterparts.
10113 (For simplicity, the '..._GFp' prefix has been dropped from
10114 various internal method names.)
10116 An internal 'field_div' method (similar to 'field_mul' and
10117 'field_sqr') has been added; this is used only for binary fields.
10119 *Sheueling Chang Shantz and Douglas Stebila
10120 (Sun Microsystems Laboratories)*
10122 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10123 through methods ('mul', 'precompute_mult').
10125 The generic implementations (now internally called 'ec_wNAF_mul'
10126 and 'ec_wNAF_precomputed_mult') remain the default if these
10127 methods are undefined.
10129 *Sheueling Chang Shantz and Douglas Stebila
10130 (Sun Microsystems Laboratories)*
10132 * New function EC_GROUP_get_degree, which is defined through
10133 EC_METHOD. For curves over prime fields, this returns the bit
10134 length of the modulus.
10136 *Sheueling Chang Shantz and Douglas Stebila
10137 (Sun Microsystems Laboratories)*
10139 * New functions EC_GROUP_dup, EC_POINT_dup.
10140 (These simply call ..._new and ..._copy).
10142 *Sheueling Chang Shantz and Douglas Stebila
10143 (Sun Microsystems Laboratories)*
10145 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10146 Polynomials are represented as BIGNUMs (where the sign bit is not
10147 used) in the following functions [macros]:
10150 BN_GF2m_sub [= BN_GF2m_add]
10151 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10152 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10153 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10155 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10156 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10157 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10158 BN_GF2m_cmp [= BN_ucmp]
10160 (Note that only the 'mod' functions are actually for fields GF(2^m).
10161 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10163 For some functions, an the irreducible polynomial defining a
10164 field can be given as an 'unsigned int[]' with strictly
10165 decreasing elements giving the indices of those bits that are set;
10166 i.e., p[] represents the polynomial
10167 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10169 p[0] > p[1] > ... > p[k] = 0.
10170 This applies to the following functions:
10173 BN_GF2m_mod_mul_arr
10174 BN_GF2m_mod_sqr_arr
10175 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10176 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10177 BN_GF2m_mod_exp_arr
10178 BN_GF2m_mod_sqrt_arr
10179 BN_GF2m_mod_solve_quad_arr
10183 Conversion can be performed by the following functions:
10188 bntest.c has additional tests for binary polynomial arithmetic.
10190 Two implementations for BN_GF2m_mod_div() are available.
10191 The default algorithm simply uses BN_GF2m_mod_inv() and
10192 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10193 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10194 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10196 *Sheueling Chang Shantz and Douglas Stebila
10197 (Sun Microsystems Laboratories)*
10199 * Add new error code 'ERR_R_DISABLED' that can be used when some
10200 functionality is disabled at compile-time.
10202 *Douglas Stebila <douglas.stebila@sun.com>*
10204 * Change default behaviour of 'openssl asn1parse' so that more
10205 information is visible when viewing, e.g., a certificate:
10207 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10208 mode the content of non-printable OCTET STRINGs is output in a
10209 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10210 avoid the appearance of a printable string.
10212 *Nils Larsch <nla@trustcenter.de>*
10214 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10216 EC_GROUP_set_asn1_flag()
10217 EC_GROUP_get_asn1_flag()
10218 EC_GROUP_set_point_conversion_form()
10219 EC_GROUP_get_point_conversion_form()
10220 These control ASN1 encoding details:
10221 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10222 has been set to OPENSSL_EC_NAMED_CURVE.
10223 - Points are encoded in uncompressed form by default; options for
10224 asn1_for are as for point2oct, namely
10225 POINT_CONVERSION_COMPRESSED
10226 POINT_CONVERSION_UNCOMPRESSED
10227 POINT_CONVERSION_HYBRID
10229 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10231 EC_GROUP_set_seed()
10232 EC_GROUP_get0_seed()
10233 EC_GROUP_get_seed_len()
10234 This is used only for ASN1 purposes (so far).
10236 *Nils Larsch <nla@trustcenter.de>*
10238 * Add 'field_type' member to EC_METHOD, which holds the NID
10239 of the appropriate field type OID. The new function
10240 EC_METHOD_get_field_type() returns this value.
10242 *Nils Larsch <nla@trustcenter.de>*
10245 EC_POINT_point2bn()
10246 EC_POINT_bn2point()
10247 EC_POINT_point2hex()
10248 EC_POINT_hex2point()
10249 providing useful interfaces to EC_POINT_point2oct() and
10250 EC_POINT_oct2point().
10252 *Nils Larsch <nla@trustcenter.de>*
10254 * Change internals of the EC library so that the functions
10255 EC_GROUP_set_generator()
10256 EC_GROUP_get_generator()
10257 EC_GROUP_get_order()
10258 EC_GROUP_get_cofactor()
10259 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10260 to methods, which would lead to unnecessary code duplication when
10261 adding different types of curves.
10263 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10265 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10266 arithmetic, and such that modified wNAFs are generated
10267 (which avoid length expansion in many cases).
10271 * Add a function EC_GROUP_check_discriminant() (defined via
10272 EC_METHOD) that verifies that the curve discriminant is non-zero.
10274 Add a function EC_GROUP_check() that makes some sanity tests
10275 on a EC_GROUP, its generator and order. This includes
10276 EC_GROUP_check_discriminant().
10278 *Nils Larsch <nla@trustcenter.de>*
10280 * Add ECDSA in new directory crypto/ecdsa/.
10282 Add applications 'openssl ecparam' and 'openssl ecdsa'
10283 (these are based on 'openssl dsaparam' and 'openssl dsa').
10285 ECDSA support is also included in various other files across the
10286 library. Most notably,
10287 - 'openssl req' now has a '-newkey ecdsa:file' option;
10288 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10289 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10290 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10291 them suitable for ECDSA where domain parameters must be
10292 extracted before the specific public key;
10293 - ECDSA engine support has been added.
10295 *Nils Larsch <nla@trustcenter.de>*
10297 * Include some named elliptic curves, and add OIDs from X9.62,
10298 SECG, and WAP/WTLS. Each curve can be obtained from the new
10300 EC_GROUP_new_by_curve_name(),
10301 and the list of available named curves can be obtained with
10302 EC_get_builtin_curves().
10303 Also add a 'curve_name' member to EC_GROUP objects, which can be
10305 EC_GROUP_set_curve_name()
10306 EC_GROUP_get_curve_name()
10308 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10310 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10311 was actually never needed) and in BN_mul(). The removal in BN_mul()
10312 required a small change in bn_mul_part_recursive() and the addition
10313 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10314 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10315 bn_sub_words() and bn_add_words() except they take arrays with
10320 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007] ###
10322 * Cleanse PEM buffers before freeing them since they may contain
10325 *Benjamin Bennett <ben@psc.edu>*
10327 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10328 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10329 authentication-only ciphersuites.
10333 * Since AES128 and AES256 share a single mask bit in the logic of
10334 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10335 kludge to work properly if AES128 is available and AES256 isn't.
10339 * Expand security boundary to match 1.1.1 module.
10343 * Remove redundant features: hash file source, editing of test vectors
10344 modify fipsld to use external fips_premain.c signature.
10348 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10349 run algorithm test programs.
10353 * Make algorithm test programs more tolerant of whitespace.
10357 * Have SSL/TLS server implementation tolerate "mismatched" record
10358 protocol version while receiving ClientHello even if the
10359 ClientHello is fragmented. (The server can't insist on the
10360 particular protocol version it has chosen before the ServerHello
10361 message has informed the client about his choice.)
10365 * Load error codes if they are not already present instead of using a
10366 static variable. This allows them to be cleanly unloaded and reloaded.
10370 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006] ###
10372 * Introduce limits to prevent malicious keys being able to
10373 cause a denial of service. [CVE-2006-2940][]
10375 *Steve Henson, Bodo Moeller*
10377 * Fix ASN.1 parsing of certain invalid structures that can result
10378 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10380 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10381 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10383 * Fix SSL client code which could crash if connecting to a
10384 malicious SSLv2 server. [CVE-2006-4343][]
10386 *Tavis Ormandy and Will Drewry, Google Security Team*
10388 * Change ciphersuite string processing so that an explicit
10389 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10390 will no longer include "AES128-SHA"), and any other similar
10391 ciphersuite (same bitmap) from *other* protocol versions (so that
10392 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10393 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10394 changes from 0.9.8b and 0.9.8d.
10398 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006] ###
10400 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10401 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10403 * Change the Unix randomness entropy gathering to use poll() when
10404 possible instead of select(), since the latter has some
10405 undesirable limitations.
10407 *Darryl Miles via Richard Levitte and Bodo Moeller*
10409 * Disable rogue ciphersuites:
10411 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10412 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10413 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10415 The latter two were purportedly from
10416 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10419 Also deactivate the remaining ciphersuites from
10420 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10421 unofficial, and the ID has long expired.
10425 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10426 dual-core machines) and other potential thread-safety issues.
10430 ### Changes between 0.9.7i and 0.9.7j [04 May 2006] ###
10432 * Adapt fipsld and the build system to link against the validated FIPS
10433 module in FIPS mode.
10437 * Fixes for VC++ 2005 build under Windows.
10441 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10442 from a Windows bash shell such as MSYS. It is autodetected from the
10443 "config" script when run from a VC++ environment. Modify standard VC++
10444 build to use fipscanister.o from the GNU make build.
10448 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005] ###
10450 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10451 The value now differs depending on if you build for FIPS or not.
10452 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10453 safely run with a non-FIPSed libcrypto, as it may crash because of
10454 the difference induced by this change.
10458 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005] ###
10460 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10461 (part of SSL_OP_ALL). This option used to disable the
10462 countermeasure against man-in-the-middle protocol-version
10463 rollback in the SSL 2.0 server implementation, which is a bad
10464 idea. [CVE-2005-2969][]
10466 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10467 for Information Security, National Institute of Advanced Industrial
10468 Science and Technology [AIST*, Japan)]
10470 * Minimal support for X9.31 signatures and PSS padding modes. This is
10471 mainly for FIPS compliance and not fully integrated at this stage.
10475 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10476 the exponentiation using a fixed-length exponent. (Otherwise,
10477 the information leaked through timing could expose the secret key
10478 after many signatures; cf. Bleichenbacher's attack on DSA with
10483 * Make a new fixed-window mod_exp implementation the default for
10484 RSA, DSA, and DH private-key operations so that the sequence of
10485 squares and multiplies and the memory access pattern are
10486 independent of the particular secret key. This will mitigate
10487 cache-timing and potential related attacks.
10489 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10490 and this is automatically used by BN_mod_exp_mont() if the new flag
10491 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10492 will use this BN flag for private exponents unless the flag
10493 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10494 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10497 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10499 * Change the client implementation for SSLv23_method() and
10500 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10501 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10502 (Previously, the SSL 2.0 backwards compatible Client Hello
10503 message format would be used even with SSL_OP_NO_SSLv2.)
10507 * Add support for smime-type MIME parameter in S/MIME messages which some
10512 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10513 a threadsafe manner. Modify rsa code to use new function and add calls
10514 to dsa and dh code (which had race conditions before).
10518 * Include the fixed error library code in the C error file definitions
10519 instead of fixing them up at runtime. This keeps the error code
10520 structures constant.
10524 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005] ###
10526 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10529 * Fixes for newer kerberos headers. NB: the casts are needed because
10530 the 'length' field is signed on one version and unsigned on another
10531 with no (?) obvious way to tell the difference, without these VC++
10532 complains. Also the "definition" of FAR (blank) is no longer included
10533 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10534 some needed definitions.
10538 * Undo Cygwin change.
10542 * Added support for proxy certificates according to RFC 3820.
10543 Because they may be a security thread to unaware applications,
10544 they must be explicitly allowed in run-time. See
10545 docs/HOWTO/proxy_certificates.txt for further information.
10549 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005] ###
10551 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10552 server and client random values. Previously
10553 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10554 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10556 This change has negligible security impact because:
10558 1. Server and client random values still have 24 bytes of pseudo random
10561 2. Server and client random values are sent in the clear in the initial
10564 3. The master secret is derived using the premaster secret (48 bytes in
10565 size for static RSA ciphersuites) as well as client server and random
10568 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10571 *Stephen Henson, reported by UK NISCC*
10573 * Use Windows randomness collection on Cygwin.
10577 * Fix hang in EGD/PRNGD query when communication socket is closed
10578 prematurely by EGD/PRNGD.
10580 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10582 * Prompt for pass phrases when appropriate for PKCS12 input format.
10586 * Back-port of selected performance improvements from development
10587 branch, as well as improved support for PowerPC platforms.
10591 * Add lots of checks for memory allocation failure, error codes to indicate
10592 failure and freeing up memory if a failure occurs.
10594 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10596 * Add new -passin argument to dgst.
10600 * Perform some character comparisons of different types in X509_NAME_cmp:
10601 this is needed for some certificates that re-encode DNs into UTF8Strings
10602 (in violation of RFC3280) and can't or won't issue name rollover
10607 * Make an explicit check during certificate validation to see that
10608 the CA setting in each certificate on the chain is correct. As a
10609 side effect always do the following basic checks on extensions,
10610 not just when there's an associated purpose to the check:
10612 - if there is an unhandled critical extension (unless the user
10613 has chosen to ignore this fault)
10614 - if the path length has been exceeded (if one is set at all)
10615 - that certain extensions fit the associated purpose (if one has
10620 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004] ###
10622 * Avoid a race condition when CRLs are checked in a multi threaded
10623 environment. This would happen due to the reordering of the revoked
10624 entries during signature checking and serial number lookup. Now the
10625 encoding is cached and the serial number sort performed under a lock.
10626 Add new STACK function sk_is_sorted().
10630 * Add Delta CRL to the extension code.
10634 * Various fixes to s3_pkt.c so alerts are sent properly.
10636 *David Holmes <d.holmes@f5.com>*
10638 * Reduce the chances of duplicate issuer name and serial numbers (in
10639 violation of RFC3280) using the OpenSSL certificate creation utilities.
10640 This is done by creating a random 64 bit value for the initial serial
10641 number when a serial number file is created or when a self signed
10642 certificate is created using 'openssl req -x509'. The initial serial
10643 number file is created using 'openssl x509 -next_serial' in CA.pl
10644 rather than being initialized to 1.
10648 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004] ###
10650 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10651 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10653 *Joe Orton, Steve Henson*
10655 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10658 *Joe Orton, Steve Henson*
10660 * Make it possible to have multiple active certificates with the same
10661 subject in the CA index file. This is done only if the keyword
10662 'unique_subject' is set to 'no' in the main CA section (default
10663 if 'CA_default') of the configuration file. The value is saved
10664 with the database itself in a separate index attribute file,
10665 named like the index file with '.attr' appended to the name.
10669 * X509 verify fixes. Disable broken certificate workarounds when
10670 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10671 keyUsage extension present. Don't accept CRLs with unhandled critical
10672 extensions: since verify currently doesn't process CRL extensions this
10673 rejects a CRL with *any* critical extensions. Add new verify error codes
10678 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10679 A clarification of RFC2560 will require the use of OCTET STRINGs and
10680 some implementations cannot handle the current raw format. Since OpenSSL
10681 copies and compares OCSP nonces as opaque blobs without any attempt at
10682 parsing them this should not create any compatibility issues.
10686 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10687 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10688 this HMAC (and other) operations are several times slower than OpenSSL
10693 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10695 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10697 * Use the correct content when signing type "other".
10701 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003] ###
10703 * Fix various bugs revealed by running the NISCC test suite:
10705 Stop out of bounds reads in the ASN1 code when presented with
10706 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10708 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10710 If verify callback ignores invalid public key errors don't try to check
10711 certificate signature with the NULL public key.
10716 * New -ignore_err option in ocsp application to stop the server
10717 exiting on the first error in a request.
10721 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10722 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10727 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10728 extra data after the compression methods not only for TLS 1.0
10729 but also for SSL 3.0 (as required by the specification).
10731 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10733 * Change X509_certificate_type() to mark the key as exported/exportable
10734 when it's 512 *bits* long, not 512 bytes.
10738 * Change AES_cbc_encrypt() so it outputs exact multiple of
10739 blocks during encryption.
10743 * Various fixes to base64 BIO and non blocking I/O. On write
10744 flushes were not handled properly if the BIO retried. On read
10745 data was not being buffered properly and had various logic bugs.
10746 This also affects blocking I/O when the data being decoded is a
10751 * Various S/MIME bugfixes and compatibility changes:
10752 output correct application/pkcs7 MIME type if
10753 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10754 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10755 of files as .eml work). Correctly handle very long lines in MIME
10760 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003] ###
10762 * Countermeasure against the Klima-Pokorny-Rosa extension of
10763 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10764 a protocol version number mismatch like a decryption error
10765 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10769 * Turn on RSA blinding by default in the default implementation
10770 to avoid a timing attack. Applications that don't want it can call
10771 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10772 They would be ill-advised to do so in most cases.
10774 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10776 * Change RSA blinding code so that it works when the PRNG is not
10777 seeded (in this case, the secret RSA exponent is abused as
10778 an unpredictable seed -- if it is not unpredictable, there
10779 is no point in blinding anyway). Make RSA blinding thread-safe
10780 by remembering the creator's thread ID in rsa->blinding and
10781 having all other threads use local one-time blinding factors
10782 (this requires more computation than sharing rsa->blinding, but
10783 avoids excessive locking; and if an RSA object is not shared
10784 between threads, blinding will still be very fast).
10788 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10789 ENGINE as defaults for all supported algorithms irrespective of
10790 the 'flags' parameter. 'flags' is now honoured, so applications
10791 should make sure they are passing it correctly.
10795 * Target "mingw" now allows native Windows code to be generated in
10796 the Cygwin environment as well as with the MinGW compiler.
10800 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003] ###
10802 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10803 via timing by performing a MAC computation even if incorrect
10804 block cipher padding has been found. This is a countermeasure
10805 against active attacks where the attacker has to distinguish
10806 between bad padding and a MAC verification error. [CVE-2003-0078][]
10808 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10809 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10810 Martin Vuagnoux (EPFL, Ilion)*
10812 * Make the no-err option work as intended. The intention with no-err
10813 is not to have the whole error stack handling routines removed from
10814 libcrypto, it's only intended to remove all the function name and
10815 reason texts, thereby removing some of the footprint that may not
10816 be interesting if those errors aren't displayed anyway.
10818 NOTE: it's still possible for any application or module to have its
10819 own set of error texts inserted. The routines are there, just not
10820 used by default when no-err is given.
10824 * Add support for FreeBSD on IA64.
10826 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10828 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10829 Kerberos function mit_des_cbc_cksum(). Before this change,
10830 the value returned by DES_cbc_cksum() was like the one from
10831 mit_des_cbc_cksum(), except the bytes were swapped.
10833 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10835 * Allow an application to disable the automatic SSL chain building.
10836 Before this a rather primitive chain build was always performed in
10837 ssl3_output_cert_chain(): an application had no way to send the
10838 correct chain if the automatic operation produced an incorrect result.
10840 Now the chain builder is disabled if either:
10842 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10844 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10846 The reasoning behind this is that an application would not want the
10847 auto chain building to take place if extra chain certificates are
10848 present and it might also want a means of sending no additional
10849 certificates (for example the chain has two certificates and the
10854 * Add the possibility to build without the ENGINE framework.
10856 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10858 * Under Win32 gmtime() can return NULL: check return value in
10859 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10863 * DSA routines: under certain error conditions uninitialized BN objects
10864 could be freed. Solution: make sure initialization is performed early
10865 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10866 Nils Larsch <nla@trustcenter.de> via PR#459)
10870 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10871 checked on reconnect on the client side, therefore session resumption
10872 could still fail with a "ssl session id is different" error. This
10873 behaviour is masked when SSL_OP_ALL is used due to
10874 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10875 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10876 followup to PR #377.
10880 * IA-32 assembler support enhancements: unified ELF targets, support
10881 for SCO/Caldera platforms, fix for Cygwin shared build.
10885 * Add support for FreeBSD on sparc64. As a consequence, support for
10886 FreeBSD on non-x86 processors is separate from x86 processors on
10887 the config script, much like the NetBSD support.
10889 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10891 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002] ###
10893 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10896 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10897 code (06) was taken as the first octet of the session ID and the last
10898 octet was ignored consequently. As a result SSLv2 client side session
10899 caching could not have worked due to the session ID mismatch between
10901 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10906 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10907 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10912 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10913 seems that in spite of existing for more than a year, many application
10914 author have done nothing to provide the necessary callbacks, which
10915 means that this particular engine will not work properly anywhere.
10916 This is a very unfortunate situation which forces us, in the name
10917 of usability, to give the hw_ncipher.c a static lock, which is part
10919 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10920 appear in 0.9.8 or later. We EXPECT application authors to have
10921 dealt properly with this when 0.9.8 is released (unless we actually
10922 make such changes in the libcrypto locking code that changes will
10923 have to be made anyway).
10927 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
10928 octets have been read, EOF or an error occurs. Without this change
10929 some truncated ASN1 structures will not produce an error.
10933 * Disable Heimdal support, since it hasn't been fully implemented.
10934 Still give the possibility to force the use of Heimdal, but with
10935 warnings and a request that patches get sent to openssl-dev.
10939 * Add the VC-CE target, introduce the WINCE sysname, and add
10940 INSTALL.WCE and appropriate conditionals to make it build.
10942 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10944 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
10945 cygssl-x.y.z.dll, where x, y and z are the major, minor and
10946 edit numbers of the version.
10948 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10950 * Introduce safe string copy and catenation functions
10951 (BUF_strlcpy() and BUF_strlcat()).
10953 *Ben Laurie (CHATS) and Richard Levitte*
10955 * Avoid using fixed-size buffers for one-line DNs.
10957 *Ben Laurie (CHATS)*
10959 * Add BUF_MEM_grow_clean() to avoid information leakage when
10960 resizing buffers containing secrets, and use where appropriate.
10962 *Ben Laurie (CHATS)*
10964 * Avoid using fixed size buffers for configuration file location.
10966 *Ben Laurie (CHATS)*
10968 * Avoid filename truncation for various CA files.
10970 *Ben Laurie (CHATS)*
10972 * Use sizeof in preference to magic numbers.
10974 *Ben Laurie (CHATS)*
10976 * Avoid filename truncation in cert requests.
10978 *Ben Laurie (CHATS)*
10980 * Add assertions to check for (supposedly impossible) buffer
10983 *Ben Laurie (CHATS)*
10985 * Don't cache truncated DNS entries in the local cache (this could
10986 potentially lead to a spoofing attack).
10988 *Ben Laurie (CHATS)*
10990 * Fix various buffers to be large enough for hex/decimal
10991 representations in a platform independent manner.
10993 *Ben Laurie (CHATS)*
10995 * Add CRYPTO_realloc_clean() to avoid information leakage when
10996 resizing buffers containing secrets, and use where appropriate.
10998 *Ben Laurie (CHATS)*
11000 * Add BIO_indent() to avoid much slightly worrying code to do
11003 *Ben Laurie (CHATS)*
11005 * Convert sprintf()/BIO_puts() to BIO_printf().
11007 *Ben Laurie (CHATS)*
11009 * buffer_gets() could terminate with the buffer only half
11012 *Ben Laurie (CHATS)*
11014 * Add assertions to prevent user-supplied crypto functions from
11015 overflowing internal buffers by having large block sizes, etc.
11017 *Ben Laurie (CHATS)*
11019 * New OPENSSL_assert() macro (similar to assert(), but enabled
11022 *Ben Laurie (CHATS)*
11024 * Eliminate unused copy of key in RC4.
11026 *Ben Laurie (CHATS)*
11028 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11030 *Ben Laurie (CHATS)*
11032 * Fix off-by-one error in EGD path.
11034 *Ben Laurie (CHATS)*
11036 * If RANDFILE path is too long, ignore instead of truncating.
11038 *Ben Laurie (CHATS)*
11040 * Eliminate unused and incorrectly sized X.509 structure
11043 *Ben Laurie (CHATS)*
11045 * Eliminate unused and dangerous function knumber().
11047 *Ben Laurie (CHATS)*
11049 * Eliminate unused and dangerous structure, KSSL_ERR.
11051 *Ben Laurie (CHATS)*
11053 * Protect against overlong session ID context length in an encoded
11054 session object. Since these are local, this does not appear to be
11057 *Ben Laurie (CHATS)*
11059 * Change from security patch (see 0.9.6e below) that did not affect
11060 the 0.9.6 release series:
11062 Remote buffer overflow in SSL3 protocol - an attacker could
11063 supply an oversized master key in Kerberos-enabled versions.
11066 *Ben Laurie (CHATS)*
11068 * Change the SSL kerb5 codes to match RFC 2712.
11072 * Make -nameopt work fully for req and add -reqopt switch.
11074 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11076 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11078 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11080 * Make sure tests can be performed even if the corresponding algorithms
11081 have been removed entirely. This was also the last step to make
11082 OpenSSL compilable with DJGPP under all reasonable conditions.
11084 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11086 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11087 to allow version independent disabling of normally unselected ciphers,
11088 which may be activated as a side-effect of selecting a single cipher.
11090 (E.g., cipher list string "RSA" enables ciphersuites that are left
11091 out of "ALL" because they do not provide symmetric encryption.
11092 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11094 *Lutz Jaenicke, Bodo Moeller*
11096 * Add appropriate support for separate platform-dependent build
11097 directories. The recommended way to make a platform-dependent
11098 build directory is the following (tested on Linux), maybe with
11101 # Place yourself outside of the OpenSSL source tree. In
11102 # this example, the environment variable OPENSSL_SOURCE
11103 # is assumed to contain the absolute OpenSSL source directory.
