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 * `ASN1_verify()`, `ASN1_digest()` and `ASN1_sign()` have been deprecated.
28 They are old functions that we don't use, and that you could disable with
29 the macro `NO_ASN1_OLD`. This goes all the way back to OpenSSL 0.9.7.
33 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
34 have been converted to Markdown with the goal to produce documents
35 which not only look pretty when viewed online in the browser, but
36 remain well readable inside a plain text editor.
38 To achieve this goal, a 'minimalistic' Markdown style has been applied
39 which avoids formatting elements that interfere too much with the
40 reading flow in the text file. For example, it
42 * avoids [ATX headings][] and uses [setext headings][] instead
43 (which works for `<h1>` and `<h2>` headings only).
44 * avoids [inline links][] and uses [reference links][] instead.
45 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
47 [ATX headings]: https://github.github.com/gfm/#atx-headings
48 [setext headings]: https://github.github.com/gfm/#setext-headings
49 [inline links]: https://github.github.com/gfm/#inline-link
50 [reference links]: https://github.github.com/gfm/#reference-link
51 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
52 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
57 * The test suite is changed to preserve results of each test recipe.
58 A new directory test-runs/ with subdirectories named like the
59 test recipes are created in the build tree for this purpose.
63 * The command line utilities ecparam and ec have been deprecated. Instead
64 use the pkeyparam, pkey and genpkey programs.
68 * All of the low level RSA functions have been deprecated including:
70 RSA_new_method, RSA_bits, RSA_size, RSA_security_bits,
71 RSA_get0_pss_params, RSA_get_version, RSA_get0_engine,
72 RSA_generate_key_ex, RSA_generate_multi_prime_key,
73 RSA_X931_derive_ex, RSA_X931_generate_key_ex, RSA_check_key,
74 RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
75 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
76 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
77 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
78 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING,
79 RSA_blinding_on, RSA_blinding_off, RSA_setup_blinding,
80 RSA_padding_add_PKCS1_type_1, RSA_padding_check_PKCS1_type_1,
81 RSA_padding_add_PKCS1_type_2, RSA_padding_check_PKCS1_type_2,
82 PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP, RSA_padding_check_PKCS1_OAEP,
83 RSA_padding_add_PKCS1_OAEP_mgf1, RSA_padding_check_PKCS1_OAEP_mgf1,
84 RSA_padding_add_SSLv23, RSA_padding_check_SSLv23,
85 RSA_padding_add_none, RSA_padding_check_none, RSA_padding_add_X931,
86 RSA_padding_check_X931, RSA_X931_hash_id, RSA_verify_PKCS1_PSS,
87 RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
88 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
89 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
90 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
91 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
92 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
93 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
94 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
95 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
96 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
97 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
98 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
99 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
101 Use of these low level functions has been informally discouraged for a long
102 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
103 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
104 L<EVP_PKEY_decrypt(3)>.
108 * X509 certificates signed using SHA1 are no longer allowed at security
110 In TLS/SSL the default security level is 1. It can be set either
111 using the cipher string with @SECLEVEL, or calling
112 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
113 a call to SSL_CTX_use_certificate() will fail if the security level is not
115 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
116 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
121 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
122 deprecated. Instead use the pkeyparam, pkey, genpkey and pkeyparam
123 programs respectively.
127 * All of the low level DH functions have been deprecated including:
129 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
130 DH_new_method, DH_bits, DH_size, DH_security_bits, DH_get_ex_new_index,
131 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
132 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
133 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
134 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
135 DH_KDF_X9_42, DH_get0_engine, DH_get_length, DH_set_length, DH_meth_new,
136 DH_meth_free, DH_meth_dup, DH_meth_get0_name, DH_meth_set1_name,
137 DH_meth_get_flags, DH_meth_set_flags, DH_meth_get0_app_data,
138 DH_meth_set0_app_data, DH_meth_get_generate_key,
139 DH_meth_set_generate_key, DH_meth_get_compute_key,
140 DH_meth_set_compute_key, DH_meth_get_bn_mod_exp,
141 DH_meth_set_bn_mod_exp, DH_meth_get_init, DH_meth_set_init,
142 DH_meth_get_finish, DH_meth_set_finish, DH_meth_get_generate_params
143 and DH_meth_set_generate_params.
145 Use of these low level functions has been informally discouraged for a long
146 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
147 and L<EVP_PKEY_derive(3)>.
151 * All of the low level DSA functions have been deprecated including:
153 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
154 DSA_get_default_method, DSA_set_method, DSA_get_method, DSA_new_method,
155 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
156 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
157 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
158 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
159 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
160 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
161 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
162 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
163 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
164 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
165 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
167 Use of these low level functions has been informally discouraged for a long
168 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
169 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
173 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
174 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
175 This means that applications don't have to look at the curve NID and
176 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
177 However, they still can, that EVP_PKEY_set_alias_type() call acts as
178 a no-op when the EVP_PKEY is already of the given type.
180 Parameter and key generation is also reworked to make it possible
181 to generate EVP_PKEY_SM2 parameters and keys without having to go
182 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
183 However, code that does the latter will still work as before.
187 * Deprecated low level ECDH and ECDSA functions. These include:
189 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
190 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
193 Use of these low level functions has been informally discouraged for a long
194 time. Instead applications should use the EVP_PKEY_derive(3),
195 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
199 * Deprecated the EC_KEY_METHOD functions. These include:
201 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
202 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
203 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
204 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
205 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
206 EC_KEY_METHOD_get_verify.
208 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
212 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
213 and EVP_PKEY_decrypt() instead.
214 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
215 and EVP_PKEY_encrypt() instead.
219 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
220 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
221 a new formulation to include all the things it can be used for,
222 as well as words of caution.
226 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
227 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
231 * All of the low level HMAC functions have been deprecated including:
233 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
234 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
237 Use of these low level functions has been informally discouraged for a long
238 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
239 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
240 and L<EVP_MAC_final(3)>.
244 * Over two thousand fixes