11104 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11105 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11106 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11107 mkdir -p `dirname $F`
11108 ln -s $OPENSSL_SOURCE/$F $F
11111 To be absolutely sure not to disturb the source tree, a "make clean"
11112 is a good thing. If it isn't successful, don't worry about it,
11113 it probably means the source directory is very clean.
11117 * Make sure any ENGINE control commands make local copies of string
11118 pointers passed to them whenever necessary. Otherwise it is possible
11119 the caller may have overwritten (or deallocated) the original string
11120 data when a later ENGINE operation tries to use the stored values.
11122 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11124 * Improve diagnostics in file reading and command-line digests.
11126 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11128 * Add AES modes CFB and OFB to the object database. Correct an
11129 error in AES-CFB decryption.
11133 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11134 allows existing EVP_CIPHER_CTX structures to be reused after
11135 calling EVP_*Final(). This behaviour is used by encryption
11136 BIOs and some applications. This has the side effect that
11137 applications must explicitly clean up cipher contexts with
11138 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11142 * Check the values of dna and dnb in bn_mul_recursive before calling
11143 bn_mul_comba (a non zero value means the a or b arrays do not contain
11144 n2 elements) and fallback to bn_mul_normal if either is not zero.
11148 * Fix escaping of non-ASCII characters when using the -subj option
11149 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11153 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11154 form for "surname", serialNumber has no short form.
11155 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11156 therefore remove "mail" short name for "internet 7".
11157 The OID for unique identifiers in X509 certificates is
11158 x500UniqueIdentifier, not uniqueIdentifier.
11159 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11163 * Add an "init" command to the ENGINE config module and auto initialize
11164 ENGINEs. Without any "init" command the ENGINE will be initialized
11165 after all ctrl commands have been executed on it. If init=1 the
11166 ENGINE is initialized at that point (ctrls before that point are run
11167 on the uninitialized ENGINE and after on the initialized one). If
11168 init=0 then the ENGINE will not be initialized at all.
11172 * Fix the 'app_verify_callback' interface so that the user-defined
11173 argument is actually passed to the callback: In the
11174 SSL_CTX_set_cert_verify_callback() prototype, the callback
11175 declaration has been changed from
11178 int (*cb)(X509_STORE_CTX *,void *);
11179 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11180 i=s->ctx->app_verify_callback(&ctx)
11181 has been changed into
11182 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11184 To update applications using SSL_CTX_set_cert_verify_callback(),
11185 a dummy argument can be added to their callback functions.
11187 *D. K. Smetters <smetters@parc.xerox.com>*
11189 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11191 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11193 * Add and OPENSSL_LOAD_CONF define which will cause
11194 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11195 This allows older applications to transparently support certain
11196 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11197 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11198 load the config file and OPENSSL_add_all_algorithms_conf() which will
11199 always load it have also been added.
11203 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11204 Adjust NIDs and EVP layer.
11206 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11208 * Config modules support in openssl utility.
11210 Most commands now load modules from the config file,
11211 though in a few (such as version) this isn't done
11212 because it couldn't be used for anything.
11214 In the case of ca and req the config file used is
11215 the same as the utility itself: that is the -config
11216 command line option can be used to specify an
11221 * Move default behaviour from OPENSSL_config(). If appname is NULL
11222 use "openssl_conf" if filename is NULL use default openssl config file.
11226 * Add an argument to OPENSSL_config() to allow the use of an alternative
11227 config section name. Add a new flag to tolerate a missing config file
11228 and move code to CONF_modules_load_file().
11232 * Support for crypto accelerator cards from Accelerated Encryption
11233 Processing, www.aep.ie. (Use engine 'aep')
11234 The support was copied from 0.9.6c [engine] and adapted/corrected
11235 to work with the new engine framework.
11237 *AEP Inc. and Richard Levitte*
11239 * Support for SureWare crypto accelerator cards from Baltimore
11240 Technologies. (Use engine 'sureware')
11241 The support was copied from 0.9.6c [engine] and adapted
11242 to work with the new engine framework.
11246 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11247 make the newer ENGINE framework commands for the CHIL engine work.
11249 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11251 * Make it possible to produce shared libraries on ReliantUNIX.
11253 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11255 * Add the configuration target debug-linux-ppro.
11256 Make 'openssl rsa' use the general key loading routines
11257 implemented in apps.c, and make those routines able to
11258 handle the key format FORMAT_NETSCAPE and the variant
11261 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11263 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11265 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11267 * Add -keyform to rsautl, and document -engine.
11269 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11271 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11272 BIO_R_NO_SUCH_FILE error code rather than the generic
11273 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11277 * Add new functions
11278 ERR_peek_last_error
11279 ERR_peek_last_error_line
11280 ERR_peek_last_error_line_data.
11281 These are similar to
11283 ERR_peek_error_line
11284 ERR_peek_error_line_data,
11285 but report on the latest error recorded rather than the first one
11286 still in the error queue.
11288 *Ben Laurie, Bodo Moeller*
11290 * default_algorithms option in ENGINE config module. This allows things
11292 default_algorithms = ALL
11293 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11297 * Preliminary ENGINE config module.
11301 * New experimental application configuration code.
11305 * Change the AES code to follow the same name structure as all other
11306 symmetric ciphers, and behave the same way. Move everything to
11307 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11309 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11311 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11313 *Ben Laurie and Theo de Raadt*
11315 * Add option to output public keys in req command.
11317 *Massimiliano Pala madwolf@openca.org*
11319 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11320 (up to about 10% better than before for P-192 and P-224).
11324 * New functions/macros
11326 SSL_CTX_set_msg_callback(ctx, cb)
11327 SSL_CTX_set_msg_callback_arg(ctx, arg)
11328 SSL_set_msg_callback(ssl, cb)
11329 SSL_set_msg_callback_arg(ssl, arg)
11331 to request calling a callback function
11333 void cb(int write_p, int version, int content_type,
11334 const void *buf, size_t len, SSL *ssl, void *arg)
11336 whenever a protocol message has been completely received
11337 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11338 protocol version according to which the SSL library interprets
11339 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11340 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11341 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11342 specification (change_cipher_spec(20), alert(21), handshake(22)).
11343 'buf' and 'len' point to the actual message, 'ssl' to the
11344 SSL object, and 'arg' is the application-defined value set by
11345 SSL[_CTX]_set_msg_callback_arg().
11347 'openssl s_client' and 'openssl s_server' have new '-msg' options
11348 to enable a callback that displays all protocol messages.
11352 * Change the shared library support so shared libraries are built as
11353 soon as the corresponding static library is finished, and thereby get
11354 openssl and the test programs linked against the shared library.
11355 This still only happens when the keyword "shard" has been given to
11356 the configuration scripts.
11358 NOTE: shared library support is still an experimental thing, and
11359 backward binary compatibility is still not guaranteed.
11361 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11363 * Add support for Subject Information Access extension.
11365 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11367 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11368 additional bytes when new memory had to be allocated, not just
11369 when reusing an existing buffer.
11373 * New command line and configuration option 'utf8' for the req command.
11374 This allows field values to be specified as UTF8 strings.
11378 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11379 runs for the former and machine-readable output for the latter.
11383 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11384 of the e-mail address in the DN (i.e., it will go into a certificate
11385 extension only). The new configuration file option 'email_in_dn = no'
11386 has the same effect.
11388 *Massimiliano Pala madwolf@openca.org*
11390 * Change all functions with names starting with des_ to be starting
11391 with DES_ instead. Add wrappers that are compatible with libdes,
11392 but are named _ossl_old_des_*. Finally, add macros that map the
11393 des_* symbols to the corresponding _ossl_old_des_* if libdes
11394 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11395 desired, the des_* symbols will be mapped to DES_*, with one
11398 Since we provide two compatibility mappings, the user needs to
11399 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11400 compatibility is desired. The default (i.e., when that macro
11401 isn't defined) is OpenSSL 0.9.6c compatibility.
11403 There are also macros that enable and disable the support of old
11404 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11405 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11406 are defined, the default will apply: to support the old des routines.
11408 In either case, one must include openssl/des.h to get the correct
11409 definitions. Do not try to just include openssl/des_old.h, that
11412 NOTE: This is a major break of an old API into a new one. Software
11413 authors are encouraged to switch to the DES_ style functions. Some
11414 time in the future, des_old.h and the libdes compatibility functions
11415 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11416 default), and then completely removed.
11420 * Test for certificates which contain unsupported critical extensions.
11421 If such a certificate is found during a verify operation it is
11422 rejected by default: this behaviour can be overridden by either
11423 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11424 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11425 X509_supported_extension() has also been added which returns 1 if a
11426 particular extension is supported.
11430 * Modify the behaviour of EVP cipher functions in similar way to digests
11431 to retain compatibility with existing code.
11435 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11436 compatibility with existing code. In particular the 'ctx' parameter does
11437 not have to be to be initialized before the call to EVP_DigestInit() and
11438 it is tidied up after a call to EVP_DigestFinal(). New function
11439 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11440 EVP_MD_CTX_copy() changed to not require the destination to be
11441 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11442 requires the destination to be valid.
11444 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11445 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11449 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11450 so that complete 'Handshake' protocol structures are kept in memory
11451 instead of overwriting 'msg_type' and 'length' with 'body' data.
11455 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11457 *Massimo Santin via Richard Levitte*
11459 * Major restructuring to the underlying ENGINE code. This includes
11460 reduction of linker bloat, separation of pure "ENGINE" manipulation
11461 (initialisation, etc) from functionality dealing with implementations
11462 of specific crypto interfaces. This change also introduces integrated
11463 support for symmetric ciphers and digest implementations - so ENGINEs
11464 can now accelerate these by providing EVP_CIPHER and EVP_MD
11465 implementations of their own. This is detailed in crypto/engine/README
11466 as it couldn't be adequately described here. However, there are a few
11467 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11468 were changed in the original introduction of ENGINE code have now
11469 reverted back - the hooking from this code to ENGINE is now a good
11470 deal more passive and at run-time, operations deal directly with
11471 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11472 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11473 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
11474 they were not being used by the framework as there is no concept of a
11475 BIGNUM_METHOD and they could not be generalised to the new
11476 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11477 ENGINE_cpy() has been removed as it cannot be consistently defined in
11482 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11486 * Change mkdef.pl to sort symbols that get the same entry number,
11487 and make sure the automatically generated functions ERR_load_*
11488 become part of libeay.num as well.
11492 * New function SSL_renegotiate_pending(). This returns true once
11493 renegotiation has been requested (either SSL_renegotiate() call
11494 or HelloRequest/ClientHello received from the peer) and becomes
11495 false once a handshake has been completed.
11496 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11497 sends a HelloRequest, but does not ensure that a handshake takes
11498 place. SSL_renegotiate_pending() is useful for checking if the
11499 client has followed the request.)
11503 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11504 By default, clients may request session resumption even during
11505 renegotiation (if session ID contexts permit); with this option,
11506 session resumption is possible only in the first handshake.
11508 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11509 more bits available for options that should not be part of
11510 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11514 * Add some demos for certificate and certificate request creation.
11518 * Make maximum certificate chain size accepted from the peer application
11519 settable (SSL*_get/set_max_cert_list()), as proposed by
11520 "Douglas E. Engert" <deengert@anl.gov>.
11524 * Add support for shared libraries for Unixware-7
11525 (Boyd Lynn Gerber <gerberb@zenez.com>).
11529 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11530 be done prior to destruction. Use this to unload error strings from
11531 ENGINEs that load their own error strings. NB: This adds two new API
11532 functions to "get" and "set" this destroy handler in an ENGINE.
11536 * Alter all existing ENGINE implementations (except "openssl" and
11537 "openbsd") to dynamically instantiate their own error strings. This
11538 makes them more flexible to be built both as statically-linked ENGINEs
11539 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11540 Also, add stub code to each that makes building them as self-contained
11541 shared-libraries easier (see README.ENGINE).
11545 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11546 implementations into applications that are completely implemented in
11547 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11548 commands that can be used to configure what shared-library to load and
11549 to control aspects of the way it is handled. Also, made an update to
11550 the README.ENGINE file that brings its information up-to-date and
11551 provides some information and instructions on the "dynamic" ENGINE
11552 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11556 * Make it possible to unload ranges of ERR strings with a new
11557 "ERR_unload_strings" function.
11561 * Add a copy() function to EVP_MD.
11565 * Make EVP_MD routines take a context pointer instead of just the
11566 md_data void pointer.
11570 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11571 that the digest can only process a single chunk of data
11572 (typically because it is provided by a piece of
11573 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11574 is only going to provide a single chunk of data, and hence the
11575 framework needn't accumulate the data for oneshot drivers.
11579 * As with "ERR", make it possible to replace the underlying "ex_data"
11580 functions. This change also alters the storage and management of global
11581 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11582 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11583 index counters. The API functions that use this state have been changed
11584 to take a "class_index" rather than pointers to the class's local STACK
11585 and counter, and there is now an API function to dynamically create new
11586 classes. This centralisation allows us to (a) plug a lot of the
11587 thread-safety problems that existed, and (b) makes it possible to clean
11588 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11589 such data would previously have always leaked in application code and
11590 workarounds were in place to make the memory debugging turn a blind eye
11591 to it. Application code that doesn't use this new function will still
11592 leak as before, but their memory debugging output will announce it now
11593 rather than letting it slide.
11595 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11596 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11597 has a return value to indicate success or failure.
11601 * Make it possible to replace the underlying "ERR" functions such that the
11602 global state (2 LHASH tables and 2 locks) is only used by the "default"
11603 implementation. This change also adds two functions to "get" and "set"
11604 the implementation prior to it being automatically set the first time
11605 any other ERR function takes place. Ie. an application can call "get",
11606 pass the return value to a module it has just loaded, and that module
11607 can call its own "set" function using that value. This means the
11608 module's "ERR" operations will use (and modify) the error state in the
11609 application and not in its own statically linked copy of OpenSSL code.
11613 * Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
11614 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11615 the operation, and provides a more encapsulated way for external code
11616 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11617 to use these functions rather than manually incrementing the counts.
11619 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11623 * Add EVP test program.
11627 * Add symmetric cipher support to ENGINE. Expect the API to change!
11631 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11632 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11633 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11634 These allow a CRL to be built without having to access X509_CRL fields
11635 directly. Modify 'ca' application to use new functions.
11639 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11640 bug workarounds. Rollback attack detection is a security feature.
11641 The problem will only arise on OpenSSL servers when TLSv1 is not
11642 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11643 Software authors not wanting to support TLSv1 will have special reasons
11644 for their choice and can explicitly enable this option.
11646 *Bodo Moeller, Lutz Jaenicke*
11648 * Rationalise EVP so it can be extended: don't include a union of
11649 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11650 (similar to those existing for EVP_CIPHER_CTX).
11655 EVP_MD_CTX_init(&md); /* new function call */
11656 EVP_DigestInit(&md, EVP_sha1());
11657 EVP_DigestUpdate(&md, in, len);
11658 EVP_DigestFinal(&md, out, NULL);
11659 EVP_MD_CTX_cleanup(&md); /* new function call */
11664 * Make DES key schedule conform to the usual scheme, as well as
11665 correcting its structure. This means that calls to DES functions
11666 now have to pass a pointer to a des_key_schedule instead of a
11667 plain des_key_schedule (which was actually always a pointer
11670 des_key_schedule ks;
11672 des_set_key_checked(..., &ks);
11673 des_ncbc_encrypt(..., &ks, ...);
11675 (Note that a later change renames 'des_...' into 'DES_...'.)
11679 * Initial reduction of linker bloat: the use of some functions, such as
11680 PEM causes large amounts of unused functions to be linked in due to
11681 poor organisation. For example pem_all.c contains every PEM function
11682 which has a knock on effect of linking in large amounts of (unused)
11683 ASN1 code. Grouping together similar functions and splitting unrelated
11684 functions prevents this.
11688 * Cleanup of EVP macros.
11692 * Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
11693 correct _ecb suffix.
11697 * Add initial OCSP responder support to ocsp application. The
11698 revocation information is handled using the text based index
11699 use by the ca application. The responder can either handle
11700 requests generated internally, supplied in files (for example
11701 via a CGI script) or using an internal minimal server.
11705 * Add configuration choices to get zlib compression for TLS.
11709 * Changes to Kerberos SSL for RFC 2712 compliance:
11710 1. Implemented real KerberosWrapper, instead of just using
11711 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11712 2. Implemented optional authenticator field of KerberosWrapper.
11714 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11715 and authenticator structs; see crypto/krb5/.
11717 Generalized Kerberos calls to support multiple Kerberos libraries.
11718 *Vern Staats <staatsvr@asc.hpc.mil>,
11719 Jeffrey Altman <jaltman@columbia.edu>
11720 via Richard Levitte*
11722 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11723 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11724 values for each of the key sizes rather than having just
11725 parameters (and 'speed' generating keys each time).
11729 * Speed up EVP routines.
11732 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11733 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11734 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11735 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11737 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11738 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11739 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11742 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11744 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11748 * Added the OS2-EMX target.
11750 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11752 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11753 to support NCONF routines in extension code. New function CONF_set_nconf()
11754 to allow functions which take an NCONF to also handle the old LHASH
11755 structure: this means that the old CONF compatible routines can be
11756 retained (in particular wrt extensions) without having to duplicate the
11757 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11761 * Enhance the general user interface with mechanisms for inner control
11762 and with possibilities to have yes/no kind of prompts.
11766 * Change all calls to low level digest routines in the library and
11767 applications to use EVP. Add missing calls to HMAC_cleanup() and
11768 don't assume HMAC_CTX can be copied using memcpy().
11770 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11772 * Add the possibility to control engines through control names but with
11773 arbitrary arguments instead of just a string.
11774 Change the key loaders to take a UI_METHOD instead of a callback
11775 function pointer. NOTE: this breaks binary compatibility with earlier
11776 versions of OpenSSL [engine].
11777 Adapt the nCipher code for these new conditions and add a card insertion
11782 * Enhance the general user interface with mechanisms to better support
11783 dialog box interfaces, application-defined prompts, the possibility
11784 to use defaults (for example default passwords from somewhere else)
11785 and interrupts/cancellations.
11789 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11790 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11794 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11795 tidy up some unnecessarily weird code in 'sk_new()').
11797 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11799 * Change the key loading routines for ENGINEs to use the same kind
11800 callback (pem_password_cb) as all other routines that need this
11805 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11806 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11807 than this minimum value is recommended.
11811 * New random seeder for OpenVMS, using the system process statistics
11812 that are easily reachable.
11816 * Windows apparently can't transparently handle global
11817 variables defined in DLLs. Initialisations such as:
11819 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11821 won't compile. This is used by the any applications that need to
11822 declare their own ASN1 modules. This was fixed by adding the option
11823 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11824 needed for static libraries under Win32.
11828 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11829 setting of purpose and trust fields. New X509_STORE trust and
11830 purpose functions and tidy up setting in other SSL functions.
11834 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11835 structure. These are inherited by X509_STORE_CTX when it is
11836 initialised. This allows various defaults to be set in the
11837 X509_STORE structure (such as flags for CRL checking and custom
11838 purpose or trust settings) for functions which only use X509_STORE_CTX
11839 internally such as S/MIME.
11841 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11842 trust settings if they are not set in X509_STORE. This allows X509_STORE
11843 purposes and trust (in S/MIME for example) to override any set by default.
11845 Add command line options for CRL checking to smime, s_client and s_server
11850 * Initial CRL based revocation checking. If the CRL checking flag(s)
11851 are set then the CRL is looked up in the X509_STORE structure and
11852 its validity and signature checked, then if the certificate is found
11853 in the CRL the verify fails with a revoked error.
11855 Various new CRL related callbacks added to X509_STORE_CTX structure.
11857 Command line options added to 'verify' application to support this.
11859 This needs some additional work, such as being able to handle multiple
11860 CRLs with different times, extension based lookup (rather than just
11861 by subject name) and ultimately more complete V2 CRL extension
11866 * Add a general user interface API (crypto/ui/). This is designed
11867 to replace things like des_read_password and friends (backward
11868 compatibility functions using this new API are provided).
11869 The purpose is to remove prompting functions from the DES code
11870 section as well as provide for prompting through dialog boxes in
11871 a window system and the like.
11875 * Add "ex_data" support to ENGINE so implementations can add state at a
11876 per-structure level rather than having to store it globally.
11880 * Make it possible for ENGINE structures to be copied when retrieved by
11881 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11882 This causes the "original" ENGINE structure to act like a template,
11883 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11884 operational state can be localised to each ENGINE structure, despite the
11885 fact they all share the same "methods". New ENGINE structures returned in
11886 this case have no functional references and the return value is the single
11887 structural reference. This matches the single structural reference returned
11888 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11893 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11894 needs to match any other type at all we need to manually clear the
11899 * Changes to the "openssl engine" utility to include;
11900 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11901 about an ENGINE's available control commands.
11902 - executing control commands from command line arguments using the
11903 '-pre' and '-post' switches. '-post' is only used if '-t' is
11904 specified and the ENGINE is successfully initialised. The syntax for
11905 the individual commands are colon-separated, for example;
11906 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11910 * New dynamic control command support for ENGINEs. ENGINEs can now
11911 declare their own commands (numbers), names (strings), descriptions,
11912 and input types for run-time discovery by calling applications. A
11913 subset of these commands are implicitly classed as "executable"
11914 depending on their input type, and only these can be invoked through
11915 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11916 can be based on user input, config files, etc). The distinction is
11917 that "executable" commands cannot return anything other than a boolean
11918 result and can only support numeric or string input, whereas some
11919 discoverable commands may only be for direct use through
11920 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11921 pointers, or other custom uses. The "executable" commands are to
11922 support parameterisations of ENGINE behaviour that can be
11923 unambiguously defined by ENGINEs and used consistently across any
11924 OpenSSL-based application. Commands have been added to all the
11925 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
11926 control over shared-library paths without source code alterations.
11930 * Changed all ENGINE implementations to dynamically allocate their
11931 ENGINEs rather than declaring them statically. Apart from this being
11932 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
11933 this also allows the implementations to compile without using the
11934 internal engine_int.h header.
11938 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
11939 'const' value. Any code that should be able to modify a RAND_METHOD
11940 should already have non-const pointers to it (ie. they should only
11941 modify their own ones).
11945 * Made a variety of little tweaks to the ENGINE code.
11946 - "atalla" and "ubsec" string definitions were moved from header files
11947 to C code. "nuron" string definitions were placed in variables
11948 rather than hard-coded - allowing parameterisation of these values
11949 later on via ctrl() commands.
11950 - Removed unused "#if 0"'d code.
11951 - Fixed engine list iteration code so it uses ENGINE_free() to release
11952 structural references.
11953 - Constified the RAND_METHOD element of ENGINE structures.
11954 - Constified various get/set functions as appropriate and added
11955 missing functions (including a catch-all ENGINE_cpy that duplicates
11956 all ENGINE values onto a new ENGINE except reference counts/state).
11957 - Removed NULL parameter checks in get/set functions. Setting a method
11958 or function to NULL is a way of cancelling out a previously set
11959 value. Passing a NULL ENGINE parameter is just plain stupid anyway
11960 and doesn't justify the extra error symbols and code.
11961 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
11962 flags from engine_int.h to engine.h.
11963 - Changed prototypes for ENGINE handler functions (init(), finish(),
11964 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
11968 * Implement binary inversion algorithm for BN_mod_inverse in addition
11969 to the algorithm using long division. The binary algorithm can be
11970 used only if the modulus is odd. On 32-bit systems, it is faster
11971 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
11972 roughly 5-15% for 256-bit moduli), so we use it only for moduli
11973 up to 450 bits. In 64-bit environments, the binary algorithm
11974 appears to be advantageous for much longer moduli; here we use it
11975 for moduli up to 2048 bits.
11979 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
11980 could not support the combine flag in choice fields.
11984 * Add a 'copy_extensions' option to the 'ca' utility. This copies
11985 extensions from a certificate request to the certificate.
11989 * Allow multiple 'certopt' and 'nameopt' options to be separated
11990 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
11991 file: this allows the display of the certificate about to be
11992 signed to be customised, to allow certain fields to be included
11993 or excluded and extension details. The old system didn't display
11994 multicharacter strings properly, omitted fields not in the policy
11995 and couldn't display additional details such as extensions.
11999 * Function EC_POINTs_mul for multiple scalar multiplication
12000 of an arbitrary number of elliptic curve points
12001 \sum scalars[i]*points[i],
12002 optionally including the generator defined for the EC_GROUP:
12003 scalar*generator + \sum scalars[i]*points[i].
12005 EC_POINT_mul is a simple wrapper function for the typical case
12006 that the point list has just one item (besides the optional
12011 * First EC_METHODs for curves over GF(p):
12013 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12014 operations and provides various method functions that can also
12015 operate with faster implementations of modular arithmetic.
12017 EC_GFp_mont_method() reuses most functions that are part of
12018 EC_GFp_simple_method, but uses Montgomery arithmetic.
12020 *Bodo Moeller; point addition and point doubling
12021 implementation directly derived from source code provided by
12022 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12024 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12025 crypto/ec/ec_lib.c):
12027 Curves are EC_GROUP objects (with an optional group generator)
12028 based on EC_METHODs that are built into the library.
12030 Points are EC_POINT objects based on EC_GROUP objects.
12032 Most of the framework would be able to handle curves over arbitrary
12033 finite fields, but as there are no obvious types for fields other
12034 than GF(p), some functions are limited to that for now.
12038 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12039 that the file contains a complete HTTP response.
12043 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12044 change the def and num file printf format specifier from "%-40sXXX"
12045 to "%-39s XXX". The latter will always guarantee a space after the
12046 field while the former will cause them to run together if the field
12047 is 40 of more characters long.
12051 * Constify the cipher and digest 'method' functions and structures
12052 and modify related functions to take constant EVP_MD and EVP_CIPHER
12057 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12058 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12062 * Modify EVP_Digest*() routines so they now return values. Although the
12063 internal software routines can never fail additional hardware versions
12068 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12070 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12071 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12074 ERR_R_NESTED_ASN1_ERROR
12076 ERR_R_MISSING_ASN1_EOS
12077 were 4 .. 9, conflicting with
12078 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12080 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12081 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12083 Add new error code 'ERR_R_INTERNAL_ERROR'.
12087 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12092 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12093 sets the subject name for a new request or supersedes the
12094 subject name in a given request. Formats that can be parsed are
12095 'CN=Some Name, OU=myOU, C=IT'
12097 'CN=Some Name/OU=myOU/C=IT'.
12099 Add options '-batch' and '-verbose' to 'openssl req'.
12101 *Massimiliano Pala <madwolf@hackmasters.net>*
12103 * Introduce the possibility to access global variables through
12104 functions on platform were that's the best way to handle exporting
12105 global variables in shared libraries. To enable this functionality,
12106 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12107 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12108 is normally done by Configure or something similar).
12110 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12111 in the source file (foo.c) like this:
12113 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12114 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12116 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12117 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12119 OPENSSL_DECLARE_GLOBAL(int,foo);
12120 #define foo OPENSSL_GLOBAL_REF(foo)
12121 OPENSSL_DECLARE_GLOBAL(double,bar);
12122 #define bar OPENSSL_GLOBAL_REF(bar)
12124 The #defines are very important, and therefore so is including the
12125 header file everywhere where the defined globals are used.
12127 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12128 of ASN.1 items, but that structure is a bit different.
12130 The largest change is in util/mkdef.pl which has been enhanced with
12131 better and easier to understand logic to choose which symbols should
12132 go into the Windows .def files as well as a number of fixes and code
12133 cleanup (among others, algorithm keywords are now sorted
12134 lexicographically to avoid constant rewrites).
12138 * In BN_div() keep a copy of the sign of 'num' before writing the
12139 result to 'rm' because if rm==num the value will be overwritten
12140 and produce the wrong result if 'num' is negative: this caused
12141 problems with BN_mod() and BN_nnmod().
12145 * Function OCSP_request_verify(). This checks the signature on an
12146 OCSP request and verifies the signer certificate. The signer
12147 certificate is just checked for a generic purpose and OCSP request
12152 * Add OCSP_check_validity() function to check the validity of OCSP
12153 responses. OCSP responses are prepared in real time and may only
12154 be a few seconds old. Simply checking that the current time lies
12155 between thisUpdate and nextUpdate max reject otherwise valid responses
12156 caused by either OCSP responder or client clock inaccuracy. Instead
12157 we allow thisUpdate and nextUpdate to fall within a certain period of
12158 the current time. The age of the response can also optionally be
12159 checked. Two new options -validity_period and -status_age added to
12164 * If signature or public key algorithm is unrecognized print out its
12165 OID rather that just UNKNOWN.
12169 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12170 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12171 ID to be generated from the issuer certificate alone which can then be
12172 passed to OCSP_id_issuer_cmp().
12176 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12177 ASN1 modules to export functions returning ASN1_ITEM pointers
12178 instead of the ASN1_ITEM structures themselves. This adds several
12179 new macros which allow the underlying ASN1 function/structure to
12180 be accessed transparently. As a result code should not use ASN1_ITEM
12181 references directly (such as &X509_it) but instead use the relevant
12182 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12183 use of the new ASN1 code on platforms where exporting structures
12184 is problematical (for example in shared libraries) but exporting
12185 functions returning pointers to structures is not.
12189 * Add support for overriding the generation of SSL/TLS session IDs.
12190 These callbacks can be registered either in an SSL_CTX or per SSL.
12191 The purpose of this is to allow applications to control, if they wish,
12192 the arbitrary values chosen for use as session IDs, particularly as it
12193 can be useful for session caching in multiple-server environments. A
12194 command-line switch for testing this (and any client code that wishes
12195 to use such a feature) has been added to "s_server".
12197 *Geoff Thorpe, Lutz Jaenicke*
12199 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12200 of the form '#if defined(...) || defined(...) || ...' and
12201 '#if !defined(...) && !defined(...) && ...'. This also avoids
12202 the growing number of special cases it was previously handling.
12206 * Make all configuration macros available for application by making
12207 sure they are available in opensslconf.h, by giving them names starting
12208 with "OPENSSL_" to avoid conflicts with other packages and by making
12209 sure e_os2.h will cover all platform-specific cases together with
12211 Additionally, it is now possible to define configuration/platform-
12212 specific names (called "system identities"). In the C code, these
12213 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
12214 macro with the name beginning with "OPENSSL_SYS_", which is determined
12215 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
12220 * New option -set_serial to 'req' and 'x509' this allows the serial
12221 number to use to be specified on the command line. Previously self
12222 signed certificates were hard coded with serial number 0 and the
12223 CA options of 'x509' had to use a serial number in a file which was
12228 * New options to 'ca' utility to support V2 CRL entry extensions.
12229 Currently CRL reason, invalidity date and hold instruction are
12230 supported. Add new CRL extensions to V3 code and some new objects.
12234 * New function EVP_CIPHER_CTX_set_padding() this is used to
12235 disable standard block padding (aka PKCS#5 padding) in the EVP
12236 API, which was previously mandatory. This means that the data is
12237 not padded in any way and so the total length much be a multiple
12238 of the block size, otherwise an error occurs.
12242 * Initial (incomplete) OCSP SSL support.
12246 * New function OCSP_parse_url(). This splits up a URL into its host,
12247 port and path components: primarily to parse OCSP URLs. New -url
12248 option to ocsp utility.
12252 * New nonce behavior. The return value of OCSP_check_nonce() now
12253 reflects the various checks performed. Applications can decide
12254 whether to tolerate certain situations such as an absent nonce
12255 in a response when one was present in a request: the ocsp application
12256 just prints out a warning. New function OCSP_add1_basic_nonce()
12257 this is to allow responders to include a nonce in a response even if
12258 the request is nonce-less.
12262 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12263 skipped when using openssl x509 multiple times on a single input file,
12264 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
12268 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12269 set string type: to handle setting ASN1_TIME structures. Fix ca
12270 utility to correctly initialize revocation date of CRLs.
12274 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12275 the clients preferred ciphersuites and rather use its own preferences.
12276 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12277 Internet Explorer by ensuring unchanged hash method during stepup.
12278 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12282 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12283 to aes and add a new 'exist' option to print out symbols that don't
12288 * Additional options to ocsp utility to allow flags to be set and
12289 additional certificates supplied.
12293 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12294 OCSP client a number of certificate to only verify the response
12299 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12300 handle the new API. Currently only ECB, CBC modes supported. Add new
12303 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12304 Encryption Standard (AES) Ciphersuites for Transport Layer
12305 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12306 not enabled by default and were not part of the "ALL" ciphersuite
12307 alias because they were not yet official; they could be
12308 explicitly requested by specifying the "AESdraft" ciphersuite
12309 group alias. In the final release of OpenSSL 0.9.7, the group
12310 alias is called "AES" and is part of "ALL".)
12312 *Ben Laurie, Steve Henson, Bodo Moeller*
12314 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12315 request to response.
12319 * Functions for OCSP responders. OCSP_request_onereq_count(),
12320 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12321 extract information from a certificate request. OCSP_response_create()
12322 creates a response and optionally adds a basic response structure.
12323 OCSP_basic_add1_status() adds a complete single response to a basic
12324 response and returns the OCSP_SINGLERESP structure just added (to allow
12325 extensions to be included for example). OCSP_basic_add1_cert() adds a
12326 certificate to a basic response and OCSP_basic_sign() signs a basic
12327 response with various flags. New helper functions ASN1_TIME_check()
12328 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12329 (converts ASN1_TIME to GeneralizedTime).
12333 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12334 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12335 structure from a certificate. X509_pubkey_digest() digests the public_key
12336 contents: this is used in various key identifiers.
12340 * Make sk_sort() tolerate a NULL argument.
12342 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12344 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12345 passed by the function are trusted implicitly. If any of them signed the
12346 response then it is assumed to be valid and is not verified.
12350 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12351 to data. This was previously part of the PKCS7 ASN1 code. This
12352 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12353 *Steve Henson, reported by Kenneth R. Robinette
12354 <support@securenetterm.com>*
12356 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12357 routines: without these tracing memory leaks is very painful.
12358 Fix leaks in PKCS12 and PKCS7 routines.
12362 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12363 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12364 effectively meant GeneralizedTime would never be used. Now it
12365 is initialised to -1 but X509_time_adj() now has to check the value
12366 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12367 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12368 *Steve Henson, reported by Kenneth R. Robinette
12369 <support@securenetterm.com>*
12371 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12372 result in a zero length in the ASN1_INTEGER structure which was
12373 not consistent with the structure when d2i_ASN1_INTEGER() was used
12374 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12375 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12376 where it did not print out a minus for negative ASN1_INTEGER.
12380 * Add summary printout to ocsp utility. The various functions which
12381 convert status values to strings have been renamed to:
12382 OCSP_response_status_str(), OCSP_cert_status_str() and
12383 OCSP_crl_reason_str() and are no longer static. New options
12384 to verify nonce values and to disable verification. OCSP response
12385 printout format cleaned up.
12389 * Add additional OCSP certificate checks. These are those specified
12390 in RFC2560. This consists of two separate checks: the CA of the
12391 certificate being checked must either be the OCSP signer certificate
12392 or the issuer of the OCSP signer certificate. In the latter case the
12393 OCSP signer certificate must contain the OCSP signing extended key
12394 usage. This check is performed by attempting to match the OCSP
12395 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12396 in the OCSP_CERTID structures of the response.
12400 * Initial OCSP certificate verification added to OCSP_basic_verify()
12401 and related routines. This uses the standard OpenSSL certificate
12402 verify routines to perform initial checks (just CA validity) and
12403 to obtain the certificate chain. Then additional checks will be
12404 performed on the chain. Currently the root CA is checked to see
12405 if it is explicitly trusted for OCSP signing. This is used to set
12406 a root CA as a global signing root: that is any certificate that
12407 chains to that CA is an acceptable OCSP signing certificate.
12411 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12412 extensions from a separate configuration file.
12413 As when reading extensions from the main configuration file,
12414 the '-extensions ...' option may be used for specifying the
12417 *Massimiliano Pala <madwolf@comune.modena.it>*
12419 * New OCSP utility. Allows OCSP requests to be generated or
12420 read. The request can be sent to a responder and the output
12421 parsed, outputted or printed in text form. Not complete yet:
12422 still needs to check the OCSP response validity.
12426 * New subcommands for 'openssl ca':
12427 'openssl ca -status <serial>' prints the status of the cert with
12428 the given serial number (according to the index file).
12429 'openssl ca -updatedb' updates the expiry status of certificates
12432 *Massimiliano Pala <madwolf@comune.modena.it>*
12434 * New '-newreq-nodes' command option to CA.pl. This is like
12435 '-newreq', but calls 'openssl req' with the '-nodes' option
12436 so that the resulting key is not encrypted.
12438 *Damien Miller <djm@mindrot.org>*
12440 * New configuration for the GNU Hurd.
12442 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12444 * Initial code to implement OCSP basic response verify. This
12445 is currently incomplete. Currently just finds the signer's
12446 certificate and verifies the signature on the response.
12450 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12451 value of OPENSSLDIR. This is available via the new '-d' option
12452 to 'openssl version', and is also included in 'openssl version -a'.
12456 * Allowing defining memory allocation callbacks that will be given
12457 file name and line number information in additional arguments
12458 (a const char* and an int). The basic functionality remains, as
12459 well as the original possibility to just replace malloc(),
12460 realloc() and free() by functions that do not know about these
12461 additional arguments. To register and find out the current
12462 settings for extended allocation functions, the following
12463 functions are provided:
12465 CRYPTO_set_mem_ex_functions
12466 CRYPTO_set_locked_mem_ex_functions
12467 CRYPTO_get_mem_ex_functions
12468 CRYPTO_get_locked_mem_ex_functions
12470 These work the same way as CRYPTO_set_mem_functions and friends.
12471 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
12472 extended allocation function is enabled.
12473 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
12474 a conventional allocation function is enabled.
12476 *Richard Levitte, Bodo Moeller*
12478 * Finish off removing the remaining LHASH function pointer casts.
12479 There should no longer be any prototype-casting required when using
12480 the LHASH abstraction, and any casts that remain are "bugs". See
12481 the callback types and macros at the head of lhash.h for details
12482 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12486 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12487 If /dev/[u]random devices are not available or do not return enough
12488 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12490 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12491 /etc/entropy will be queried once each in this sequence, querying stops
12492 when enough entropy was collected without querying more sockets.
12496 * Change the Unix RAND_poll() variant to be able to poll several
12497 random devices, as specified by DEVRANDOM, until a sufficient amount
12498 of data has been collected. We spend at most 10 ms on each file
12499 (select timeout) and read in non-blocking mode. DEVRANDOM now
12500 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12501 (previously it was just the string "/dev/urandom"), so on typical
12502 platforms the 10 ms delay will never occur.
12503 Also separate out the Unix variant to its own file, rand_unix.c.
12504 For VMS, there's a currently-empty rand_vms.c.
12508 * Move OCSP client related routines to ocsp_cl.c. These
12509 provide utility functions which an application needing
12510 to issue a request to an OCSP responder and analyse the
12511 response will typically need: as opposed to those which an
12512 OCSP responder itself would need which will be added later.
12514 OCSP_request_sign() signs an OCSP request with an API similar
12515 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12516 response. OCSP_response_get1_basic() extracts basic response
12517 from response. OCSP_resp_find_status(): finds and extracts status
12518 information from an OCSP_CERTID structure (which will be created
12519 when the request structure is built). These are built from lower
12520 level functions which work on OCSP_SINGLERESP structures but
12521 won't normally be used unless the application wishes to examine
12522 extensions in the OCSP response for example.
12524 Replace nonce routines with a pair of functions.
12525 OCSP_request_add1_nonce() adds a nonce value and optionally
12526 generates a random value. OCSP_check_nonce() checks the
12527 validity of the nonce in an OCSP response.
12531 * Change function OCSP_request_add() to OCSP_request_add0_id().
12532 This doesn't copy the supplied OCSP_CERTID and avoids the
12533 need to free up the newly created id. Change return type
12534 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12535 This can then be used to add extensions to the request.
12536 Deleted OCSP_request_new(), since most of its functionality
12537 is now in OCSP_REQUEST_new() (and the case insensitive name
12538 clash) apart from the ability to set the request name which
12539 will be added elsewhere.
12543 * Update OCSP API. Remove obsolete extensions argument from
12544 various functions. Extensions are now handled using the new
12545 OCSP extension code. New simple OCSP HTTP function which
12546 can be used to send requests and parse the response.
12550 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12551 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12552 uses the special reorder version of SET OF to sort the attributes
12553 and reorder them to match the encoded order. This resolves a long
12554 standing problem: a verify on a PKCS7 structure just after signing
12555 it used to fail because the attribute order did not match the
12556 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12557 it uses the received order. This is necessary to tolerate some broken
12558 software that does not order SET OF. This is handled by encoding
12559 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12560 to produce the required SET OF.
12564 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12565 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12566 files to get correct declarations of the ASN.1 item variables.
12570 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12571 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12572 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12573 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12574 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12575 ASN1_ITEM and no wrapper functions.
12579 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12580 replace the old function pointer based I/O routines. Change most of
12581 the *_d2i_bio() and *_d2i_fp() functions to use these.
12585 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12586 lines, recognize more "algorithms" that can be deselected, and make
12587 it complain about algorithm deselection that isn't recognised.
12591 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12592 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12593 to use new functions. Add NO_ASN1_OLD which can be set to remove
12594 some old style ASN1 functions: this can be used to determine if old
12595 code will still work when these eventually go away.
12599 * New extension functions for OCSP structures, these follow the
12600 same conventions as certificates and CRLs.
12604 * New function X509V3_add1_i2d(). This automatically encodes and
12605 adds an extension. Its behaviour can be customised with various
12606 flags to append, replace or delete. Various wrappers added for
12607 certificates and CRLs.
12611 * Fix to avoid calling the underlying ASN1 print routine when
12612 an extension cannot be parsed. Correct a typo in the
12613 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12617 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12618 entries for variables.
12622 * Add functionality to apps/openssl.c for detecting locking
12623 problems: As the program is single-threaded, all we have
12624 to do is register a locking callback using an array for
12625 storing which locks are currently held by the program.
12629 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12630 SSL_get_ex_data_X509_STORE_idx(), which is used in
12631 ssl_verify_cert_chain() and thus can be called at any time
12632 during TLS/SSL handshakes so that thread-safety is essential.
12633 Unfortunately, the ex_data design is not at all suited
12634 for multi-threaded use, so it probably should be abolished.
12638 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12640 *Broadcom, tweaked and integrated by Geoff Thorpe*
12642 * Move common extension printing code to new function
12643 X509V3_print_extensions(). Reorganise OCSP print routines and
12644 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12648 * New function X509_signature_print() to remove duplication in some
12653 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12654 set (this was treated exactly the same as SET OF previously). This
12655 is used to reorder the STACK representing the structure to match the
12656 encoding. This will be used to get round a problem where a PKCS7
12657 structure which was signed could not be verified because the STACK
12658 order did not reflect the encoded order.
12662 * Reimplement the OCSP ASN1 module using the new code.
12666 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12667 for its ASN1 operations. The old style function pointers still exist
12668 for now but they will eventually go away.
12672 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12673 completely replaces the old ASN1 functionality with a table driven
12674 encoder and decoder which interprets an ASN1_ITEM structure describing
12675 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12676 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12677 has also been converted to the new form.
12681 * Change BN_mod_exp_recp so that negative moduli are tolerated
12682 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12683 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12684 for negative moduli.
12688 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12689 of not touching the result's sign bit.
12693 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12698 * Changed the LHASH code to use prototypes for callbacks, and created
12699 macros to declare and implement thin (optionally static) functions
12700 that provide type-safety and avoid function pointer casting for the
12701 type-specific callbacks.
12705 * Added Kerberos Cipher Suites to be used with TLS, as written in
12707 *Veers Staats <staatsvr@asc.hpc.mil>,
12708 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12710 * Reformat the FAQ so the different questions and answers can be divided
12711 in sections depending on the subject.
12715 * Have the zlib compression code load ZLIB.DLL dynamically under
12720 * New function BN_mod_sqrt for computing square roots modulo a prime
12721 (using the probabilistic Tonelli-Shanks algorithm unless
12722 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12723 be handled deterministically).
12725 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12727 * Make BN_mod_inverse faster by explicitly handling small quotients
12728 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12729 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12733 * New function BN_kronecker.
12737 * Fix BN_gcd so that it works on negative inputs; the result is
12738 positive unless both parameters are zero.
12739 Previously something reasonably close to an infinite loop was
12740 possible because numbers could be growing instead of shrinking
12741 in the implementation of Euclid's algorithm.
12745 * Fix BN_is_word() and BN_is_one() macros to take into account the
12746 sign of the number in question.
12748 Fix BN_is_word(a,w) to work correctly for w == 0.
12750 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12751 because its test if the absolute value of 'a' equals 'w'.
12752 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12753 it exists mostly for use in the implementations of BN_is_zero(),
12754 BN_is_one(), and BN_is_word().
12758 * New function BN_swap.
12762 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12763 the exponentiation functions are more likely to produce reasonable
12764 results on negative inputs.
12768 * Change BN_mod_mul so that the result is always non-negative.
12769 Previously, it could be negative if one of the factors was negative;
12770 I don't think anyone really wanted that behaviour.
12774 * Move BN_mod_... functions into new file crypto/bn/bn_mod.c
12775 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12776 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12777 and add new functions:
12786 BN_mod_lshift1_quick
12788 BN_mod_lshift_quick
12790 These functions always generate non-negative results.
12792 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12793 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12795 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12796 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12797 be reduced modulo m.
12799 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12802 The following entry accidentally appeared in the CHANGES file
12803 distributed with OpenSSL 0.9.7. The modifications described in
12804 it do *not* apply to OpenSSL 0.9.7.
12806 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12807 was actually never needed) and in BN_mul(). The removal in BN_mul()
12808 required a small change in bn_mul_part_recursive() and the addition
12809 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12810 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12811 bn_sub_words() and bn_add_words() except they take arrays with
12817 * In 'openssl passwd', verify passwords read from the terminal
12818 unless the '-salt' option is used (which usually means that
12819 verification would just waste user's time since the resulting
12820 hash is going to be compared with some given password hash)
12821 or the new '-noverify' option is used.
12823 This is an incompatible change, but it does not affect
12824 non-interactive use of 'openssl passwd' (passwords on the command
12825 line, '-stdin' option, '-in ...' option) and thus should not
12826 cause any problems.
12830 * Remove all references to RSAref, since there's no more need for it.
12834 * Make DSO load along a path given through an environment variable
12835 (SHLIB_PATH) with shl_load().
12839 * Constify the ENGINE code as a result of BIGNUM constification.
12840 Also constify the RSA code and most things related to it. In a
12841 few places, most notable in the depth of the ASN.1 code, ugly
12842 casts back to non-const were required (to be solved at a later
12847 * Make it so the openssl application has all engines loaded by default.
12851 * Constify the BIGNUM routines a little more.
12855 * Add the following functions:
12857 ENGINE_load_cswift()
12859 ENGINE_load_atalla()
12860 ENGINE_load_nuron()
12861 ENGINE_load_builtin_engines()
12863 That way, an application can itself choose if external engines that
12864 are built-in in OpenSSL shall ever be used or not. The benefit is
12865 that applications won't have to be linked with libdl or other dso
12866 libraries unless it's really needed.
12868 Changed 'openssl engine' to load all engines on demand.
12869 Changed the engine header files to avoid the duplication of some
12870 declarations (they differed!).
12874 * 'openssl engine' can now list capabilities.
12878 * Better error reporting in 'openssl engine'.
12882 * Never call load_dh_param(NULL) in s_server.
12886 * Add engine application. It can currently list engines by name and
12887 identity, and test if they are actually available.
12891 * Improve RPM specification file by forcing symbolic linking and making
12892 sure the installed documentation is also owned by root.root.
12894 *Damien Miller <djm@mindrot.org>*
12896 * Give the OpenSSL applications more possibilities to make use of
12897 keys (public as well as private) handled by engines.
12901 * Add OCSP code that comes from CertCo.
12905 * Add VMS support for the Rijndael code.
12909 * Added untested support for Nuron crypto accelerator.
12913 * Add support for external cryptographic devices. This code was
12914 previously distributed separately as the "engine" branch.
12916 *Geoff Thorpe, Richard Levitte*
12918 * Rework the filename-translation in the DSO code. It is now possible to
12919 have far greater control over how a "name" is turned into a filename
12920 depending on the operating environment and any oddities about the
12921 different shared library filenames on each system.
12925 * Support threads on FreeBSD-elf in Configure.
12929 * Fix for SHA1 assembly problem with MASM: it produces
12930 warnings about corrupt line number information when assembling
12931 with debugging information. This is caused by the overlapping
12934 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
12937 NCONF_get_number() has no error checking at all. As a replacement,
12938 NCONF_get_number_e() is defined (_e for "error checking") and is
12939 promoted strongly. The old NCONF_get_number is kept around for
12940 binary backward compatibility.
12941 Make it possible for methods to load from something other than a BIO,
12942 by providing a function pointer that is given a name instead of a BIO.
12943 For example, this could be used to load configuration data from an
12948 * Fix for non blocking accept BIOs. Added new I/O special reason
12949 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
12950 with non blocking I/O was not possible because no retry code was
12951 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
12956 * Added the beginnings of Rijndael support.
12960 * Fix for bug in DirectoryString mask setting. Add support for
12961 X509_NAME_print_ex() in 'req' and X509_print_ex() function
12962 to allow certificate printing to more controllable, additional
12963 'certopt' option to 'x509' to allow new printing options to be
12968 * Clean old EAY MD5 hack from e_os.h.
12972 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004] ###
12974 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12975 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
12977 *Joe Orton, Steve Henson*
12979 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003] ###
12981 * Fix additional bug revealed by the NISCC test suite:
12983 Stop bug triggering large recursion when presented with
12984 certain ASN.1 tags [CVE-2003-0851][]
12988 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003] ###
12990 * Fix various bugs revealed by running the NISCC test suite:
12992 Stop out of bounds reads in the ASN1 code when presented with
12993 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12995 If verify callback ignores invalid public key errors don't try to check
12996 certificate signature with the NULL public key.
13001 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13002 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13007 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13008 extra data after the compression methods not only for TLS 1.0
13009 but also for SSL 3.0 (as required by the specification).
13011 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13013 * Change X509_certificate_type() to mark the key as exported/exportable
13014 when it's 512 *bits* long, not 512 bytes.
13018 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003] ###
13020 * Countermeasure against the Klima-Pokorny-Rosa extension of
13021 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13022 a protocol version number mismatch like a decryption error
13023 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13027 * Turn on RSA blinding by default in the default implementation
13028 to avoid a timing attack. Applications that don't want it can call
13029 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13030 They would be ill-advised to do so in most cases.
13032 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13034 * Change RSA blinding code so that it works when the PRNG is not
13035 seeded (in this case, the secret RSA exponent is abused as
13036 an unpredictable seed -- if it is not unpredictable, there
13037 is no point in blinding anyway). Make RSA blinding thread-safe
13038 by remembering the creator's thread ID in rsa->blinding and
13039 having all other threads use local one-time blinding factors
13040 (this requires more computation than sharing rsa->blinding, but
13041 avoids excessive locking; and if an RSA object is not shared
13042 between threads, blinding will still be very fast).
13046 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003] ###
13048 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13049 via timing by performing a MAC computation even if incorrect
13050 block cipher padding has been found. This is a countermeasure
13051 against active attacks where the attacker has to distinguish
13052 between bad padding and a MAC verification error. [CVE-2003-0078][]
13054 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13055 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13056 Martin Vuagnoux (EPFL, Ilion)*
13058 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002] ###
13060 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13061 memory from its contents. This is done with a counter that will
13062 place alternating values in each byte. This can be used to solve
13063 two issues: 1) the removal of calls to memset() by highly optimizing
13064 compilers, and 2) cleansing with other values than 0, since those can
13065 be read through on certain media, for example a swap space on disk.
13069 * Bugfix: client side session caching did not work with external caching,
13070 because the session->cipher setting was not restored when reloading
13071 from the external cache. This problem was masked, when
13072 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13073 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13077 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13078 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13080 *Zeev Lieber <zeev-l@yahoo.com>*
13082 * Undo an undocumented change introduced in 0.9.6e which caused
13083 repeated calls to OpenSSL_add_all_ciphers() and
13084 OpenSSL_add_all_digests() to be ignored, even after calling
13089 * Change the default configuration reader to deal with last line not
13090 being properly terminated.
13094 * Change X509_NAME_cmp() so it applies the special rules on handling
13095 DN values that are of type PrintableString, as well as RDNs of type
13096 emailAddress where the value has the type ia5String.
13098 *stefank@valicert.com via Richard Levitte*
13100 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13101 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13102 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13103 the bitwise-OR of the two for use by the majority of applications
13104 wanting this behaviour, and update the docs. The documented
13105 behaviour and actual behaviour were inconsistent and had been
13106 changing anyway, so this is more a bug-fix than a behavioural
13109 *Geoff Thorpe, diagnosed by Nadav Har'El*
13111 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13112 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13116 * Fix initialization code race conditions in
13117 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13118 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13119 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13120 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13121 ssl2_get_cipher_by_char(),
13122 ssl3_get_cipher_by_char().
13124 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13126 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13127 the cached sessions are flushed, as the remove_cb() might use ex_data
13128 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13129 (see [openssl.org #212]).
13131 *Geoff Thorpe, Lutz Jaenicke*
13133 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13134 length, instead of the encoding length to d2i_ASN1_OBJECT.
13138 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002] ###
13140 * [In 0.9.6g-engine release:]
13141 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
13143 *Lynn Gazis <lgazis@rainbow.com>*
13145 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002] ###
13147 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13148 and get fix the header length calculation.
13149 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13150 Alon Kantor <alonk@checkpoint.com> (and others),
13153 * Use proper error handling instead of 'assertions' in buffer
13154 overflow checks added in 0.9.6e. This prevents DoS (the
13155 assertions could call abort()).
13157 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13159 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002] ###
13161 * Add various sanity checks to asn1_get_length() to reject
13162 the ASN1 length bytes if they exceed sizeof(long), will appear
13163 negative or the content length exceeds the length of the
13166 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13168 * Fix cipher selection routines: ciphers without encryption had no flags
13169 for the cipher strength set and where therefore not handled correctly
13170 by the selection routines (PR #130).
13174 * Fix EVP_dsa_sha macro.
13179 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13180 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13181 that was added in OpenSSL 0.9.6d.
13183 As the countermeasure turned out to be incompatible with some
13184 broken SSL implementations, the new option is part of SSL_OP_ALL.
13185 SSL_OP_ALL is usually employed when compatibility with weird SSL
13186 implementations is desired (e.g. '-bugs' option to 's_client' and
13187 's_server'), so the new option is automatically set in many
13192 * Changes in security patch:
13194 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13195 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13196 Air Force Materiel Command, USAF, under agreement number
13199 * Add various sanity checks to asn1_get_length() to reject
13200 the ASN1 length bytes if they exceed sizeof(long), will appear
13201 negative or the content length exceeds the length of the
13202 supplied buffer. [CVE-2002-0659][]
13204 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13206 * Assertions for various potential buffer overflows, not known to
13207 happen in practice.
13209 *Ben Laurie (CHATS)*
13211 * Various temporary buffers to hold ASCII versions of integers were
13212 too small for 64 bit platforms. [CVE-2002-0655][]
13213 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
13215 * Remote buffer overflow in SSL3 protocol - an attacker could
13216 supply an oversized session ID to a client. [CVE-2002-0656][]
13218 *Ben Laurie (CHATS)*
13220 * Remote buffer overflow in SSL2 protocol - an attacker could
13221 supply an oversized client master key. [CVE-2002-0656][]
13223 *Ben Laurie (CHATS)*
13225 ### Changes between 0.9.6c and 0.9.6d [9 May 2002] ###
13227 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13228 encoded as NULL) with id-dsa-with-sha1.
13230 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13232 * Check various X509_...() return values in apps/req.c.
13234 *Nils Larsch <nla@trustcenter.de>*
13236 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13237 an end-of-file condition would erroneously be flagged, when the CRLF
13238 was just at the end of a processed block. The bug was discovered when
13239 processing data through a buffering memory BIO handing the data to a
13240 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13241 <ptsekov@syntrex.com> and Nedelcho Stanev.
13245 * Implement a countermeasure against a vulnerability recently found
13246 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13247 before application data chunks to avoid the use of known IVs
13248 with data potentially chosen by the attacker.
13252 * Fix length checks in ssl3_get_client_hello().
13256 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13257 to prevent ssl3_read_internal() from incorrectly assuming that
13258 ssl3_read_bytes() found application data while handshake
13259 processing was enabled when in fact s->s3->in_read_app_data was
13260 merely automatically cleared during the initial handshake.
13262 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13264 * Fix object definitions for Private and Enterprise: they were not
13265 recognized in their shortname (=lowercase) representation. Extend
13266 obj_dat.pl to issue an error when using undefined keywords instead
13267 of silently ignoring the problem (Svenning Sorensen
13268 <sss@sss.dnsalias.net>).
13272 * Fix DH_generate_parameters() so that it works for 'non-standard'
13273 generators, i.e. generators other than 2 and 5. (Previously, the
13274 code did not properly initialise the 'add' and 'rem' values to
13275 BN_generate_prime().)
13277 In the new general case, we do not insist that 'generator' is
13278 actually a primitive root: This requirement is rather pointless;
13279 a generator of the order-q subgroup is just as good, if not
13284 * Map new X509 verification errors to alerts. Discovered and submitted by
13285 Tom Wu <tom@arcot.com>.
13289 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13290 returning non-zero before the data has been completely received
13291 when using non-blocking I/O.
13293 *Bodo Moeller; problem pointed out by John Hughes*
13295 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13297 *Ben Laurie, Lutz Jaenicke*
13299 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13300 Yoram Zahavi <YoramZ@gilian.com>).
13304 * Add information about CygWin 1.3 and on, and preserve proper
13305 configuration for the versions before that.
13307 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13309 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13310 check whether we deal with a copy of a session and do not delete from
13311 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13312 <izhar@checkpoint.com>.
13316 * Do not store session data into the internal session cache, if it
13317 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13318 flag is set). Proposed by Aslam <aslam@funk.com>.
13322 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13327 * [In 0.9.6d-engine release:]
13328 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13330 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13332 * Add the configuration target linux-s390x.
13334 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13336 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13337 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13338 variable as an indication that a ClientHello message has been
13339 received. As the flag value will be lost between multiple
13340 invocations of ssl3_accept when using non-blocking I/O, the
13341 function may not be aware that a handshake has actually taken
13342 place, thus preventing a new session from being added to the
13345 To avoid this problem, we now set s->new_session to 2 instead of
13346 using a local variable.
13348 *Lutz Jaenicke, Bodo Moeller*
13350 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13351 if the SSL_R_LENGTH_MISMATCH error is detected.
13353 *Geoff Thorpe, Bodo Moeller*
13355 * New 'shared_ldflag' column in Configure platform table.
13359 * Fix EVP_CIPHER_mode macro.
13361 *"Dan S. Camper" <dan@bti.net>*
13363 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13364 type, we must throw them away by setting rr->length to 0.
13366 *D P Chang <dpc@qualys.com>*
13368 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001] ###
13370 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13371 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13372 worked incorrectly for those cases where range = 10..._2 and
13373 3*range is two bits longer than range.)
13377 * Only add signing time to PKCS7 structures if it is not already
13382 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13383 OBJ_ld_ce should be OBJ_id_ce.
13384 Also some ip-pda OIDs in crypto/objects/objects.txt were
13385 incorrect (cf. RFC 3039).
13387 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13389 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13390 returns early because it has nothing to do.
13392 *Andy Schneider <andy.schneider@bjss.co.uk>*
13394 * [In 0.9.6c-engine release:]
13395 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13397 *Andy Schneider <andy.schneider@bjss.co.uk>*
13399 * [In 0.9.6c-engine release:]
13400 Add support for Cryptographic Appliance's keyserver technology.
13401 (Use engine 'keyclient')
13403 *Cryptographic Appliances and Geoff Thorpe*
13405 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13406 is called via tools/c89.sh because arguments have to be
13407 rearranged (all '-L' options must appear before the first object
13410 *Richard Shapiro <rshapiro@abinitio.com>*
13412 * [In 0.9.6c-engine release:]
13413 Add support for Broadcom crypto accelerator cards, backported
13416 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13418 * [In 0.9.6c-engine release:]
13419 Add support for SureWare crypto accelerator cards from
13420 Baltimore Technologies. (Use engine 'sureware')
13422 *Baltimore Technologies and Mark Cox*
13424 * [In 0.9.6c-engine release:]
13425 Add support for crypto accelerator cards from Accelerated
13426 Encryption Processing, www.aep.ie. (Use engine 'aep')
13428 *AEP Inc. and Mark Cox*
13430 * Add a configuration entry for gcc on UnixWare.
13432 *Gary Benson <gbenson@redhat.com>*
13434 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13435 messages are stored in a single piece (fixed-length part and
13436 variable-length part combined) and fix various bugs found on the way.
13440 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13441 instead. BIO_gethostbyname() does not know what timeouts are
13442 appropriate, so entries would stay in cache even when they have
13444 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
13446 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13447 faced with a pathologically small ClientHello fragment that does
13448 not contain client_version: Instead of aborting with an error,
13449 simply choose the highest available protocol version (i.e.,
13450 TLS 1.0 unless it is disabled). In practice, ClientHello
13451 messages are never sent like this, but this change gives us
13452 strictly correct behaviour at least for TLS.
13456 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13457 never resets s->method to s->ctx->method when called from within
13458 one of the SSL handshake functions.
13460 *Bodo Moeller; problem pointed out by Niko Baric*
13462 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13463 (sent using the client's version number) if client_version is
13464 smaller than the protocol version in use. Also change
13465 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13466 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13467 the client will at least see that alert.
13471 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13476 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13477 client receives HelloRequest while in a handshake.
13479 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13481 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13482 should end in 'break', not 'goto end' which circumvents various
13483 cleanups done in state SSL_ST_OK. But session related stuff
13484 must be disabled for SSL_ST_OK in the case that we just sent a
13487 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13488 before just sending a HelloRequest.
13490 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13492 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13493 reveal whether illegal block cipher padding was found or a MAC
13494 verification error occurred. (Neither SSLerr() codes nor alerts
13495 are directly visible to potential attackers, but the information
13496 may leak via logfiles.)
13498 Similar changes are not required for the SSL 2.0 implementation
13499 because the number of padding bytes is sent in clear for SSL 2.0,
13500 and the extra bytes are just ignored. However ssl/s2_pkt.c
13501 failed to verify that the purported number of padding bytes is in
13506 * Add OpenUNIX-8 support including shared libraries
13507 (Boyd Lynn Gerber <gerberb@zenez.com>).
13511 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13512 'wristwatch attack' using huge encoding parameters (cf.
13513 James H. Manger's CRYPTO 2001 paper). Note that the
13514 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13515 encoding parameters and hence was not vulnerable.
13519 * BN_sqr() bug fix.
13521 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13523 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13524 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13525 followed by modular reduction.
13527 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13529 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13530 equivalent based on BN_pseudo_rand() instead of BN_rand().
13534 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13535 This function was broken, as the check for a new client hello message
13536 to handle SGC did not allow these large messages.
13537 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13541 * Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
13545 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13546 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13550 * Rework the configuration and shared library support for Tru64 Unix.
13551 The configuration part makes use of modern compiler features and
13552 still retains old compiler behavior for those that run older versions
13553 of the OS. The shared library support part includes a variant that
13554 uses the RPATH feature, and is available through the special
13555 configuration target "alpha-cc-rpath", which will never be selected
13558 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13560 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13561 with the same message size as in ssl3_get_certificate_request().
13562 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13563 messages might inadvertently be reject as too long.
13565 *Petr Lampa <lampa@fee.vutbr.cz>*
13567 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13571 * Modified SSL library such that the verify_callback that has been set
13572 specifically for an SSL object with SSL_set_verify() is actually being
13573 used. Before the change, a verify_callback set with this function was
13574 ignored and the verify_callback() set in the SSL_CTX at the time of
13575 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13576 to allow the necessary settings.
13580 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13581 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13582 done automatically (in contradiction to the requirements of the C
13583 standard). This made problems when used from OpenSSH.
13587 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13588 dh->length and always used
13590 BN_rand_range(priv_key, dh->p).
13592 BN_rand_range() is not necessary for Diffie-Hellman, and this
13593 specific range makes Diffie-Hellman unnecessarily inefficient if
13594 dh->length (recommended exponent length) is much smaller than the
13595 length of dh->p. We could use BN_rand_range() if the order of
13596 the subgroup was stored in the DH structure, but we only have
13601 BN_rand(priv_key, l, ...)
13603 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13610 RSA_eay_public_encrypt
13611 RSA_eay_private_decrypt
13612 RSA_eay_private_encrypt (signing)
13613 RSA_eay_public_decrypt (signature verification)
13615 (default implementations for RSA_public_encrypt,
13616 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13617 always reject numbers >= n.
13621 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13622 to synchronize access to 'locking_thread'. This is necessary on
13623 systems where access to 'locking_thread' (an 'unsigned long'
13624 variable) is not atomic.
13628 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13629 *before* setting the 'crypto_lock_rand' flag. The previous code had
13630 a race condition if 0 is a valid thread ID.
13632 *Travis Vitek <vitek@roguewave.com>*
13634 * Add support for shared libraries under Irix.
13636 *Albert Chin-A-Young <china@thewrittenword.com>*
13638 * Add configuration option to build on Linux on both big-endian and
13639 little-endian MIPS.
13641 *Ralf Baechle <ralf@uni-koblenz.de>*
13643 * Add the possibility to create shared libraries on HP-UX.
13647 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001] ###
13649 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13650 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13651 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13652 PRNG state recovery was possible based on the output of
13653 one PRNG request appropriately sized to gain knowledge on
13654 'md' followed by enough consecutive 1-byte PRNG requests
13655 to traverse all of 'state'.
13657 1. When updating 'md_local' (the current thread's copy of 'md')
13658 during PRNG output generation, hash all of the previous
13659 'md_local' value, not just the half used for PRNG output.
13661 2. Make the number of bytes from 'state' included into the hash
13662 independent from the number of PRNG bytes requested.
13664 The first measure alone would be sufficient to avoid
13665 Markku-Juhani's attack. (Actually it had never occurred
13666 to me that the half of 'md_local' used for chaining was the
13667 half from which PRNG output bytes were taken -- I had always
13668 assumed that the secret half would be used.) The second
13669 measure makes sure that additional data from 'state' is never
13670 mixed into 'md_local' in small portions; this heuristically
13671 further strengthens the PRNG.
13675 * Fix crypto/bn/asm/mips3.s.
13679 * When only the key is given to "enc", the IV is undefined. Print out
13680 an error message in this case.
13684 * Handle special case when X509_NAME is empty in X509 printing routines.
13688 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13689 positive and less than q.
13693 * Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
13694 used: it isn't thread safe and the add_lock_callback should handle
13697 *Paul Rose <Paul.Rose@bridge.com>*
13699 * Verify that incoming data obeys the block size in
13700 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13706 *Ulf Möller, Bodo Möller*
13708 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13709 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13710 when fixing the server behaviour for backwards-compatible 'client
13711 hello' messages. (Note that the attack is impractical against
13712 SSL 3.0 and TLS 1.0 anyway because length and version checking
13713 means that the probability of guessing a valid ciphertext is
13714 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13717 Before 0.9.5, the countermeasure (hide the error by generating a
13718 random 'decryption result') did not work properly because
13719 ERR_clear_error() was missing, meaning that SSL_get_error() would
13720 detect the supposedly ignored error.
13722 Both problems are now fixed.
13726 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13727 (previously it was 1024).
13731 * Fix for compatibility mode trust settings: ignore trust settings
13732 unless some valid trust or reject settings are present.
13736 * Fix for blowfish EVP: its a variable length cipher.
13740 * Fix various bugs related to DSA S/MIME verification. Handle missing
13741 parameters in DSA public key structures and return an error in the
13742 DSA routines if parameters are absent.
13746 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13747 in the current directory if neither $RANDFILE nor $HOME was set.
13748 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13749 caused some confusion to Windows users who haven't defined $HOME.
13750 Thus RAND_file_name() is changed again: e_os.h can define a
13751 DEFAULT_HOME, which will be used if $HOME is not set.
13752 For Windows, we use "C:"; on other platforms, we still require
13753 environment variables.
13755 * Move 'if (!initialized) RAND_poll()' into regions protected by
13756 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13757 having multiple threads call RAND_poll() concurrently.
13761 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13762 combination of a flag and a thread ID variable.
13763 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13764 flag), *other* threads can enter ssleay_add_bytes without obeying
13765 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13766 that they do not hold after the first thread unsets add_do_not_lock).
13770 * Change bctest again: '-x' expressions are not available in all
13771 versions of 'test'.
13775 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001] ###
13777 * Fix a couple of memory leaks in PKCS7_dataDecode()
13779 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13781 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13782 the default extension for executables, if any. Also, make the perl
13783 scripts that use symlink() to test if it really exists and use "cp"
13784 if it doesn't. All this made OpenSSL compilable and installable in
13789 * Fix for asn1_GetSequence() for indefinite length constructed data.
13790 If SEQUENCE is length is indefinite just set c->slen to the total
13791 amount of data available.
13793 *Steve Henson, reported by shige@FreeBSD.org*
13795 *This change does not apply to 0.9.7.*
13797 * Change bctest to avoid here-documents inside command substitution
13798 (workaround for FreeBSD /bin/sh bug).
13799 For compatibility with Ultrix, avoid shell functions (introduced
13800 in the bctest version that searches along $PATH).
13804 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13805 with des_encrypt() defined on some operating systems, like Solaris
13810 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13811 On the Importance of Eliminating Errors in Cryptographic
13812 Computations, J. Cryptology 14 (2001) 2, 101-119,
13813 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
13817 * MIPS assembler BIGNUM division bug fix.
13821 * Disabled incorrect Alpha assembler code.
13825 * Fix PKCS#7 decode routines so they correctly update the length
13826 after reading an EOC for the EXPLICIT tag.
13830 *This change does not apply to 0.9.7.*
13832 * Fix bug in PKCS#12 key generation routines. This was triggered
13833 if a 3DES key was generated with a 0 initial byte. Include
13834 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13835 (but broken) behaviour.
13839 * Enhance bctest to search for a working bc along $PATH and print
13842 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13844 * Fix memory leaks in err.c: free err_data string if necessary;
13845 don't write to the wrong index in ERR_set_error_data.
13849 * Implement ssl23_peek (analogous to ssl23_read), which previously
13854 * Replace rdtsc with _emit statements for VC++ version 5.
13856 *Jeremy Cooper <jeremy@baymoo.org>*
13858 * Make it possible to reuse SSLv2 sessions.
13862 * In copy_email() check for >= 0 as a return value for
13863 X509_NAME_get_index_by_NID() since 0 is a valid index.
13865 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13867 * Avoid coredump with unsupported or invalid public keys by checking if
13868 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13869 PKCS7_verify() fails with non detached data.
13873 * Don't use getenv in library functions when run as setuid/setgid.
13874 New function OPENSSL_issetugid().
13878 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13879 due to incorrect handling of multi-threading:
13881 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13883 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13885 3. Count how many times MemCheck_off() has been called so that
13886 nested use can be treated correctly. This also avoids
13887 inband-signalling in the previous code (which relied on the
13888 assumption that thread ID 0 is impossible).
13892 * Add "-rand" option also to s_client and s_server.
13896 * Fix CPU detection on Irix 6.x.
13897 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13898 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13900 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13905 *This change does not apply to 0.9.7.*
13907 * Use the cached encoding of an X509_NAME structure rather than
13908 copying it. This is apparently the reason for the libsafe "errors"
13909 but the code is actually correct.
13913 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13914 Bleichenbacher's DSA attack.
13915 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13916 to be set and top=0 forces the highest bit to be set; top=-1 is new
13917 and leaves the highest bit random.
13919 *Ulf Moeller, Bodo Moeller*
13921 * In the NCONF_...-based implementations for CONF_... queries
13922 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13923 a temporary CONF structure with the data component set to NULL
13924 (which gives segmentation faults in lh_retrieve).
13925 Instead, use NULL for the CONF pointer in CONF_get_string and
13926 CONF_get_number (which may use environment variables) and directly
13927 return NULL from CONF_get_section.
13931 * Fix potential buffer overrun for EBCDIC.
13935 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
13936 keyUsage if basicConstraints absent for a CA.
13940 * Make SMIME_write_PKCS7() write mail header values with a format that
13941 is more generally accepted (no spaces before the semicolon), since
13942 some programs can't parse those values properly otherwise. Also make
13943 sure BIO's that break lines after each write do not create invalid
13948 * Make the CRL encoding routines work with empty SEQUENCE OF. The
13949 macros previously used would not encode an empty SEQUENCE OF
13950 and break the signature.
13954 *This change does not apply to 0.9.7.*
13956 * Zero the premaster secret after deriving the master secret in
13961 * Add some EVP_add_digest_alias registrations (as found in
13962 OpenSSL_add_all_digests()) to SSL_library_init()
13963 aka OpenSSL_add_ssl_algorithms(). This provides improved
13964 compatibility with peers using X.509 certificates
13965 with unconventional AlgorithmIdentifier OIDs.
13969 * Fix for Irix with NO_ASM.
13971 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13973 * ./config script fixes.
13975 *Ulf Moeller, Richard Levitte*
13977 * Fix 'openssl passwd -1'.
13981 * Change PKCS12_key_gen_asc() so it can cope with non null
13982 terminated strings whose length is passed in the passlen
13983 parameter, for example from PEM callbacks. This was done
13984 by adding an extra length parameter to asc2uni().
13986 *Steve Henson, reported by <oddissey@samsung.co.kr>*
13988 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
13989 call failed, free the DSA structure.
13993 * Fix to uni2asc() to cope with zero length Unicode strings.
13994 These are present in some PKCS#12 files.
13998 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
13999 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14000 when writing a 32767 byte record.
14002 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14004 * In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
14005 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
14007 (RSA objects have a reference count access to which is protected
14008 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14009 so they are meant to be shared between threads.)
14010 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14011 "Reddie, Steven" <Steven.Reddie@ca.com>*
14013 * Fix a deadlock in CRYPTO_mem_leaks().
14017 * Use better test patterns in bntest.
14021 * rand_win.c fix for Borland C.
14025 * BN_rshift bugfix for n == 0.
14029 * Add a 'bctest' script that checks for some known 'bc' bugs
14030 so that 'make test' does not abort just because 'bc' is broken.
14034 * Store verify_result within SSL_SESSION also for client side to
14035 avoid potential security hole. (Re-used sessions on the client side
14036 always resulted in verify_result==X509_V_OK, not using the original
14037 result of the server certificate verification.)
14041 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14042 SSL3_RT_APPLICATION_DATA, return 0.
14043 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14048 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14049 releases, have been re-implemented by renaming the previous
14050 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14051 and ssl3_read_internal, respectively, and adding 'peek' parameters
14052 to them. The new ssl[23]_{read,peek} functions are calls to
14053 ssl[23]_read_internal with the 'peek' flag set appropriately.
14054 A 'peek' parameter has also been added to ssl3_read_bytes, which
14055 does the actual work for ssl3_read_internal.
14059 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14060 the method-specific "init()" handler. Also clean up ex_data after
14061 calling the method-specific "finish()" handler. Previously, this was
14062 happening the other way round.
14066 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14067 The previous value, 12, was not always sufficient for BN_mod_exp().
14071 * Make sure that shared libraries get the internal name engine with
14072 the full version number and not just 0. This should mark the
14073 shared libraries as not backward compatible. Of course, this should
14074 be changed again when we can guarantee backward binary compatibility.
14078 * Fix typo in get_cert_by_subject() in by_dir.c
14080 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14082 * Rework the system to generate shared libraries:
14084 - Make note of the expected extension for the shared libraries and
14085 if there is a need for symbolic links from for example libcrypto.so.0
14086 to libcrypto.so.0.9.7. There is extended info in Configure for
14089 - Make as few rebuilds of the shared libraries as possible.
14091 - Still avoid linking the OpenSSL programs with the shared libraries.
14093 - When installing, install the shared libraries separately from the
14098 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14100 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14101 and not in SSL_clear because the latter is also used by the
14102 accept/connect functions; previously, the settings made by
14103 SSL_set_read_ahead would be lost during the handshake.
14105 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14107 * Correct util/mkdef.pl to be selective about disabled algorithms.
14108 Previously, it would create entries for disabled algorithms no
14113 * Added several new manual pages for SSL_* function.
14117 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000] ###
14119 * In ssl23_get_client_hello, generate an error message when faced
14120 with an initial SSL 3.0/TLS record that is too small to contain the
14121 first two bytes of the ClientHello message, i.e. client_version.
14122 (Note that this is a pathologic case that probably has never happened
14123 in real life.) The previous approach was to use the version number
14124 from the record header as a substitute; but our protocol choice
14125 should not depend on that one because it is not authenticated
14126 by the Finished messages.
14130 * More robust randomness gathering functions for Windows.
14132 *Jeffrey Altman <jaltman@columbia.edu>*
14134 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14135 not set then we don't setup the error code for issuer check errors
14136 to avoid possibly overwriting other errors which the callback does
14137 handle. If an application does set the flag then we assume it knows
14138 what it is doing and can handle the new informational codes
14143 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14144 a general "ANY" type, as such it should be able to decode anything
14145 including tagged types. However it didn't check the class so it would
14146 wrongly interpret tagged types in the same way as their universal
14147 counterpart and unknown types were just rejected. Changed so that the
14148 tagged and unknown types are handled in the same way as a SEQUENCE:
14149 that is the encoding is stored intact. There is also a new type
14150 "V_ASN1_OTHER" which is used when the class is not universal, in this
14151 case we have no idea what the actual type is so we just lump them all
14156 * On VMS, stdout may very well lead to a file that is written to
14157 in a record-oriented fashion. That means that every write() will
14158 write a separate record, which will be read separately by the
14159 programs trying to read from it. This can be very confusing.
14161 The solution is to put a BIO filter in the way that will buffer
14162 text until a linefeed is reached, and then write everything a
14163 line at a time, so every record written will be an actual line,
14164 not chunks of lines and not (usually doesn't happen, but I've
14165 seen it once) several lines in one record. BIO_f_linebuffer() is
14168 Currently, it's a VMS-only method, because that's where it has
14169 been tested well enough.
14173 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14174 it can return incorrect results.
14175 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14176 but it was in 0.9.6-beta[12].)
14180 * Disable the check for content being present when verifying detached
14181 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14182 include zero length content when signing messages.
14186 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14187 BIO_ctrl (for BIO pairs).
14191 * Add DSO method for VMS.
14195 * Bug fix: Montgomery multiplication could produce results with the
14200 * Add RPM specification openssl.spec and modify it to build three
14201 packages. The default package contains applications, application
14202 documentation and run-time libraries. The devel package contains
14203 include files, static libraries and function documentation. The
14204 doc package contains the contents of the doc directory. The original
14205 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14209 * Add a large number of documentation files for many SSL routines.
14211 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14213 * Add a configuration entry for Sony News 4.
14215 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14217 * Don't set the two most significant bits to one when generating a
14218 random number < q in the DSA library.
14222 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14223 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14224 the underlying transport is blocking) if a handshake took place.
14225 (The default behaviour is needed by applications such as s_client
14226 and s_server that use select() to determine when to use SSL_read;
14227 but for applications that know in advance when to expect data, it
14228 just makes things more complicated.)
14232 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14237 * Add a few more EBCDIC conditionals that make `req' and `x509'
14238 work better on such systems.
14240 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14242 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14243 Update PKCS12_parse() so it copies the friendlyName and the
14244 keyid to the certificates aux info.
14248 * Fix bug in PKCS7_verify() which caused an infinite loop
14249 if there was more than one signature.
14251 *Sven Uszpelkat <su@celocom.de>*
14253 * Major change in util/mkdef.pl to include extra information
14254 about each symbol, as well as presenting variables as well
14255 as functions. This change means that there's n more need
14256 to rebuild the .num files when some algorithms are excluded.
14260 * Allow the verify time to be set by an application,
14261 rather than always using the current time.
14265 * Phase 2 verify code reorganisation. The certificate
14266 verify code now looks up an issuer certificate by a
14267 number of criteria: subject name, authority key id
14268 and key usage. It also verifies self signed certificates
14269 by the same criteria. The main comparison function is
14270 X509_check_issued() which performs these checks.
14272 Lot of changes were necessary in order to support this
14273 without completely rewriting the lookup code.
14275 Authority and subject key identifier are now cached.
14277 The LHASH 'certs' is X509_STORE has now been replaced
14278 by a STACK_OF(X509_OBJECT). This is mainly because an
14279 LHASH can't store or retrieve multiple objects with
14280 the same hash value.
14282 As a result various functions (which were all internal
14283 use only) have changed to handle the new X509_STORE
14284 structure. This will break anything that messed round
14285 with X509_STORE internally.
14287 The functions X509_STORE_add_cert() now checks for an
14288 exact match, rather than just subject name.
14290 The X509_STORE API doesn't directly support the retrieval
14291 of multiple certificates matching a given criteria, however
14292 this can be worked round by performing a lookup first
14293 (which will fill the cache with candidate certificates)
14294 and then examining the cache for matches. This is probably
14295 the best we can do without throwing out X509_LOOKUP
14296 entirely (maybe later...).
14298 The X509_VERIFY_CTX structure has been enhanced considerably.
14300 All certificate lookup operations now go via a get_issuer()
14301 callback. Although this currently uses an X509_STORE it
14302 can be replaced by custom lookups. This is a simple way
14303 to bypass the X509_STORE hackery necessary to make this
14304 work and makes it possible to use more efficient techniques
14305 in future. A very simple version which uses a simple
14306 STACK for its trusted certificate store is also provided
14307 using X509_STORE_CTX_trusted_stack().
14309 The verify_cb() and verify() callbacks now have equivalents
14310 in the X509_STORE_CTX structure.
14312 X509_STORE_CTX also has a 'flags' field which can be used
14313 to customise the verify behaviour.
14317 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14318 excludes S/MIME capabilities.
14322 * When a certificate request is read in keep a copy of the
14323 original encoding of the signed data and use it when outputting
14324 again. Signatures then use the original encoding rather than
14325 a decoded, encoded version which may cause problems if the
14326 request is improperly encoded.
14330 * For consistency with other BIO_puts implementations, call
14331 buffer_write(b, ...) directly in buffer_puts instead of calling
14334 In BIO_puts, increment b->num_write as in BIO_write.
14336 *Peter.Sylvester@EdelWeb.fr*
14338 * Fix BN_mul_word for the case where the word is 0. (We have to use
14339 BN_zero, we may not return a BIGNUM with an array consisting of
14340 words set to zero.)
14344 * Avoid calling abort() from within the library when problems are
14345 detected, except if preprocessor symbols have been defined
14346 (such as REF_CHECK, BN_DEBUG etc.).
14350 * New openssl application 'rsautl'. This utility can be
14351 used for low level RSA operations. DER public key
14352 BIO/fp routines also added.
14356 * New Configure entry and patches for compiling on QNX 4.
14358 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14360 * A demo state-machine implementation was sponsored by
14361 Nuron (http://www.nuron.com/) and is now available in
14362 demos/state_machine.
14366 * New options added to the 'dgst' utility for signature
14367 generation and verification.
14371 * Unrecognized PKCS#7 content types are now handled via a
14372 catch all ASN1_TYPE structure. This allows unsupported
14373 types to be stored as a "blob" and an application can
14374 encode and decode it manually.
14378 * Fix various signed/unsigned issues to make a_strex.c
14379 compile under VC++.
14381 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14383 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14384 length if passed a buffer. ASN1_INTEGER_to_BN failed
14385 if passed a NULL BN and its argument was negative.
14387 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14389 * Modification to PKCS#7 encoding routines to output definite
14390 length encoding. Since currently the whole structures are in
14391 memory there's not real point in using indefinite length
14392 constructed encoding. However if OpenSSL is compiled with
14393 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14397 * Added BIO_vprintf() and BIO_vsnprintf().
14401 * Added more prefixes to parse for in the strings written
14402 through a logging bio, to cover all the levels that are available
14403 through syslog. The prefixes are now:
14405 PANIC, EMERG, EMR => LOG_EMERG
14406 ALERT, ALR => LOG_ALERT
14407 CRIT, CRI => LOG_CRIT
14408 ERROR, ERR => LOG_ERR
14409 WARNING, WARN, WAR => LOG_WARNING
14410 NOTICE, NOTE, NOT => LOG_NOTICE
14411 INFO, INF => LOG_INFO
14412 DEBUG, DBG => LOG_DEBUG
14414 and as before, if none of those prefixes are present at the
14415 beginning of the string, LOG_ERR is chosen.
14417 On Win32, the LOG_* levels are mapped according to this:
14419 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14420 LOG_WARNING => EVENTLOG_WARNING_TYPE
14421 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14426 * Made it possible to reconfigure with just the configuration
14427 argument "reconf" or "reconfigure". The command line arguments
14428 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14429 and are retrieved from there when reconfiguring.
14435 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14437 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14441 * The obj_dat.pl script was messing up the sorting of object
14442 names. The reason was that it compared the quoted version
14443 of strings as a result "OCSP" > "OCSP Signing" because
14444 " > SPACE. Changed script to store unquoted versions of
14445 names and add quotes on output. It was also omitting some
14446 names from the lookup table if they were given a default
14447 value (that is if SN is missing it is given the same
14448 value as LN and vice versa), these are now added on the
14449 grounds that if an object has a name we should be able to
14450 look it up. Finally added warning output when duplicate
14451 short or long names are found.
14455 * Changes needed for Tandem NSK.
14457 *Scott Uroff <scott@xypro.com>*
14459 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14460 RSA_padding_check_SSLv23(), special padding was never detected
14461 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14462 version rollback attacks was not effective.
14464 In s23_clnt.c, don't use special rollback-attack detection padding
14465 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14466 client; similarly, in s23_srvr.c, don't do the rollback check if
14467 SSL 2.0 is the only protocol enabled in the server.
14471 * Make it possible to get hexdumps of unprintable data with 'openssl
14472 asn1parse'. By implication, the functions ASN1_parse_dump() and
14473 BIO_dump_indent() are added.
14477 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14478 these print out strings and name structures based on various
14479 flags including RFC2253 support and proper handling of
14480 multibyte characters. Added options to the 'x509' utility
14481 to allow the various flags to be set.
14485 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14486 Also change the functions X509_cmp_current_time() and
14487 X509_gmtime_adj() work with an ASN1_TIME structure,
14488 this will enable certificates using GeneralizedTime in validity
14489 dates to be checked.
14493 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14494 negative public key encodings) on by default,
14495 NO_NEG_PUBKEY_BUG can be set to disable it.
14499 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14500 content octets. An i2c_ASN1_OBJECT is unnecessary because
14501 the encoding can be trivially obtained from the structure.
14505 * crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
14506 not read locks (CRYPTO_r_[un]lock).
14510 * A first attempt at creating official support for shared
14511 libraries through configuration. I've kept it so the
14512 default is static libraries only, and the OpenSSL programs
14513 are always statically linked for now, but there are
14514 preparations for dynamic linking in place.
14515 This has been tested on Linux and Tru64.
14519 * Randomness polling function for Win9x, as described in:
14520 Peter Gutmann, Software Generation of Practically Strong
14525 * Fix so PRNG is seeded in req if using an already existing
14530 * New options to smime application. -inform and -outform
14531 allow alternative formats for the S/MIME message including
14532 PEM and DER. The -content option allows the content to be
14533 specified separately. This should allow things like Netscape
14534 form signing output easier to verify.
14538 * Fix the ASN1 encoding of tags using the 'long form'.
14542 * New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
14543 STRING types. These convert content octets to and from the
14544 underlying type. The actual tag and length octets are
14545 already assumed to have been read in and checked. These
14546 are needed because all other string types have virtually
14547 identical handling apart from the tag. By having versions
14548 of the ASN1 functions that just operate on content octets
14549 IMPLICIT tagging can be handled properly. It also allows
14550 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14551 and ASN1_INTEGER are identical apart from the tag.
14555 * Change the handling of OID objects as follows:
14557 - New object identifiers are inserted in objects.txt, following
14558 the syntax given in objects.README.
14559 - objects.pl is used to process obj_mac.num and create a new
14561 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14564 This is currently kind of a hack, and the perl code in objects.pl
14565 isn't very elegant, but it works as I intended. The simplest way
14566 to check that it worked correctly is to look in obj_dat.h and
14567 check the array nid_objs and make sure the objects haven't moved
14568 around (this is important!). Additions are OK, as well as
14569 consistent name changes.
14573 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14577 * Addition of the command line parameter '-rand file' to 'openssl req'.
14578 The given file adds to whatever has already been seeded into the
14579 random pool through the RANDFILE configuration file option or
14580 environment variable, or the default random state file.
14584 * mkstack.pl now sorts each macro group into lexical order.
14585 Previously the output order depended on the order the files
14586 appeared in the directory, resulting in needless rewriting
14591 * Patches to make OpenSSL compile under Win32 again. Mostly
14592 work arounds for the VC++ problem that it treats func() as
14593 func(void). Also stripped out the parts of mkdef.pl that
14594 added extra typesafe functions: these no longer exist.
14598 * Reorganisation of the stack code. The macros are now all
14599 collected in safestack.h . Each macro is defined in terms of
14600 a "stack macro" of the form SKM_<name>(type, a, b). The
14601 DEBUG_SAFESTACK is now handled in terms of function casts,
14602 this has the advantage of retaining type safety without the
14603 use of additional functions. If DEBUG_SAFESTACK is not defined
14604 then the non typesafe macros are used instead. Also modified the
14605 mkstack.pl script to handle the new form. Needs testing to see
14606 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14607 the default if no major problems. Similar behaviour for ASN1_SET_OF
14608 and PKCS12_STACK_OF.
14612 * When some versions of IIS use the 'NET' form of private key the
14613 key derivation algorithm is different. Normally MD5(password) is
14614 used as a 128 bit RC4 key. In the modified case
14615 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14616 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14617 as the old Netscape_RSA functions except they have an additional
14618 'sgckey' parameter which uses the modified algorithm. Also added
14619 an -sgckey command line option to the rsa utility. Thanks to
14620 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14621 algorithm to openssl-dev.
14625 * The evp_local.h macros were using 'c.##kname' which resulted in
14626 invalid expansion on some systems (SCO 5.0.5 for example).
14627 Corrected to 'c.kname'.
14629 *Phillip Porch <root@theporch.com>*
14631 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14632 a STACK of email addresses from a certificate or request, these look
14633 in the subject name and the subject alternative name extensions and
14634 omit any duplicate addresses.
14638 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14639 This makes DSA verification about 2 % faster.
14643 * Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
14644 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14645 plus overhead for 1024 bit moduli).
14646 This makes exponentiations about 0.5 % faster for 1024 bit
14647 exponents (as measured by "openssl speed rsa2048").
14651 * Rename memory handling macros to avoid conflicts with other
14653 Malloc => OPENSSL_malloc
14654 Malloc_locked => OPENSSL_malloc_locked
14655 Realloc => OPENSSL_realloc
14656 Free => OPENSSL_free
14660 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14661 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14665 * CygWin32 support.
14667 *John Jarvie <jjarvie@newsguy.com>*
14669 * The type-safe stack code has been rejigged. It is now only compiled
14670 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14671 by default all type-specific stack functions are "#define"d back to
14672 standard stack functions. This results in more streamlined output
14673 but retains the type-safety checking possibilities of the original
14678 * The STACK code has been cleaned up, and certain type declarations
14679 that didn't make a lot of sense have been brought in line. This has
14680 also involved a cleanup of sorts in safestack.h to more correctly
14681 map type-safe stack functions onto their plain stack counterparts.
14682 This work has also resulted in a variety of "const"ifications of
14683 lots of the code, especially "_cmp" operations which should normally
14684 be prototyped with "const" parameters anyway.
14688 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14689 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14690 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14691 where all of 'md' is used each time the PRNG is used, but 'state'
14692 is used only indexed by a cyclic counter. As entropy may not be
14693 well distributed from the beginning, 'md' is important as a
14694 chaining variable. However, the output function chains only half
14695 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14696 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14697 in all of 'state' being rewritten, with the new values depending
14698 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14702 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14703 the handshake is continued after ssl_verify_cert_chain();
14704 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14705 can lead to 'unexplainable' connection aborts later.
14707 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14709 * Major EVP API cipher revision.
14710 Add hooks for extra EVP features. This allows various cipher
14711 parameters to be set in the EVP interface. Support added for variable
14712 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14713 setting of RC2 and RC5 parameters.
14715 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14718 Remove lots of duplicated code from the EVP library. For example *every*
14719 cipher init() function handles the 'iv' in the same way according to the
14720 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14721 for CFB and OFB modes they zero ctx->num.
14723 New functionality allows removal of S/MIME code RC2 hack.
14725 Most of the routines have the same form and so can be declared in terms
14728 By shifting this to the top level EVP_CipherInit() it can be removed from
14729 all individual ciphers. If the cipher wants to handle IVs or keys
14730 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14733 Change lots of functions like EVP_EncryptUpdate() to now return a
14734 value: although software versions of the algorithms cannot fail
14735 any installed hardware versions can.
14739 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14740 this option is set, tolerate broken clients that send the negotiated
14741 protocol version number instead of the requested protocol version
14746 * Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
14747 i.e. non-zero for export ciphersuites, zero otherwise.
14748 Previous versions had this flag inverted, inconsistent with
14749 rsa_tmp_cb (..._TMP_RSA_CB).
14751 *Bodo Moeller; problem reported by Amit Chopra*
14753 * Add missing DSA library text string. Work around for some IIS
14754 key files with invalid SEQUENCE encoding.
14758 * Add a document (doc/standards.txt) that list all kinds of standards
14759 and so on that are implemented in OpenSSL.
14763 * Enhance c_rehash script. Old version would mishandle certificates
14764 with the same subject name hash and wouldn't handle CRLs at all.
14765 Added -fingerprint option to crl utility, to support new c_rehash
14770 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14774 * Fix for SSL server purpose checking. Server checking was
14775 rejecting certificates which had extended key usage present
14776 but no ssl client purpose.
14778 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14780 * Make PKCS#12 code work with no password. The PKCS#12 spec
14781 is a little unclear about how a blank password is handled.
14782 Since the password in encoded as a BMPString with terminating
14783 double NULL a zero length password would end up as just the
14784 double NULL. However no password at all is different and is
14785 handled differently in the PKCS#12 key generation code. NS
14786 treats a blank password as zero length. MSIE treats it as no
14787 password on export: but it will try both on import. We now do
14788 the same: PKCS12_parse() tries zero length and no password if
14789 the password is set to "" or NULL (NULL is now a valid password:
14790 it wasn't before) as does the pkcs12 application.
14794 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14795 perror when PEM_read_bio_X509_REQ fails, the error message must
14796 be obtained from the error queue.
14800 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14801 it in ERR_remove_state if appropriate, and change ERR_get_state
14802 accordingly to avoid race conditions (this is necessary because
14803 thread_hash is no longer constant once set).
14807 * Bugfix for linux-elf makefile.one.
14811 * RSA_get_default_method() will now cause a default
14812 RSA_METHOD to be chosen if one doesn't exist already.
14813 Previously this was only set during a call to RSA_new()
14814 or RSA_new_method(NULL) meaning it was possible for
14815 RSA_get_default_method() to return NULL.
14819 * Added native name translation to the existing DSO code
14820 that will convert (if the flag to do so is set) filenames
14821 that are sufficiently small and have no path information
14822 into a canonical native form. Eg. "blah" converted to
14823 "libblah.so" or "blah.dll" etc.
14827 * New function ERR_error_string_n(e, buf, len) which is like
14828 ERR_error_string(e, buf), but writes at most 'len' bytes
14829 including the 0 terminator. For ERR_error_string_n, 'buf'
14832 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14834 * CONF library reworked to become more general. A new CONF
14835 configuration file reader "class" is implemented as well as a
14836 new functions (NCONF_*, for "New CONF") to handle it. The now
14837 old CONF_* functions are still there, but are reimplemented to
14838 work in terms of the new functions. Also, a set of functions
14839 to handle the internal storage of the configuration data is
14840 provided to make it easier to write new configuration file
14841 reader "classes" (I can definitely see something reading a
14842 configuration file in XML format, for example), called _CONF_*,
14843 or "the configuration storage API"...
14845 The new configuration file reading functions are:
14847 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14848 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14850 NCONF_default, NCONF_WIN32
14852 NCONF_dump_fp, NCONF_dump_bio
14854 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14855 NCONF_new creates a new CONF object. This works in the same way
14856 as other interfaces in OpenSSL, like the BIO interface.
14857 NCONF_dump_* dump the internal storage of the configuration file,
14858 which is useful for debugging. All other functions take the same
14859 arguments as the old CONF_* functions with the exception of the
14860 first that must be a `CONF *' instead of a `LHASH *'.
14862 To make it easier to use the new classes with the old CONF_* functions,
14863 the function CONF_set_default_method is provided.
14867 * Add '-tls1' option to 'openssl ciphers', which was already
14868 mentioned in the documentation but had not been implemented.
14869 (This option is not yet really useful because even the additional
14870 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14874 * Initial DSO code added into libcrypto for letting OpenSSL (and
14875 OpenSSL-based applications) load shared libraries and bind to
14876 them in a portable way.
14878 *Geoff Thorpe, with contributions from Richard Levitte*
14880 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000] ###
14882 * Make sure _lrotl and _lrotr are only used with MSVC.
14884 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14885 (the default implementation of RAND_status).
14887 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14888 to '-clrext' (= clear extensions), as intended and documented.
14889 *Bodo Moeller; inconsistency pointed out by Michael Attili
14890 <attili@amaxo.com>*
14892 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14893 was larger than the MD block size.
14895 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14897 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14898 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14899 using the passed key: if the passed key was a private key the result
14900 of X509_print(), for example, would be to print out all the private key
14905 * des_quad_cksum() byte order bug fix.
14906 *Ulf Möller, using the problem description in krb4-0.9.7, where
14907 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14909 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14912 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14914 * For easily testing in shell scripts whether some command
14915 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14916 returns with exit code 0 iff no command of the given name is available.
14917 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14918 the output goes to stdout and nothing is printed to stderr.
14919 Additional arguments are always ignored.
14921 Since for each cipher there is a command of the same name,
14922 the 'no-cipher' compilation switches can be tested this way.
14924 ('openssl no-XXX' is not able to detect pseudo-commands such
14925 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
14929 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
14933 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
14934 is set; it will be thrown away anyway because each handshake creates
14936 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
14937 to parameters -- in previous versions (since OpenSSL 0.9.3) the
14938 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
14939 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
14943 * New s_client option -ign_eof: EOF at stdin is ignored, and
14944 'Q' and 'R' lose their special meanings (quit/renegotiate).
14945 This is part of what -quiet does; unlike -quiet, -ign_eof
14946 does not suppress any output.
14950 * Add compatibility options to the purpose and trust code. The
14951 purpose X509_PURPOSE_ANY is "any purpose" which automatically
14952 accepts a certificate or CA, this was the previous behaviour,
14953 with all the associated security issues.
14955 X509_TRUST_COMPAT is the old trust behaviour: only and
14956 automatically trust self signed roots in certificate store. A
14957 new trust setting X509_TRUST_DEFAULT is used to specify that
14958 a purpose has no associated trust setting and it should instead
14959 use the value in the default purpose.
14963 * Fix the PKCS#8 DSA private key code so it decodes keys again
14964 and fix a memory leak.
14968 * In util/mkerr.pl (which implements 'make errors'), preserve
14969 reason strings from the previous version of the .c file, as
14970 the default to have only downcase letters (and digits) in
14971 automatically generated reasons codes is not always appropriate.
14975 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
14976 using strerror. Previously, ERR_reason_error_string() returned
14977 library names as reason strings for SYSerr; but SYSerr is a special
14978 case where small numbers are errno values, not library numbers.
14982 * Add '-dsaparam' option to 'openssl dhparam' application. This
14983 converts DSA parameters into DH parameters. (When creating parameters,
14984 DSA_generate_parameters is used.)
14988 * Include 'length' (recommended exponent length) in C code generated
14989 by 'openssl dhparam -C'.
14993 * The second argument to set_label in perlasm was already being used
14994 so couldn't be used as a "file scope" flag. Moved to third argument
14999 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15000 instead of RAND_bytes for encryption IVs and salts.
15004 * Include RAND_status() into RAND_METHOD instead of implementing
15005 it only for md_rand.c Otherwise replacing the PRNG by calling
15006 RAND_set_rand_method would be impossible.
15010 * Don't let DSA_generate_key() enter an infinite loop if the random
15011 number generation fails.
15015 * New 'rand' application for creating pseudo-random output.
15019 * Added configuration support for Linux/IA64
15021 *Rolf Haberrecker <rolf@suse.de>*
15023 * Assembler module support for Mingw32.
15027 * Shared library support for HPUX (in shlib/).
15029 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15031 * Shared library support for Solaris gcc.
15033 *Lutz Behnke <behnke@trustcenter.de>*
15035 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000] ###
15037 * PKCS7_encrypt() was adding text MIME headers twice because they
15038 were added manually and by SMIME_crlf_copy().
15042 * In bntest.c don't call BN_rand with zero bits argument.
15044 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15046 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15047 case was implemented. This caused BN_div_recp() to fail occasionally.
15051 * Add an optional second argument to the set_label() in the perl
15052 assembly language builder. If this argument exists and is set
15053 to 1 it signals that the assembler should use a symbol whose
15054 scope is the entire file, not just the current function. This
15055 is needed with MASM which uses the format label:: for this scope.
15057 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15059 * Change the ASN1 types so they are typedefs by default. Before
15060 almost all types were #define'd to ASN1_STRING which was causing
15061 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15066 * Change names of new functions to the new get1/get0 naming
15067 convention: After 'get1', the caller owns a reference count
15068 and has to call ..._free; 'get0' returns a pointer to some
15069 data structure without incrementing reference counters.
15070 (Some of the existing 'get' functions increment a reference
15071 counter, some don't.)
15072 Similarly, 'set1' and 'add1' functions increase reference
15073 counters or duplicate objects.
15077 * Allow for the possibility of temp RSA key generation failure:
15078 the code used to assume it always worked and crashed on failure.
15082 * Fix potential buffer overrun problem in BIO_printf().
15083 *Ulf Möller, using public domain code by Patrick Powell; problem
15084 pointed out by David Sacerdote <das33@cornell.edu>*
15086 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15087 RAND_egd() and RAND_status(). In the command line application,
15088 the EGD socket can be specified like a seed file using RANDFILE
15093 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15094 Some CAs (e.g. Verisign) distribute certificates in this form.
15098 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15099 list to exclude them. This means that no special compilation option
15100 is needed to use anonymous DH: it just needs to be included in the
15105 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15106 EVP_MD_type. The old functionality is available in a new macro called
15107 EVP_MD_md(). Change code that uses it and update docs.
15111 * ..._ctrl functions now have corresponding ..._callback_ctrl functions
15112 where the 'void *' argument is replaced by a function pointer argument.
15113 Previously 'void *' was abused to point to functions, which works on
15114 many platforms, but is not correct. As these functions are usually
15115 called by macros defined in OpenSSL header files, most source code
15116 should work without changes.
15120 * <openssl/opensslconf.h> (which is created by Configure) now contains
15121 sections with information on -D... compiler switches used for
15122 compiling the library so that applications can see them. To enable
15123 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
15124 must be defined. E.g.,
15125 #define OPENSSL_ALGORITHM_DEFINES
15126 #include <openssl/opensslconf.h>
15127 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
15129 *Richard Levitte, Ulf and Bodo Möller*
15131 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15136 * Change the 'other' type in certificate aux info to a STACK_OF
15137 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15138 the required ASN1 format: arbitrary types determined by an OID.
15142 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15143 argument to 'req'. This is not because the function is newer or
15144 better than others it just uses the work 'NEW' in the certificate
15145 request header lines. Some software needs this.
15149 * Reorganise password command line arguments: now passwords can be
15150 obtained from various sources. Delete the PEM_cb function and make
15151 it the default behaviour: i.e. if the callback is NULL and the
15152 usrdata argument is not NULL interpret it as a null terminated pass
15153 phrase. If usrdata and the callback are NULL then the pass phrase
15154 is prompted for as usual.
15158 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15159 the support is automatically enabled. The resulting binaries will
15160 autodetect the card and use it if present.
15162 *Ben Laurie and Compaq Inc.*
15164 * Work around for Netscape hang bug. This sends certificate request
15165 and server done in one record. Since this is perfectly legal in the
15166 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15167 the bugs/SSLv3 entry for more info.
15171 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15175 * Add -rand argument to smime and pkcs12 applications and read/write
15180 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15184 * Add command line password options to the remaining applications.
15188 * Bug fix for BN_div_recp() for numerators with an even number of
15193 * More tests in bntest.c, and changed test_bn output.
15197 * ./config recognizes MacOS X now.
15201 * Bug fix for BN_div() when the first words of num and divisor are
15202 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
15206 * Add support for various broken PKCS#8 formats, and command line
15207 options to produce them.
15211 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15212 get temporary BIGNUMs from a BN_CTX.
15216 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15221 * Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
15222 include a #define from the old name to the new. The original intent
15223 was that statically linked binaries could for example just call
15224 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15225 link with digests. This never worked because SSLeay_add_all_digests()
15226 and SSLeay_add_all_ciphers() were in the same source file so calling
15227 one would link with the other. They are now in separate source files.
15231 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15235 * Use a less unusual form of the Miller-Rabin primality test (it used
15236 a binary algorithm for exponentiation integrated into the Miller-Rabin
15237 loop, our standard modexp algorithms are faster).
15241 * Support for the EBCDIC character set completed.
15243 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15245 * Source code cleanups: use const where appropriate, eliminate casts,
15246 use void * instead of char * in lhash.
15250 * Bugfix: ssl3_send_server_key_exchange was not restartable
15251 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15252 this the server could overwrite ephemeral keys that the client
15257 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15258 using 50 iterations of the Rabin-Miller test.
15260 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15261 iterations of the Rabin-Miller test as required by the appendix
15262 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15263 As BN_is_prime_fasttest includes trial division, DSA parameter
15264 generation becomes much faster.
15266 This implies a change for the callback functions in DSA_is_prime
15267 and DSA_generate_parameters: The callback function is called once
15268 for each positive witness in the Rabin-Miller test, not just
15269 occasionally in the inner loop; and the parameters to the
15270 callback function now provide an iteration count for the outer
15271 loop rather than for the current invocation of the inner loop.
15272 DSA_generate_parameters additionally can call the callback
15273 function with an 'iteration count' of -1, meaning that a
15274 candidate has passed the trial division test (when q is generated
15275 from an application-provided seed, trial division is skipped).
15279 * New function BN_is_prime_fasttest that optionally does trial
15280 division before starting the Rabin-Miller test and has
15281 an additional BN_CTX * argument (whereas BN_is_prime always
15282 has to allocate at least one BN_CTX).
15283 'callback(1, -1, cb_arg)' is called when a number has passed the
15284 trial division stage.
15288 * Fix for bug in CRL encoding. The validity dates weren't being handled
15293 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15297 * New function BN_pseudo_rand().
15301 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15302 bignum version of BN_from_montgomery() with the working code from
15303 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15308 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15309 made it impossible to use the same SSL_SESSION data structure in
15310 SSL2 clients in multiple threads.
15314 * The return value of RAND_load_file() no longer counts bytes obtained
15315 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15316 to seed the PRNG (previously an explicit byte count was required).
15318 *Ulf Möller, Bodo Möller*
15320 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15321 used (char *) instead of (void *) and had casts all over the place.
15325 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15329 * Retain source code compatibility for BN_prime_checks macro:
15330 BN_is_prime(..., BN_prime_checks, ...) now uses
15331 BN_prime_checks_for_size to determine the appropriate number of
15332 Rabin-Miller iterations.
15336 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15337 DH_CHECK_P_NOT_SAFE_PRIME.
15338 (Check if this is true? OpenPGP calls them "strong".)
15342 * Merge the functionality of "dh" and "gendh" programs into a new program
15343 "dhparam". The old programs are retained for now but will handle DH keys
15344 (instead of parameters) in future.
15348 * Make the ciphers, s_server and s_client programs check the return values
15349 when a new cipher list is set.
15353 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15354 ciphers. Before when the 56bit ciphers were enabled the sorting was
15357 The syntax for the cipher sorting has been extended to support sorting by
15358 cipher-strength (using the strength_bits hard coded in the tables).
15359 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15361 Fix a bug in the cipher-command parser: when supplying a cipher command
15362 string with an "undefined" symbol (neither command nor alphanumeric
15363 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15364 an error is flagged.
15366 Due to the strength-sorting extension, the code of the
15367 ssl_create_cipher_list() function was completely rearranged. I hope that
15368 the readability was also increased :-)
15370 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15372 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15373 for the first serial number and places 2 in the serial number file. This
15374 avoids problems when the root CA is created with serial number zero and
15375 the first user certificate has the same issuer name and serial number
15380 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15381 the new code. Add documentation for this stuff.
15385 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15386 X509_*() to X509at_*() on the grounds that they don't handle X509
15387 structures and behave in an analogous way to the X509v3 functions:
15388 they shouldn't be called directly but wrapper functions should be used
15391 So we also now have some wrapper functions that call the X509at functions
15392 when passed certificate requests. (TO DO: similar things can be done with
15393 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15394 things. Some of these need some d2i or i2d and print functionality
15395 because they handle more complex structures.)
15399 * Add missing #ifndefs that caused missing symbols when building libssl
15400 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15401 NO_RSA in ssl/s2*.c.
15403 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15405 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15406 has a return value which indicates the quality of the random data
15407 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15408 error queue. New function RAND_pseudo_bytes() generates output that is
15409 guaranteed to be unique but not unpredictable. RAND_add is like
15410 RAND_seed, but takes an extra argument for an entropy estimate
15411 (RAND_seed always assumes full entropy).
15415 * Do more iterations of Rabin-Miller probable prime test (specifically,
15416 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15417 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15418 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15419 false-positive rate of at most 2^-80 for random input.
15423 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15427 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15428 in the 0.9.5 release), this returns the chain
15429 from an X509_CTX structure with a dup of the stack and all
15430 the X509 reference counts upped: so the stack will exist
15431 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15434 Also make SSL_SESSION_print() print out the verify return
15439 * Add manpage for the pkcs12 command. Also change the default
15440 behaviour so MAC iteration counts are used unless the new
15441 -nomaciter option is used. This improves file security and
15442 only older versions of MSIE (4.0 for example) need it.
15446 * Honor the no-xxx Configure options when creating .DEF files.
15450 * Add PKCS#10 attributes to field table: challengePassword,
15451 unstructuredName and unstructuredAddress. These are taken from
15452 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15453 international characters are used.
15455 More changes to X509_ATTRIBUTE code: allow the setting of types
15456 based on strings. Remove the 'loc' parameter when adding
15457 attributes because these will be a SET OF encoding which is sorted
15462 * Initial changes to the 'req' utility to allow request generation
15463 automation. This will allow an application to just generate a template
15464 file containing all the field values and have req construct the
15467 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15468 used all over the place including certificate requests and PKCS#7
15469 structures. They are currently handled manually where necessary with
15470 some primitive wrappers for PKCS#7. The new functions behave in a
15471 manner analogous to the X509 extension functions: they allow
15472 attributes to be looked up by NID and added.
15474 Later something similar to the X509V3 code would be desirable to
15475 automatically handle the encoding, decoding and printing of the
15476 more complex types. The string types like challengePassword can
15477 be handled by the string table functions.
15479 Also modified the multi byte string table handling. Now there is
15480 a 'global mask' which masks out certain types. The table itself
15481 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15482 is useful when for example there is only one permissible type
15483 (as in countryName) and using the mask might result in no valid
15488 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15489 SSL_get_peer_finished to allow applications to obtain the latest
15490 Finished messages sent to the peer or expected from the peer,
15491 respectively. (SSL_get_peer_finished is usually the Finished message
15492 actually received from the peer, otherwise the protocol will be aborted.)
15494 As the Finished message are message digests of the complete handshake
15495 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15496 be used for external authentication procedures when the authentication
15497 provided by SSL/TLS is not desired or is not enough.
15501 * Enhanced support for Alpha Linux is added. Now ./config checks if
15502 the host supports BWX extension and if Compaq C is present on the
15503 $PATH. Just exploiting of the BWX extension results in 20-30%
15504 performance kick for some algorithms, e.g. DES and RC4 to mention
15505 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15510 * Add support for MS "fast SGC". This is arguably a violation of the
15511 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15512 weak crypto and after checking the certificate is SGC a second one
15513 with strong crypto. MS SGC stops the first handshake after receiving
15514 the server certificate message and sends a second client hello. Since
15515 a server will typically do all the time consuming operations before
15516 expecting any further messages from the client (server key exchange
15517 is the most expensive) there is little difference between the two.
15519 To get OpenSSL to support MS SGC we have to permit a second client
15520 hello message after we have sent server done. In addition we have to
15521 reset the MAC if we do get this second client hello.
15525 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15526 if a DER encoded private key is RSA or DSA traditional format. Changed
15527 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15528 format DER encoded private key. Newer code should use PKCS#8 format which
15529 has the key type encoded in the ASN1 structure. Added DER private key
15530 support to pkcs8 application.
15534 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15535 ciphersuites has been selected (as required by the SSL 3/TLS 1
15536 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15537 is set, we interpret this as a request to violate the specification
15538 (the worst that can happen is a handshake failure, and 'correct'
15539 behaviour would result in a handshake failure anyway).
15543 * In SSL_CTX_add_session, take into account that there might be multiple
15544 SSL_SESSION structures with the same session ID (e.g. when two threads
15545 concurrently obtain them from an external cache).
15546 The internal cache can handle only one SSL_SESSION with a given ID,
15547 so if there's a conflict, we now throw out the old one to achieve
15552 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15553 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15554 some routines that use cipher OIDs: some ciphers do not have OIDs
15555 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15560 * Simplify the trust setting structure and code. Now we just have
15561 two sequences of OIDs for trusted and rejected settings. These will
15562 typically have values the same as the extended key usage extension
15563 and any application specific purposes.
15565 The trust checking code now has a default behaviour: it will just
15566 check for an object with the same NID as the passed id. Functions can
15567 be provided to override either the default behaviour or the behaviour
15568 for a given id. SSL client, server and email already have functions
15569 in place for compatibility: they check the NID and also return "trusted"
15570 if the certificate is self signed.
15574 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15575 traditional format into an EVP_PKEY structure.
15579 * Add a password callback function PEM_cb() which either prompts for
15580 a password if usr_data is NULL or otherwise assumes it is a null
15581 terminated password. Allow passwords to be passed on command line
15582 environment or config files in a few more utilities.
15586 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15587 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15588 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15589 Update documentation.
15593 * Support for ASN1 "NULL" type. This could be handled before by using
15594 ASN1_TYPE but there wasn't any function that would try to read a NULL
15595 and produce an error if it couldn't. For compatibility we also have
15596 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15597 don't allocate anything because they don't need to.
15601 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15604 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15606 * Rebuild of the memory allocation routines used by OpenSSL code and
15607 possibly others as well. The purpose is to make an interface that
15608 provide hooks so anyone can build a separate set of allocation and
15609 deallocation routines to be used by OpenSSL, for example memory
15610 pool implementations, or something else, which was previously hard
15611 since Malloc(), Realloc() and Free() were defined as macros having
15612 the values malloc, realloc and free, respectively (except for Win32
15613 compilations). The same is provided for memory debugging code.
15614 OpenSSL already comes with functionality to find memory leaks, but
15615 this gives people a chance to debug other memory problems.
15617 With these changes, a new set of functions and macros have appeared:
15619 CRYPTO_set_mem_debug_functions() [F]
15620 CRYPTO_get_mem_debug_functions() [F]
15621 CRYPTO_dbg_set_options() [F]
15622 CRYPTO_dbg_get_options() [F]
15623 CRYPTO_malloc_debug_init() [M]
15625 The memory debug functions are NULL by default, unless the library
15626 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15627 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15628 gives the standard debugging functions that come with OpenSSL) or
15629 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15630 provided by the library user) must be used. When the standard
15631 debugging functions are used, CRYPTO_dbg_set_options can be used to
15632 request additional information:
15633 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15634 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15636 Also, things like CRYPTO_set_mem_functions will always give the
15637 expected result (the new set of functions is used for allocation
15638 and deallocation) at all times, regardless of platform and compiler
15641 To finish it up, some functions that were never use in any other
15642 way than through macros have a new API and new semantic:
15644 CRYPTO_dbg_malloc()
15645 CRYPTO_dbg_realloc()
15648 All macros of value have retained their old syntax.
15650 *Richard Levitte and Bodo Moeller*
15652 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15653 ordering of SMIMECapabilities wasn't in "strength order" and there
15654 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15659 * Some ASN1 types with illegal zero length encoding (INTEGER,
15660 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15662 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15664 * Merge in my S/MIME library for OpenSSL. This provides a simple
15665 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15666 functionality to handle multipart/signed properly) and a utility
15667 called 'smime' to call all this stuff. This is based on code I
15668 originally wrote for Celo who have kindly allowed it to be
15669 included in OpenSSL.
15673 * Add variants des_set_key_checked and des_set_key_unchecked of
15674 des_set_key (aka des_key_sched). Global variable des_check_key
15675 decides which of these is called by des_set_key; this way
15676 des_check_key behaves as it always did, but applications and
15677 the library itself, which was buggy for des_check_key == 1,
15678 have a cleaner way to pick the version they need.
15682 * New function PKCS12_newpass() which changes the password of a
15687 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15688 dynamic mix. In both cases the ids can be used as an index into the
15689 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15690 functions so they accept a list of the field values and the
15691 application doesn't need to directly manipulate the X509_TRUST
15696 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15701 * Modify the way the V3 extension code looks up extensions. This now
15702 works in a similar way to the object code: we have some "standard"
15703 extensions in a static table which is searched with OBJ_bsearch()
15704 and the application can add dynamic ones if needed. The file
15705 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15706 updated whenever a new extension is added to the core code and kept
15707 in ext_nid order. There is a simple program 'tabtest.c' which checks
15708 this. New extensions are not added too often so this file can readily
15709 be maintained manually.
15711 There are two big advantages in doing things this way. The extensions
15712 can be looked up immediately and no longer need to be "added" using
15713 X509V3_add_standard_extensions(): this function now does nothing.
15714 *Side note: I get *lots* of email saying the extension code doesn't
15715 work because people forget to call this function*
15716 Also no dynamic allocation is done unless new extensions are added:
15717 so if we don't add custom extensions there is no need to call
15718 X509V3_EXT_cleanup().
15722 * Modify enc utility's salting as follows: make salting the default. Add a
15723 magic header, so unsalted files fail gracefully instead of just decrypting
15724 to garbage. This is because not salting is a big security hole, so people
15725 should be discouraged from doing it.
15729 * Fixes and enhancements to the 'x509' utility. It allowed a message
15730 digest to be passed on the command line but it only used this
15731 parameter when signing a certificate. Modified so all relevant
15732 operations are affected by the digest parameter including the
15733 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15734 DSA key was used because it didn't fix the digest.
15738 * Initial certificate chain verify code. Currently tests the untrusted
15739 certificates for consistency with the verify purpose (which is set
15740 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15742 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15743 this is because it will reject chains with invalid extensions whereas
15744 every previous version of OpenSSL and SSLeay made no checks at all.
15746 Trust code: checks the root CA for the relevant trust settings. Trust
15747 settings have an initial value consistent with the verify purpose: e.g.
15748 if the verify purpose is for SSL client use it expects the CA to be
15749 trusted for SSL client use. However the default value can be changed to
15750 permit custom trust settings: one example of this would be to only trust
15751 certificates from a specific "secure" set of CAs.
15753 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15754 which should be used for version portability: especially since the
15755 verify structure is likely to change more often now.
15757 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15758 to set them. If not set then assume SSL clients will verify SSL servers
15761 Two new options to the verify program: -untrusted allows a set of
15762 untrusted certificates to be passed in and -purpose which sets the
15763 intended purpose of the certificate. If a purpose is set then the
15764 new chain verify code is used to check extension consistency.
15768 * Support for the authority information access extension.
15772 * Modify RSA and DSA PEM read routines to transparently handle
15773 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15774 public keys in a format compatible with certificate
15775 SubjectPublicKeyInfo structures. Unfortunately there were already
15776 functions called *_PublicKey_* which used various odd formats so
15777 these are retained for compatibility: however the DSA variants were
15778 never in a public release so they have been deleted. Changed dsa/rsa
15779 utilities to handle the new format: note no releases ever handled public
15780 keys so we should be OK.
15782 The primary motivation for this change is to avoid the same fiasco
15783 that dogs private keys: there are several incompatible private key
15784 formats some of which are standard and some OpenSSL specific and
15785 require various evil hacks to allow partial transparent handling and
15786 even then it doesn't work with DER formats. Given the option anything
15787 other than PKCS#8 should be dumped: but the other formats have to
15788 stay in the name of compatibility.
15790 With public keys and the benefit of hindsight one standard format
15791 is used which works with EVP_PKEY, RSA or DSA structures: though
15792 it clearly returns an error if you try to read the wrong kind of key.
15794 Added a -pubkey option to the 'x509' utility to output the public key.
15795 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
15796 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
15797 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
15798 that do the same as the EVP_PKEY_assign_*() except they up the
15799 reference count of the added key (they don't "swallow" the
15804 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15805 CRLs would fail if the file contained no certificates or no CRLs:
15806 added a new function to read in both types and return the number
15807 read: this means that if none are read it will be an error. The
15808 DER versions of the certificate and CRL reader would always fail
15809 because it isn't possible to mix certificates and CRLs in DER format
15810 without choking one or the other routine. Changed this to just read
15811 a certificate: this is the best we can do. Also modified the code
15812 in apps/verify.c to take notice of return codes: it was previously
15813 attempting to read in certificates from NULL pointers and ignoring
15814 any errors: this is one reason why the cert and CRL reader seemed
15815 to work. It doesn't check return codes from the default certificate
15816 routines: these may well fail if the certificates aren't installed.
15820 * Code to support otherName option in GeneralName.
15824 * First update to verify code. Change the verify utility
15825 so it warns if it is passed a self signed certificate:
15826 for consistency with the normal behaviour. X509_verify
15827 has been modified to it will now verify a self signed
15828 certificate if *exactly* the same certificate appears
15829 in the store: it was previously impossible to trust a
15830 single self signed certificate. This means that:
15831 openssl verify ss.pem
15832 now gives a warning about a self signed certificate but
15833 openssl verify -CAfile ss.pem ss.pem
15838 * For servers, store verify_result in SSL_SESSION data structure
15839 (and add it to external session representation).
15840 This is needed when client certificate verifications fails,
15841 but an application-provided verification callback (set by
15842 SSL_CTX_set_cert_verify_callback) allows accepting the session
15843 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15844 but returns 1): When the session is reused, we have to set
15845 ssl->verify_result to the appropriate error code to avoid
15848 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15850 * Fix a bug in the new PKCS#7 code: it didn't consider the
15851 case in PKCS7_dataInit() where the signed PKCS7 structure
15852 didn't contain any existing data because it was being created.
15854 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15856 * Add a salt to the key derivation routines in enc.c. This
15857 forms the first 8 bytes of the encrypted file. Also add a
15858 -S option to allow a salt to be input on the command line.
15862 * New function X509_cmp(). Oddly enough there wasn't a function
15863 to compare two certificates. We do this by working out the SHA1
15864 hash and comparing that. X509_cmp() will be needed by the trust
15869 * SSL_get1_session() is like SSL_get_session(), but increments
15870 the reference count in the SSL_SESSION returned.
15872 *Geoff Thorpe <geoff@eu.c2.net>*
15874 * Fix for 'req': it was adding a null to request attributes.
15875 Also change the X509_LOOKUP and X509_INFO code to handle
15876 certificate auxiliary information.
15880 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15885 * Add the possibility to add extra information to the memory leak
15886 detecting output, to form tracebacks, showing from where each
15887 allocation was originated: CRYPTO_push_info("constant string") adds
15888 the string plus current file name and line number to a per-thread
15889 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15890 is like calling CYRPTO_pop_info() until the stack is empty.
15891 Also updated memory leak detection code to be multi-thread-safe.
15895 * Add options -text and -noout to pkcs7 utility and delete the
15896 encryption options which never did anything. Update docs.
15900 * Add options to some of the utilities to allow the pass phrase
15901 to be included on either the command line (not recommended on
15902 OSes like Unix) or read from the environment. Update the
15903 manpages and fix a few bugs.
15907 * Add a few manpages for some of the openssl commands.
15911 * Fix the -revoke option in ca. It was freeing up memory twice,
15912 leaking and not finding already revoked certificates.
15916 * Extensive changes to support certificate auxiliary information.
15917 This involves the use of X509_CERT_AUX structure and X509_AUX
15918 functions. An X509_AUX function such as PEM_read_X509_AUX()
15919 can still read in a certificate file in the usual way but it
15920 will also read in any additional "auxiliary information". By
15921 doing things this way a fair degree of compatibility can be
15922 retained: existing certificates can have this information added
15923 using the new 'x509' options.
15925 Current auxiliary information includes an "alias" and some trust
15926 settings. The trust settings will ultimately be used in enhanced
15927 certificate chain verification routines: currently a certificate
15928 can only be trusted if it is self signed and then it is trusted
15933 * Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
15934 The problem was that one of the replacement routines had not been working
15935 since SSLeay releases. For now the offending routine has been replaced
15936 with non-optimised assembler. Even so, this now gives around 95%
15937 performance improvement for 1024 bit RSA signs.
15941 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
15942 handling. Most clients have the effective key size in bits equal to
15943 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
15944 A few however don't do this and instead use the size of the decrypted key
15945 to determine the RC2 key length and the AlgorithmIdentifier to determine
15946 the effective key length. In this case the effective key length can still
15947 be 40 bits but the key length can be 168 bits for example. This is fixed
15948 by manually forcing an RC2 key into the EVP_PKEY structure because the
15949 EVP code can't currently handle unusual RC2 key sizes: it always assumes
15950 the key length and effective key length are equal.
15954 * Add a bunch of functions that should simplify the creation of
15955 X509_NAME structures. Now you should be able to do:
15956 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
15957 and have it automatically work out the correct field type and fill in
15958 the structures. The more adventurous can try:
15959 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
15960 and it will (hopefully) work out the correct multibyte encoding.
15964 * Change the 'req' utility to use the new field handling and multibyte
15965 copy routines. Before the DN field creation was handled in an ad hoc
15966 way in req, ca, and x509 which was rather broken and didn't support
15967 BMPStrings or UTF8Strings. Since some software doesn't implement
15968 BMPStrings or UTF8Strings yet, they can be enabled using the config file
15969 using the dirstring_type option. See the new comment in the default
15970 openssl.cnf for more info.
15974 * Make crypto/rand/md_rand.c more robust:
15975 - Assure unique random numbers after fork().
15976 - Make sure that concurrent threads access the global counter and
15977 md serializably so that we never lose entropy in them
15978 or use exactly the same state in multiple threads.
15979 Access to the large state is not always serializable because
15980 the additional locking could be a performance killer, and
15981 md should be large enough anyway.
15985 * New file apps/app_rand.c with commonly needed functionality
15986 for handling the random seed file.
15988 Use the random seed file in some applications that previously did not:
15990 dsaparam -genkey (which also ignored its '-rand' option),
15993 x509 (when signing).
15994 Except on systems with /dev/urandom, it is crucial to have a random
15995 seed file at least for key creation, DSA signing, and for DH exchanges;
15996 for RSA signatures we could do without one.
15998 gendh and gendsa (unlike genrsa) used to read only the first byte
15999 of each file listed in the '-rand' option. The function as previously
16000 found in genrsa is now in app_rand.c and is used by all programs
16001 that support '-rand'.
16005 * In RAND_write_file, use mode 0600 for creating files;
16006 don't just chmod when it may be too late.
16010 * Report an error from X509_STORE_load_locations
16011 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16015 * New function ASN1_mbstring_copy() this copies a string in either
16016 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16017 into an ASN1_STRING type. A mask of permissible types is passed
16018 and it chooses the "minimal" type to use or an error if not type
16023 * Add function equivalents to the various macros in asn1.h. The old
16024 macros are retained with an M_ prefix. Code inside the library can
16025 use the M_ macros. External code (including the openssl utility)
16026 should *NOT* in order to be "shared library friendly".
16030 * Add various functions that can check a certificate's extensions
16031 to see if it usable for various purposes such as SSL client,
16032 server or S/MIME and CAs of these types. This is currently
16033 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16034 verification. Also added a -purpose flag to x509 utility to
16035 print out all the purposes.
16039 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16044 * New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
16045 for, obtain and decode and extension and obtain its critical flag.
16046 This allows all the necessary extension code to be handled in a
16047 single function call.
16051 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16052 platforms. See crypto/rc4/rc4_enc.c for further details.
16056 * New -noout option to asn1parse. This causes no output to be produced
16057 its main use is when combined with -strparse and -out to extract data
16058 from a file (which may not be in ASN.1 format).
16062 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16063 when producing the local key id.
16065 *Richard Levitte <levitte@stacken.kth.se>*
16067 * New option -dhparam in s_server. This allows a DH parameter file to be
16068 stated explicitly. If it is not stated then it tries the first server
16069 certificate file. The previous behaviour hard coded the filename
16074 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16075 a public key to be input or output. For example:
16076 openssl rsa -in key.pem -pubout -out pubkey.pem
16077 Also added necessary DSA public key functions to handle this.
16081 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16082 in the message. This was handled by allowing
16083 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16085 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16087 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16088 to the end of the strings whereas this didn't. This would cause problems
16089 if strings read with d2i_ASN1_bytes() were later modified.
16091 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16093 * Fix for base64 decode bug. When a base64 bio reads only one line of
16094 data and it contains EOF it will end up returning an error. This is
16095 caused by input 46 bytes long. The cause is due to the way base64
16096 BIOs find the start of base64 encoded data. They do this by trying a
16097 trial decode on each line until they find one that works. When they
16098 do a flag is set and it starts again knowing it can pass all the
16099 data directly through the decoder. Unfortunately it doesn't reset
16100 the context it uses. This means that if EOF is reached an attempt
16101 is made to pass two EOFs through the context and this causes the
16102 resulting error. This can also cause other problems as well. As is
16103 usual with these problems it takes *ages* to find and the fix is
16104 trivial: move one line.
16106 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) *
16108 * Ugly workaround to get s_client and s_server working under Windows. The
16109 old code wouldn't work because it needed to select() on sockets and the
16110 tty (for keypresses and to see if data could be written). Win32 only
16111 supports select() on sockets so we select() with a 1s timeout on the
16112 sockets and then see if any characters are waiting to be read, if none
16113 are present then we retry, we also assume we can always write data to
16114 the tty. This isn't nice because the code then blocks until we've
16115 received a complete line of data and it is effectively polling the
16116 keyboard at 1s intervals: however it's quite a bit better than not
16117 working at all :-) A dedicated Windows application might handle this
16118 with an event loop for example.
16122 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16123 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16124 will be called when RSA_sign() and RSA_verify() are used. This is useful
16125 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16126 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16127 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16128 This necessitated the support of an extra signature type NID_md5_sha1
16129 for SSL signatures and modifications to the SSL library to use it instead
16130 of calling RSA_public_decrypt() and RSA_private_encrypt().
16134 * Add new -verify -CAfile and -CApath options to the crl program, these
16135 will lookup a CRL issuers certificate and verify the signature in a
16136 similar way to the verify program. Tidy up the crl program so it
16137 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16138 less strict. It will now permit CRL extensions even if it is not
16139 a V2 CRL: this will allow it to tolerate some broken CRLs.
16143 * Initialize all non-automatic variables each time one of the openssl
16144 sub-programs is started (this is necessary as they may be started
16145 multiple times from the "OpenSSL>" prompt).
16147 *Lennart Bang, Bodo Moeller*
16149 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16150 removing all other RSA functionality (this is what NO_RSA does). This
16151 is so (for example) those in the US can disable those operations covered
16152 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16157 * Non-copying interface to BIO pairs.
16158 (still largely untested)
16162 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16163 ASCII string. This was handled independently in various places before.
16167 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16168 UTF8 strings a character at a time.
16172 * Use client_version from client hello to select the protocol
16173 (s23_srvr.c) and for RSA client key exchange verification
16174 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16178 * Add various utility functions to handle SPKACs, these were previously
16179 handled by poking round in the structure internals. Added new function
16180 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16181 print, verify and generate SPKACs. Based on an original idea from
16182 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16186 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16190 * Allow the config file extension section to be overwritten on the
16191 command line. Based on an original idea from Massimiliano Pala
16192 <madwolf@comune.modena.it>. The new option is called -extensions
16193 and can be applied to ca, req and x509. Also -reqexts to override
16194 the request extensions in req and -crlexts to override the crl extensions
16199 * Add new feature to the SPKAC handling in ca. Now you can include
16200 the same field multiple times by preceding it by "XXXX." for example:
16203 this is the same syntax as used in the req config file.
16207 * Allow certificate extensions to be added to certificate requests. These
16208 are specified in a 'req_extensions' option of the req section of the
16209 config file. They can be printed out with the -text option to req but
16210 are otherwise ignored at present.
16214 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16215 data read consists of only the final block it would not decrypted because
16216 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16217 A misplaced 'break' also meant the decrypted final block might not be
16218 copied until the next read.
16222 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16223 a few extra parameters to the DH structure: these will be useful if
16224 for example we want the value of 'q' or implement X9.42 DH.
16228 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16229 provides hooks that allow the default DSA functions or functions on a
16230 "per key" basis to be replaced. This allows hardware acceleration and
16231 hardware key storage to be handled without major modification to the
16232 library. Also added low level modexp hooks and CRYPTO_EX structure and
16233 associated functions.
16237 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16238 as "read only": it can't be written to and the buffer it points to will
16239 not be freed. Reading from a read only BIO is much more efficient than
16240 a normal memory BIO. This was added because there are several times when
16241 an area of memory needs to be read from a BIO. The previous method was
16242 to create a memory BIO and write the data to it, this results in two
16243 copies of the data and an O(n^2) reading algorithm. There is a new
16244 function BIO_new_mem_buf() which creates a read only memory BIO from
16245 an area of memory. Also modified the PKCS#7 routines to use read only
16250 * Bugfix: ssl23_get_client_hello did not work properly when called in
16251 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16252 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16253 but a retry condition occurred while trying to read the rest.
16257 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16258 NID_pkcs7_encrypted by default: this was wrong since this should almost
16259 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16260 the encrypted data type: this is a more sensible place to put it and it
16261 allows the PKCS#12 code to be tidied up that duplicated this
16266 * Changed obj_dat.pl script so it takes its input and output files on
16267 the command line. This should avoid shell escape redirection problems
16272 * Initial support for certificate extension requests, these are included
16273 in things like Xenroll certificate requests. Included functions to allow
16274 extensions to be obtained and added.
16278 * -crlf option to s_client and s_server for sending newlines as
16279 CRLF (as required by many protocols).
16283 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999] ###
16285 * Install libRSAglue.a when OpenSSL is built with RSAref.
16287 *Ralf S. Engelschall*
16289 * A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
16291 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16293 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16298 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16299 DH parameters/keys (q is lost during that conversion, but the resulting
16300 DH parameters contain its length).
16302 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16303 much faster than DH_generate_parameters (which creates parameters
16304 where p = 2*q + 1), and also the smaller q makes DH computations
16305 much more efficient (160-bit exponentiation instead of 1024-bit
16306 exponentiation); so this provides a convenient way to support DHE
16307 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16308 utter importance to use
16309 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16311 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16312 when such DH parameters are used, because otherwise small subgroup
16313 attacks may become possible!
16317 * Avoid memory leak in i2d_DHparams.
16321 * Allow the -k option to be used more than once in the enc program:
16322 this allows the same encrypted message to be read by multiple recipients.
16326 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16327 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16328 it will always use the numerical form of the OID, even if it has a short
16333 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16334 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16335 otherwise bn_mod_exp was called. In the case of hardware keys for example
16336 no private key components need be present and it might store extra data
16337 in the RSA structure, which cannot be accessed from bn_mod_exp.
16338 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16339 private key operations.
16343 * Added support for SPARC Linux.
16347 * pem_password_cb function type incompatibly changed from
16348 typedef int pem_password_cb(char *buf, int size, int rwflag);
16350 ....(char *buf, int size, int rwflag, void *userdata);
16351 so that applications can pass data to their callbacks:
16352 The PEM[_ASN1]_{read,write}... functions and macros now take an
16353 additional void * argument, which is just handed through whenever
16354 the password callback is called.
16356 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16358 New function SSL_CTX_set_default_passwd_cb_userdata.
16360 Compatibility note: As many C implementations push function arguments
16361 onto the stack in reverse order, the new library version is likely to
16362 interoperate with programs that have been compiled with the old
16363 pem_password_cb definition (PEM_whatever takes some data that
16364 happens to be on the stack as its last argument, and the callback
16365 just ignores this garbage); but there is no guarantee whatsoever that
16368 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16369 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16370 problems not only on Windows, but also on some Unix platforms.
16371 To avoid problematic command lines, these definitions are now in an
16372 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16373 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16377 * MIPS III/IV assembler module is reimplemented.
16381 * More DES library cleanups: remove references to srand/rand and
16382 delete an unused file.
16386 * Add support for the free Netwide assembler (NASM) under Win32,
16387 since not many people have MASM (ml) and it can be hard to obtain.
16388 This is currently experimental but it seems to work OK and pass all
16389 the tests. Check out INSTALL.W32 for info.
16393 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16394 without temporary keys kept an extra copy of the server key,
16395 and connections with temporary keys did not free everything in case
16400 * New function RSA_check_key and new openssl rsa option -check
16401 for verifying the consistency of RSA keys.
16403 *Ulf Moeller, Bodo Moeller*
16405 * Various changes to make Win32 compile work:
16406 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16407 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16408 comparison" warnings.
16409 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
16413 * Add a debugging option to PKCS#5 v2 key generation function: when
16414 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16415 derived keys are printed to stderr.
16419 * Copy the flags in ASN1_STRING_dup().
16421 *Roman E. Pavlov <pre@mo.msk.ru>*
16423 * The x509 application mishandled signing requests containing DSA
16424 keys when the signing key was also DSA and the parameters didn't match.
16426 It was supposed to omit the parameters when they matched the signing key:
16427 the verifying software was then supposed to automatically use the CA's
16428 parameters if they were absent from the end user certificate.
16430 Omitting parameters is no longer recommended. The test was also
16431 the wrong way round! This was probably due to unusual behaviour in
16432 EVP_cmp_parameters() which returns 1 if the parameters match.
16433 This meant that parameters were omitted when they *didn't* match and
16434 the certificate was useless. Certificates signed with 'ca' didn't have
16437 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16439 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16440 The interface is as follows:
16441 Applications can use
16442 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16443 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16444 "off" is now the default.
16445 The library internally uses
16446 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16447 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16448 to disable memory-checking temporarily.
16450 Some inconsistent states that previously were possible (and were
16451 even the default) are now avoided.
16453 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16454 with each memory chunk allocated; this is occasionally more helpful
16455 than just having a counter.
16457 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16459 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16464 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16465 which largely parallels "options", but is for changing API behaviour,
16466 whereas "options" are about protocol behaviour.
16467 Initial "mode" flags are:
16469 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16470 a single record has been written.
16471 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16472 retries use the same buffer location.
16473 (But all of the contents must be
16478 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16481 * Fix problems with no-hmac etc.
16483 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16485 * New functions RSA_get_default_method(), RSA_set_method() and
16486 RSA_get_method(). These allows replacement of RSA_METHODs without having
16487 to mess around with the internals of an RSA structure.
16491 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16492 Also really enable memory leak checks in openssl.c and in some
16495 *Chad C. Mulligan, Bodo Moeller*
16497 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16498 up the length of negative integers. This has now been simplified to just
16499 store the length when it is first determined and use it later, rather
16500 than trying to keep track of where data is copied and updating it to
16502 *Steve Henson, reported by Brien Wheeler
16503 <bwheeler@authentica-security.com>*
16505 * Add a new function PKCS7_signatureVerify. This allows the verification
16506 of a PKCS#7 signature but with the signing certificate passed to the
16507 function itself. This contrasts with PKCS7_dataVerify which assumes the
16508 certificate is present in the PKCS#7 structure. This isn't always the
16509 case: certificates can be omitted from a PKCS#7 structure and be
16510 distributed by "out of band" means (such as a certificate database).
16514 * Complete the PEM_* macros with DECLARE_PEM versions to replace the
16515 function prototypes in pem.h, also change util/mkdef.pl to add the
16516 necessary function names.
16520 * mk1mf.pl (used by Windows builds) did not properly read the
16521 options set by Configure in the top level Makefile, and Configure
16522 was not even able to write more than one option correctly.
16523 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16527 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16528 file to be loaded from a BIO or FILE pointer. The BIO version will
16529 for example allow memory BIOs to contain config info.
16533 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16534 Whoever hopes to achieve shared-library compatibility across versions
16535 must use this, not the compile-time macro.
16536 (Exercise 0.9.4: Which is the minimum library version required by
16538 Note: All this applies only to multi-threaded programs, others don't
16543 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16544 through a BIO pair triggered the default case, i.e.
16545 SSLerr(...,SSL_R_UNKNOWN_STATE).
16549 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16550 can use the SSL library even if none of the specific BIOs is
16555 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16556 for the encoded length.
16558 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16560 * Add initial documentation of the X509V3 functions.
16564 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16565 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16566 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16567 secure PKCS#8 private key format with a high iteration count.
16571 * Fix determination of Perl interpreter: A perl or perl5
16572 _directory_ in $PATH was also accepted as the interpreter.
16574 *Ralf S. Engelschall*
16576 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16577 wrong with it but it was very old and did things like calling
16578 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16579 unusual formatting.
16583 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16584 to use the new extension code.
16588 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16589 with macros. This should make it easier to change their form, add extra
16590 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16595 * Add to configuration table a new entry that can specify an alternative
16596 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16597 according to Mark Crispin <MRC@Panda.COM>.
16602 * DES CBC did not update the IV. Weird.
16606 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16607 Changing the behaviour of the former might break existing programs --
16608 where IV updating is needed, des_ncbc_encrypt can be used.
16611 * When bntest is run from "make test" it drives bc to check its
16612 calculations, as well as internally checking them. If an internal check
16613 fails, it needs to cause bc to give a non-zero result or make test carries
16614 on without noticing the failure. Fixed.
16618 * DES library cleanups.
16622 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16623 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16624 ciphers. NOTE: although the key derivation function has been verified
16625 against some published test vectors it has not been extensively tested
16626 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16631 * Instead of "mkdir -p", which is not fully portable, use new
16632 Perl script "util/mkdir-p.pl".
16636 * Rewrite the way password based encryption (PBE) is handled. It used to
16637 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16638 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16639 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16640 the 'parameter' field of the AlgorithmIdentifier is passed to the
16641 underlying key generation function so it must do its own ASN1 parsing.
16642 This has also changed the EVP_PBE_CipherInit() function which now has a
16643 'parameter' argument instead of literal salt and iteration count values
16644 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16648 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16649 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16650 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16651 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16652 value was just used as a "magic string" and not used directly its
16653 value doesn't matter.
16657 * Introduce some semblance of const correctness to BN. Shame C doesn't
16662 * "linux-sparc64" configuration (ultrapenguin).
16664 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16665 "linux-sparc" configuration.
16667 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16669 * config now generates no-xxx options for missing ciphers.
16673 * Support the EBCDIC character set (work in progress).
16674 File ebcdic.c not yet included because it has a different license.
16676 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16678 * Support BS2000/OSD-POSIX.
16680 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16682 * Make callbacks for key generation use void * instead of char *.
16686 * Make S/MIME samples compile (not yet tested).
16690 * Additional typesafe stacks.
16694 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16699 ### Changes between 0.9.3 and 0.9.3a [29 May 1999] ###
16701 * New configuration variant "sco5-gcc".
16703 * Updated some demos.
16705 *Sean O Riordain, Wade Scholine*
16707 * Add missing BIO_free at exit of pkcs12 application.
16711 * Fix memory leak in conf.c.
16715 * Updates for Win32 to assembler version of MD5.
16719 * Set #! path to perl in apps/der_chop to where we found it
16720 instead of using a fixed path.
16724 * SHA library changes for irix64-mips4-cc.
16728 * Improvements for VMS support.
16733 ### Changes between 0.9.2b and 0.9.3 [24 May 1999] ###
16735 * Bignum library bug fix. IRIX 6 passes "make test" now!
16736 This also avoids the problems with SC4.2 and unpatched SC5.
16738 *Andy Polyakov <appro@fy.chalmers.se>*
16740 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16741 These are required because of the typesafe stack would otherwise break
16742 existing code. If old code used a structure member which used to be STACK
16743 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16744 sk_num or sk_value it would produce an error because the num, data members
16745 are not present in STACK_OF. Now it just produces a warning. sk_set
16746 replaces the old method of assigning a value to sk_value
16747 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16748 that does this will no longer work (and should use sk_set instead) but
16749 this could be regarded as a "questionable" behaviour anyway.
16753 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16754 correctly handle encrypted S/MIME data.
16758 * Change type of various DES function arguments from des_cblock
16759 (which means, in function argument declarations, pointer to char)
16760 to des_cblock * (meaning pointer to array with 8 char elements),
16761 which allows the compiler to do more typechecking; it was like
16762 that back in SSLeay, but with lots of ugly casts.
16764 Introduce new type const_des_cblock.
16768 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16769 problems: find RecipientInfo structure that matches recipient certificate
16770 and initialise the ASN1 structures properly based on passed cipher.
16774 * Belatedly make the BN tests actually check the results.
16778 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16779 to and from BNs: it was completely broken. New compilation option
16780 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16781 key elements as negative integers.
16785 * Reorganize and speed up MD5.
16787 *Andy Polyakov <appro@fy.chalmers.se>*
16791 *Richard Levitte <richard@levitte.org>*
16793 * New option -out to asn1parse to allow the parsed structure to be
16794 output to a file. This is most useful when combined with the -strparse
16795 option to examine the output of things like OCTET STRINGS.
16799 * Make SSL library a little more fool-proof by not requiring any longer
16800 that SSL_set_{accept,connect}_state be called before
16801 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
16802 in many applications because usually everything *appeared* to work as
16803 intended anyway -- now it really works as intended).
16807 * Move openssl.cnf out of lib/.
16811 * Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
16812 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16813 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
16815 *Ralf S. Engelschall*
16817 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16818 handle PKCS#7 enveloped data properly.
16820 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16822 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16823 copying pointers. The cert_st handling is changed by this in
16824 various ways (and thus what used to be known as ctx->default_cert
16825 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
16826 any longer when s->cert does not give us what we need).
16827 ssl_cert_instantiate becomes obsolete by this change.
16828 As soon as we've got the new code right (possibly it already is?),
16829 we have solved a couple of bugs of the earlier code where s->cert
16830 was used as if it could not have been shared with other SSL structures.
16832 Note that using the SSL API in certain dirty ways now will result
16833 in different behaviour than observed with earlier library versions:
16834 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
16835 does not influence s as it used to.
16837 In order to clean up things more thoroughly, inside SSL_SESSION
16838 we don't use CERT any longer, but a new structure SESS_CERT
16839 that holds per-session data (if available); currently, this is
16840 the peer's certificate chain and, for clients, the server's certificate
16841 and temporary key. CERT holds only those values that can have
16842 meaningful defaults in an SSL_CTX.
16846 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16847 from the internal representation. Various PKCS#7 fixes: remove some
16848 evil casts and set the enc_dig_alg field properly based on the signing
16853 * Allow PKCS#12 password to be set from the command line or the
16854 environment. Let 'ca' get its config file name from the environment
16855 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16860 * Allow certificate policies extension to use an IA5STRING for the
16861 organization field. This is contrary to the PKIX definition but
16862 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16867 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16868 without disallowing inline assembler and the like for non-pedantic builds.
16872 * Support Borland C++ builder.
16874 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16880 * SHA-1 cleanups and performance enhancements.
16882 *Andy Polyakov <appro@fy.chalmers.se>*
16884 * Sparc v8plus assembler for the bignum library.
16886 *Andy Polyakov <appro@fy.chalmers.se>*
16888 * Accept any -xxx and +xxx compiler options in Configure.
16892 * Update HPUX configuration.
16896 * Add missing sk_<type>_unshift() function to safestack.h
16898 *Ralf S. Engelschall*
16900 * New function SSL_CTX_use_certificate_chain_file that sets the
16901 "extra_cert"s in addition to the certificate. (This makes sense
16902 only for "PEM" format files, as chains as a whole are not
16907 * Support verify_depth from the SSL API.
16908 x509_vfy.c had what can be considered an off-by-one-error:
16909 Its depth (which was not part of the external interface)
16910 was actually counting the number of certificates in a chain;
16911 now it really counts the depth.
16915 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16916 instead of X509err, which often resulted in confusing error
16917 messages since the error codes are not globally unique
16918 (e.g. an alleged error in ssl3_accept when a certificate
16919 didn't match the private key).
16921 * New function SSL_CTX_set_session_id_context that allows to set a default
16922 value (so that you don't need SSL_set_session_id_context for each
16923 connection using the SSL_CTX).
16927 * OAEP decoding bug fix.
16931 * Support INSTALL_PREFIX for package builders, as proposed by
16936 * New Configure options "threads" and "no-threads". For systems
16937 where the proper compiler options are known (currently Solaris
16938 and Linux), "threads" is the default.
16942 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
16946 * Install various scripts to $(OPENSSLDIR)/misc, not to
16947 $(INSTALLTOP)/bin -- they shouldn't clutter directories
16948 such as /usr/local/bin.
16952 * "make linux-shared" to build shared libraries.
16954 *Niels Poppe <niels@netbox.org>*
16956 * New Configure option no-<cipher> (rsa, idea, rc5, ...).
16960 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
16961 extension adding in x509 utility.
16965 * Remove NOPROTO sections and error code comments.
16969 * Partial rewrite of the DEF file generator to now parse the ANSI
16974 * New Configure options --prefix=DIR and --openssldir=DIR.
16978 * Complete rewrite of the error code script(s). It is all now handled
16979 by one script at the top level which handles error code gathering,
16980 header rewriting and C source file generation. It should be much better
16981 than the old method: it now uses a modified version of Ulf's parser to
16982 read the ANSI prototypes in all header files (thus the old K&R definitions
16983 aren't needed for error creation any more) and do a better job of
16984 translating function codes into names. The old 'ASN1 error code embedded
16985 in a comment' is no longer necessary and it doesn't use .err files which
16986 have now been deleted. Also the error code call doesn't have to appear all
16987 on one line (which resulted in some large lines...).
16991 * Change #include filenames from <foo.h> to <openssl/foo.h>.
16995 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
16996 0 (which usually indicates a closed connection), but continue reading.
17000 * Fix some race conditions.
17004 * Add support for CRL distribution points extension. Add Certificate
17005 Policies and CRL distribution points documentation.
17009 * Move the autogenerated header file parts to crypto/opensslconf.h.
17013 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17014 8 of keying material. Merlin has also confirmed interop with this fix
17015 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17017 *Merlin Hughes <merlin@baltimore.ie>*
17019 * Fix lots of warnings.
17021 *Richard Levitte <levitte@stacken.kth.se>*
17023 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17024 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17026 *Richard Levitte <levitte@stacken.kth.se>*
17028 * Fix problems with sizeof(long) == 8.
17030 *Andy Polyakov <appro@fy.chalmers.se>*
17032 * Change functions to ANSI C.
17036 * Fix typos in error codes.
17038 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17040 * Remove defunct assembler files from Configure.
17044 * SPARC v8 assembler BIGNUM implementation.
17046 *Andy Polyakov <appro@fy.chalmers.se>*
17048 * Support for Certificate Policies extension: both print and set.
17049 Various additions to support the r2i method this uses.
17053 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17054 return a const string when you are expecting an allocated buffer.
17058 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17059 types DirectoryString and DisplayText.
17063 * Add code to allow r2i extensions to access the configuration database,
17064 add an LHASH database driver and add several ctx helper functions.
17068 * Fix an evil bug in bn_expand2() which caused various BN functions to
17069 fail when they extended the size of a BIGNUM.
17073 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17074 support typesafe stack.
17078 * Fix typo in SSL_[gs]et_options().
17080 *Nils Frostberg <nils@medcom.se>*
17082 * Delete various functions and files that belonged to the (now obsolete)
17083 old X509V3 handling code.
17087 * New Configure option "rsaref".
17091 * Don't auto-generate pem.h.
17095 * Introduce type-safe ASN.1 SETs.
17099 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17101 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17103 * Introduce type-safe STACKs. This will almost certainly break lots of code
17104 that links with OpenSSL (well at least cause lots of warnings), but fear
17105 not: the conversion is trivial, and it eliminates loads of evil casts. A
17106 few STACKed things have been converted already. Feel free to convert more.
17107 In the fullness of time, I'll do away with the STACK type altogether.
17111 * Add `openssl ca -revoke <certfile>' facility which revokes a certificate
17112 specified in <certfile> by updating the entry in the index.txt file.
17113 This way one no longer has to edit the index.txt file manually for
17114 revoking a certificate. The -revoke option does the gory details now.
17116 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17118 * Fix `openssl crl -noout -text' combination where `-noout' killed the
17119 `-text' option at all and this way the `-noout -text' combination was
17120 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
17122 *Ralf S. Engelschall*
17124 * Make sure a corresponding plain text error message exists for the
17125 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17126 verify callback function determined that a certificate was revoked.
17128 *Ralf S. Engelschall*
17130 * Bugfix: In test/testenc, don't test "openssl <cipher>" for
17131 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17132 all available ciphers including rc5, which was forgotten until now.
17133 In order to let the testing shell script know which algorithms
17134 are available, a new (up to now undocumented) command
17135 "openssl list-cipher-commands" is used.
17139 * Bugfix: s_client occasionally would sleep in select() when
17140 it should have checked SSL_pending() first.
17144 * New functions DSA_do_sign and DSA_do_verify to provide access to
17145 the raw DSA values prior to ASN.1 encoding.
17149 * Tweaks to Configure
17151 *Niels Poppe <niels@netbox.org>*
17153 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17158 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17162 * New config option to avoid instructions that are illegal on the 80386.
17163 The default code is faster, but requires at least a 486.
17167 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17168 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17169 same as SSL2_VERSION anyway.
17173 * New "-showcerts" option for s_client.
17177 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17178 application. Various cleanups and fixes.
17182 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17183 modify error routines to work internally. Add error codes and PBE init
17184 to library startup routines.
17188 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17189 packing functions to asn1 and evp. Changed function names and error
17190 codes along the way.
17194 * PKCS12 integration: and so it begins... First of several patches to
17195 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17196 objects to objects.h
17200 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17201 and display support for Thawte strong extranet extension.
17205 * Add LinuxPPC support.
17207 *Jeff Dubrule <igor@pobox.org>*
17209 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17210 bn_div_words in alpha.s.
17212 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17214 * Make sure the RSA OAEP test is skipped under -DRSAref because
17215 OAEP isn't supported when OpenSSL is built with RSAref.
17217 *Ulf Moeller <ulf@fitug.de>*
17219 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17220 so they no longer are missing under -DNOPROTO.
17222 *Soren S. Jorvang <soren@t.dk>*
17225 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999] ###
17227 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17228 doesn't work when the session is reused. Coming soon!
17232 * Fix a security hole, that allows sessions to be reused in the wrong
17233 context thus bypassing client cert protection! All software that uses
17234 client certs and session caches in multiple contexts NEEDS PATCHING to
17235 allow session reuse! A fuller solution is in the works.
17237 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17239 * Some more source tree cleanups (removed obsolete files
17240 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17241 permission on "config" script to be executable) and a fix for the INSTALL
17244 *Ulf Moeller <ulf@fitug.de>*
17246 * Remove some legacy and erroneous uses of malloc, free instead of
17249 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17251 * Make rsa_oaep_test return non-zero on error.
17253 *Ulf Moeller <ulf@fitug.de>*
17255 * Add support for native Solaris shared libraries. Configure
17256 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17257 if someone would make that last step automatic.
17259 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17261 * ctx_size was not built with the right compiler during "make links". Fixed.
17265 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17266 except NULL ciphers". This means the default cipher list will no longer
17267 enable NULL ciphers. They need to be specifically enabled e.g. with
17268 the string "DEFAULT:eNULL".
17272 * Fix to RSA private encryption routines: if p < q then it would
17273 occasionally produce an invalid result. This will only happen with
17274 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17278 * Be less restrictive and allow also `perl util/perlpath.pl
17279 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17280 because this way one can also use an interpreter named `perl5' (which is
17281 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17282 installed as `perl').
17284 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17286 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17288 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17290 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17291 advapi32.lib to Win32 build and change the pem test comparison
17292 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17293 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17294 and crypto/des/ede_cbcm_enc.c.
17298 * DES quad checksum was broken on big-endian architectures. Fixed.
17302 * Comment out two functions in bio.h that aren't implemented. Fix up the
17303 Win32 test batch file so it (might) work again. The Win32 test batch file
17304 is horrible: I feel ill....
17308 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17309 in e_os.h. Audit of header files to check ANSI and non ANSI
17310 sections: 10 functions were absent from non ANSI section and not exported
17311 from Windows DLLs. Fixed up libeay.num for new functions.
17315 * Make `openssl version' output lines consistent.
17317 *Ralf S. Engelschall*
17319 * Fix Win32 symbol export lists for BIO functions: Added
17320 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17321 to ms/libeay{16,32}.def.
17323 *Ralf S. Engelschall*
17325 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17326 fine under Unix and passes some trivial tests I've now added. But the
17327 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17328 added to make sure no one expects that this stuff really works in the
17329 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17330 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17333 *Ralf S. Engelschall*
17335 * Fix the generation of two part addresses in perl.
17337 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17339 * Add config entry for Linux on MIPS.
17341 *John Tobey <jtobey@channel1.com>*
17343 * Make links whenever Configure is run, unless we are on Windoze.
17347 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17348 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17353 * Add a useful kludge to allow package maintainers to specify compiler and
17354 other platforms details on the command line without having to patch the
17355 Configure script every time: One now can use ``perl Configure
17356 <id>:<details>'', i.e. platform ids are allowed to have details appended
17357 to them (separated by colons). This is treated as there would be a static
17358 pre-configured entry in Configure's %table under key <id> with value
17359 <details> and ``perl Configure <id>'' is called. So, when you want to
17360 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17361 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
17362 now, which overrides the FreeBSD-elf entry on-the-fly.
17364 *Ralf S. Engelschall*
17366 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17370 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17371 on the `perl Configure ...' command line. This way one can compile
17372 OpenSSL libraries with Position Independent Code (PIC) which is needed
17373 for linking it into DSOs.
17375 *Ralf S. Engelschall*
17377 * Remarkably, export ciphers were totally broken and no-one had noticed!
17382 * Cleaned up the LICENSE document: The official contact for any license
17383 questions now is the OpenSSL core team under openssl-core@openssl.org.
17384 And add a paragraph about the dual-license situation to make sure people
17385 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17386 to the OpenSSL toolkit.
17388 *Ralf S. Engelschall*
17390 * General source tree makefile cleanups: Made `making xxx in yyy...'
17391 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17392 Additionally cleaned up the `make links' target: Remove unnecessary
17393 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17394 to speed processing and no longer clutter the display with confusing
17395 stuff. Instead only the actually done links are displayed.
17397 *Ralf S. Engelschall*
17399 * Permit null encryption ciphersuites, used for authentication only. It used
17400 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17401 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17406 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17407 signed attributes when verifying signatures (this would break them),
17408 the detached data encoding was wrong and public keys obtained using
17409 X509_get_pubkey() weren't freed.
17413 * Add text documentation for the BUFFER functions. Also added a work around
17414 to a Win95 console bug. This was triggered by the password read stuff: the
17415 last character typed gets carried over to the next fread(). If you were
17416 generating a new cert request using 'req' for example then the last
17417 character of the passphrase would be CR which would then enter the first
17422 * Added the new `Includes OpenSSL Cryptography Software' button as
17423 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17424 button and can be used by applications based on OpenSSL to show the
17425 relationship to the OpenSSL project.
17427 *Ralf S. Engelschall*
17429 * Remove confusing variables in function signatures in files
17430 ssl/ssl_lib.c and ssl/ssl.h.
17432 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17434 * Don't install bss_file.c under PREFIX/include/
17436 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17438 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17439 functions that return function pointers and has support for NT specific
17440 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17441 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17442 unsigned to signed types: this was killing the Win32 compile.
17446 * Add new certificate file to stack functions,
17447 SSL_add_dir_cert_subjects_to_stack() and
17448 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17449 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17450 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17451 This means that Apache-SSL and similar packages don't have to mess around
17452 to add as many CAs as they want to the preferred list.
17456 * Experiment with doxygen documentation. Currently only partially applied to
17458 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
17459 openssl.doxy as the configuration file.
17463 * Get rid of remaining C++-style comments which strict C compilers hate.
17465 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17467 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17468 compiled in by default: it has problems with large keys.
17472 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17473 DH private keys and/or callback functions which directly correspond to
17474 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17475 is needed for applications which have to configure certificates on a
17476 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17478 For the RSA certificate situation is makes no difference, but
17479 for the DSA certificate situation this fixes the "no shared cipher"
17480 problem where the OpenSSL cipher selection procedure failed because the
17481 temporary keys were not overtaken from the context and the API provided
17482 no way to reconfigure them.
17483 The new functions now let applications reconfigure the stuff and they
17484 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17485 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17486 non-public-API function ssl_cert_instantiate() is used as a helper
17487 function and also to reduce code redundancy inside ssl_rsa.c.
17489 *Ralf S. Engelschall*
17491 * Move s_server -dcert and -dkey options out of the undocumented feature
17492 area because they are useful for the DSA situation and should be
17493 recognized by the users.
17495 *Ralf S. Engelschall*
17497 * Fix the cipher decision scheme for export ciphers: the export bits are
17498 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17499 SSL_EXP_MASK. So, the original variable has to be used instead of the
17500 already masked variable.
17502 *Richard Levitte <levitte@stacken.kth.se>*
17504 * Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
17506 *Richard Levitte <levitte@stacken.kth.se>*
17508 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17509 from `int' to `unsigned int' because it's a length and initialized by
17510 EVP_DigestFinal() which expects an `unsigned int *'.
17512 *Richard Levitte <levitte@stacken.kth.se>*
17514 * Don't hard-code path to Perl interpreter on shebang line of Configure
17515 script. Instead use the usual Shell->Perl transition trick.
17517 *Ralf S. Engelschall*
17519 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17520 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17521 -noout -modulus' as it's already the case for `openssl rsa -noout
17522 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17523 currently the public key is printed (a decision which was already done by
17524 `openssl dsa -modulus' in the past) which serves a similar purpose.
17525 Additionally the NO_RSA no longer completely removes the whole -modulus
17526 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17529 *Ralf S. Engelschall*
17531 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17532 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17534 *Arne Ansper <arne@ats.cyber.ee>*
17536 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17537 to be added. Now both 'req' and 'ca' can use new objects defined in the
17542 * Add cool BIO that does syslog (or event log on NT).
17544 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17546 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17547 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17548 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17549 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17553 * Add preliminary config info for new extension code.
17557 * Make RSA_NO_PADDING really use no padding.
17559 *Ulf Moeller <ulf@fitug.de>*
17561 * Generate errors when private/public key check is done.
17565 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17566 for some CRL extensions and new objects added.
17570 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17571 key usage extension and fuller support for authority key id.
17575 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17576 padding method for RSA, which is recommended for new applications in PKCS
17577 #1 v2.0 (RFC 2437, October 1998).
17578 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17579 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17580 against Bleichbacher's attack on RSA.
17581 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17584 * Updates to the new SSL compression code
17586 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17588 * Fix so that the version number in the master secret, when passed
17589 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17590 (because the server will not accept higher), that the version number
17591 is 0x03,0x01, not 0x03,0x00
17593 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17595 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17596 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17597 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17601 * Support for RAW extensions where an arbitrary extension can be
17602 created by including its DER encoding. See apps/openssl.cnf for
17607 * Make sure latest Perl versions don't interpret some generated C array
17608 code as Perl array code in the crypto/err/err_genc.pl script.
17610 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17612 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17613 not many people have the assembler. Various Win32 compilation fixes and
17614 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17615 build instructions.
17619 * Modify configure script 'Configure' to automatically create crypto/date.h
17620 file under Win32 and also build pem.h from pem.org. New script
17621 util/mkfiles.pl to create the MINFO file on environments that can't do a
17622 'make files': perl util/mkfiles.pl >MINFO should work.
17626 * Major rework of DES function declarations, in the pursuit of correctness
17627 and purity. As a result, many evil casts evaporated, and some weirdness,
17628 too. You may find this causes warnings in your code. Zapping your evil
17629 casts will probably fix them. Mostly.
17633 * Fix for a typo in asn1.h. Bug fix to object creation script
17634 obj_dat.pl. It considered a zero in an object definition to mean
17635 "end of object": none of the objects in objects.h have any zeros
17636 so it wasn't spotted.
17638 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17640 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17641 Masking (CBCM). In the absence of test vectors, the best I have been able
17642 to do is check that the decrypt undoes the encrypt, so far. Send me test
17643 vectors if you have them.
17647 * Correct calculation of key length for export ciphers (too much space was
17648 allocated for null ciphers). This has not been tested!
17652 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17653 message is now correct (it understands "crypto" and "ssl" on its
17654 command line). There is also now an "update" option. This will update
17655 the util/ssleay.num and util/libeay.num files with any new functions.
17657 perl util/mkdef.pl crypto ssl update
17658 it will update them.
17662 * Overhauled the Perl interface (perl/*):
17663 - ported BN stuff to OpenSSL's different BN library
17664 - made the perl/ source tree CVS-aware
17665 - renamed the package from SSLeay to OpenSSL (the files still contain
17666 their history because I've copied them in the repository)
17667 - removed obsolete files (the test scripts will be replaced
17668 by better Test::Harness variants in the future)
17670 *Ralf S. Engelschall*
17672 * First cut for a very conservative source tree cleanup:
17673 1. merge various obsolete readme texts into doc/ssleay.txt
17674 where we collect the old documents and readme texts.
17675 2. remove the first part of files where I'm already sure that we no
17676 longer need them because of three reasons: either they are just temporary
17677 files which were left by Eric or they are preserved original files where
17678 I've verified that the diff is also available in the CVS via "cvs diff
17679 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17680 the crypto/md/ stuff).
17682 *Ralf S. Engelschall*
17684 * More extension code. Incomplete support for subject and issuer alt
17685 name, issuer and authority key id. Change the i2v function parameters
17686 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17687 what that's for :-) Fix to ASN1 macro which messed up
17688 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17692 * Preliminary support for ENUMERATED type. This is largely copied from the
17697 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17699 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17701 * Make sure `make rehash' target really finds the `openssl' program.
17703 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17705 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17706 like to hear about it if this slows down other processors.
17710 * Add CygWin32 platform information to Configure script.
17712 *Alan Batie <batie@aahz.jf.intel.com>*
17714 * Fixed ms/32all.bat script: `no_asm' -> `no-asm'
17716 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17718 * New program nseq to manipulate netscape certificate sequences
17722 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17727 * Fixes to BN code. Previously the default was to define BN_RECURSION
17728 but the BN code had some problems that would cause failures when
17729 doing certificate verification and some other functions.
17731 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17733 * Add ASN1 and PEM code to support netscape certificate sequences.
17737 * Add ASN1 and PEM code to support netscape certificate sequences.
17741 * Add several PKIX and private extended key usage OIDs.
17745 * Modify the 'ca' program to handle the new extension code. Modify
17746 openssl.cnf for new extension format, add comments.
17750 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17751 and add a sample to openssl.cnf so req -x509 now adds appropriate
17756 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17757 error code, add initial support to X509_print() and x509 application.
17761 * Takes a deep breath and start adding X509 V3 extension support code. Add
17762 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17763 stuff is currently isolated and isn't even compiled yet.
17767 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17768 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17769 Removed the versions check from X509 routines when loading extensions:
17770 this allows certain broken certificates that don't set the version
17771 properly to be processed.
17775 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17776 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17777 can still be regenerated with "make depend".
17781 * Spelling mistake in C version of CAST-128.
17783 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17785 * Changes to the error generation code. The perl script err-code.pl
17786 now reads in the old error codes and retains the old numbers, only
17787 adding new ones if necessary. It also only changes the .err files if new
17788 codes are added. The makefiles have been modified to only insert errors
17789 when needed (to avoid needlessly modifying header files). This is done
17790 by only inserting errors if the .err file is newer than the auto generated
17791 C file. To rebuild all the error codes from scratch (the old behaviour)
17792 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17793 or delete all the .err files.
17797 * CAST-128 was incorrectly implemented for short keys. The C version has
17798 been fixed, but is untested. The assembler versions are also fixed, but
17799 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17800 to regenerate it if needed.
17801 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17802 Hagino <itojun@kame.net>*
17804 * File was opened incorrectly in randfile.c.
17806 *Ulf Möller <ulf@fitug.de>*
17808 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17809 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17810 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17811 al: it's just almost always a UTCTime. Note this patch adds new error
17812 codes so do a "make errors" if there are problems.
17816 * Correct Linux 1 recognition in config.
17818 *Ulf Möller <ulf@fitug.de>*
17820 * Remove pointless MD5 hash when using DSA keys in ca.
17822 *Anonymous <nobody@replay.com>*
17824 * Generate an error if given an empty string as a cert directory. Also
17825 generate an error if handed NULL (previously returned 0 to indicate an
17826 error, but didn't set one).
17828 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17830 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17834 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17835 parameters. This was causing a warning which killed off the Win32 compile.
17839 * Remove C++ style comments from crypto/bn/bn_local.h.
17841 *Neil Costigan <neil.costigan@celocom.com>*
17843 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17844 based on a text string, looking up short and long names and finally
17845 "dot" format. The "dot" format stuff didn't work. Added new function
17846 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17847 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17848 OID is not part of the table.
17852 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17853 X509_LOOKUP_by_alias().
17857 * Sort openssl functions by name.
17861 * Get the gendsa program working (hopefully) and add it to app list. Remove
17862 encryption from sample DSA keys (in case anyone is interested the password
17867 * Make _all_ *_free functions accept a NULL pointer.
17869 *Frans Heymans <fheymans@isaserver.be>*
17871 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17874 *Anonymous <nobody@replay.com>*
17876 * s_server should send the CAfile as acceptable CAs, not its own cert.
17878 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17880 * Don't blow it for numeric -newkey arguments to apps/req.
17882 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17884 * Temp key "for export" tests were wrong in s3_srvr.c.
17886 *Anonymous <nobody@replay.com>*
17888 * Add prototype for temp key callback functions
17889 SSL_CTX_set_tmp_{rsa,dh}_callback().
17893 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17894 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17898 * X509_name_add_entry() freed the wrong thing after an error.
17900 *Arne Ansper <arne@ats.cyber.ee>*
17902 * rsa_eay.c would attempt to free a NULL context.
17904 *Arne Ansper <arne@ats.cyber.ee>*
17906 * BIO_s_socket() had a broken should_retry() on Windoze.
17908 *Arne Ansper <arne@ats.cyber.ee>*
17910 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17912 *Arne Ansper <arne@ats.cyber.ee>*
17914 * Make sure the already existing X509_STORE->depth variable is initialized
17915 in X509_STORE_new(), but document the fact that this variable is still
17916 unused in the certificate verification process.
17918 *Ralf S. Engelschall*
17920 * Fix the various library and apps files to free up pkeys obtained from
17921 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17925 * Fix reference counting in X509_PUBKEY_get(). This makes
17926 demos/maurice/example2.c work, amongst others, probably.
17928 *Steve Henson and Ben Laurie*
17930 * First cut of a cleanup for apps/. First the `ssleay' program is now named
17931 `openssl' and second, the shortcut symlinks for the `openssl <command>'
17932 are no longer created. This way we have a single and consistent command
17933 line interface `openssl <command>', similar to `cvs <command>'.
17935 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
17937 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
17938 BIT STRING wrapper always have zero unused bits.
17942 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
17946 * Make the top-level INSTALL documentation easier to understand.
17950 * Makefiles updated to exit if an error occurs in a sub-directory
17951 make (including if user presses ^C) [Paul Sutton]
17953 * Make Montgomery context stuff explicit in RSA data structure.
17957 * Fix build order of pem and err to allow for generated pem.h.
17961 * Fix renumbering bug in X509_NAME_delete_entry().
17965 * Enhanced the err-ins.pl script so it makes the error library number
17966 global and can add a library name. This is needed for external ASN1 and
17967 other error libraries.
17971 * Fixed sk_insert which never worked properly.
17975 * Fix ASN1 macros so they can handle indefinite length constructed
17976 EXPLICIT tags. Some non standard certificates use these: they can now
17981 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
17982 into a single doc/ssleay.txt bundle. This way the information is still
17983 preserved but no longer messes up this directory. Now it's new room for
17984 the new set of documentation files.
17986 *Ralf S. Engelschall*
17988 * SETs were incorrectly DER encoded. This was a major pain, because they
17989 shared code with SEQUENCEs, which aren't coded the same. This means that
17990 almost everything to do with SETs or SEQUENCEs has either changed name or
17991 number of arguments.
17993 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
17995 * Fix test data to work with the above.
17999 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18000 was already fixed by Eric for 0.9.1 it seems.
18002 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18004 * Autodetect FreeBSD3.
18008 * Fix various bugs in Configure. This affects the following platforms:
18012 unixware-2.0-pentium
18017 * Eliminate generated files from CVS. Reorder tests to regenerate files
18018 before they are needed.
18022 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18027 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998] ###
18029 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18030 changed SSLeay to OpenSSL in version strings.
18032 *Ralf S. Engelschall*
18034 * Some fixups to the top-level documents.
18038 * Fixed the nasty bug where rsaref.h was not found under compile-time
18039 because the symlink to include/ was missing.
18041 *Ralf S. Engelschall*
18043 * Incorporated the popular no-RSA/DSA-only patches
18044 which allow to compile a RSA-free SSLeay.
18046 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18048 * Fixed nasty rehash problem under `make -f Makefile.ssl links'
18049 when "ssleay" is still not found.
18051 *Ralf S. Engelschall*
18053 * Added more platforms to Configure: Cray T3E, HPUX 11,
18055 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18057 * Updated the README file.
18059 *Ralf S. Engelschall*
18061 * Added various .cvsignore files in the CVS repository subdirs
18062 to make a "cvs update" really silent.
18064 *Ralf S. Engelschall*
18066 * Recompiled the error-definition header files and added
18067 missing symbols to the Win32 linker tables.
18069 *Ralf S. Engelschall*
18071 * Cleaned up the top-level documents;
18072 o new files: CHANGES and LICENSE
18073 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18074 o merged COPYRIGHT into LICENSE
18075 o removed obsolete TODO file
18076 o renamed MICROSOFT to INSTALL.W32
18078 *Ralf S. Engelschall*
18080 * Removed dummy files from the 0.9.1b source tree:
18081 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18082 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18083 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18084 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18085 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18087 *Ralf S. Engelschall*
18089 * Added various platform portability fixes.
18093 * The Genesis of the OpenSSL rpject:
18094 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18095 Young and Tim J. Hudson created while they were working for C2Net until
18098 *The OpenSSL Project*
18101 ### Changes between 0.9.0b and 0.9.1b [not released] ###
18103 * Updated a few CA certificates under certs/
18107 * Changed some BIGNUM api stuff.
18111 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18112 DGUX x86, Linux Alpha, etc.
18116 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18117 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18122 * Add -strparse option to asn1pars program which parses nested
18125 *Dr Stephen Henson <shenson@bigfoot.com>*
18127 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18131 * DSA fix for "ca" program.
18135 * Added "-genkey" option to "dsaparam" program.
18139 * Added RIPE MD160 (rmd160) message digest.
18143 * Added -a (all) option to "ssleay version" command.
18147 * Added PLATFORM define which is the id given to Configure.
18151 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18155 * Extended the ASN.1 parser routines.
18159 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18163 * Added a BN_CTX to the BN library.
18167 * Fixed the weak key values in DES library
18171 * Changed API in EVP library for cipher aliases.
18175 * Added support for RC2/64bit cipher.
18179 * Converted the lhash library to the crypto/mem.c functions.
18183 * Added more recognized ASN.1 object ids.
18187 * Added more RSA padding checks for SSL/TLS.
18191 * Added BIO proxy/filter functionality.
18195 * Added extra_certs to SSL_CTX which can be used
18196 send extra CA certificates to the client in the CA cert chain sending
18197 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18201 * Now Fortezza is denied in the authentication phase because
18202 this is key exchange mechanism is not supported by SSLeay at all.
18206 * Additional PKCS1 checks.
18210 * Support the string "TLSv1" for all TLS v1 ciphers.
18214 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18215 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18219 * Fixed a few memory leaks.
18223 * Fixed various code and comment typos.
18227 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18228 bytes sent in the client random.
18230 *Edward Bishop <ebishop@spyglass.com>*
18235 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18236 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18237 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18238 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18239 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18240 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18241 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18242 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18243 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18244 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18245 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18246 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18247 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18248 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18249 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18250 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18251 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18252 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18253 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18254 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18255 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18256 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18257 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18258 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18259 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18260 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18261 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18262 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18263 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18264 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18265 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18266 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18267 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18268 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18269 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18270 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18271 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18272 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18273 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18274 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18275 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18276 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18277 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18278 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18279 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18280 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18281 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18282 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18283 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18284 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18285 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18286 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18287 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18288 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18289 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18290 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18291 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18292 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18293 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18294 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18295 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18296 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18297 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18298 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18299 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18300 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18301 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18302 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18303 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18304 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18305 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18306 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18307 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18308 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18309 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18310 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18311 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18312 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18313 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18314 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18315 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18316 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18317 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18318 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18319 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18320 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18321 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18322 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18323 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18324 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18325 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18326 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18327 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18328 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18329 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18330 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18331 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18332 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18333 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18334 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18335 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18336 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18337 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18338 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18339 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18340 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18341 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18342 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18343 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18344 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18345 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18346 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18347 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18348 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18349 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18350 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18351 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18352 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18353 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18354 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18355 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18356 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18357 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18358 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18359 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18360 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18361 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18362 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18363 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18364 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18365 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18366 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18367 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18368 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18369 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18370 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18371 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18372 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18373 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18374 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18375 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18376 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18377 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18378 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18379 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18380 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18381 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18382 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18383 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18384 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18385 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18386 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18387 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18388 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18389 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18390 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18391 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18392 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18393 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18394 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655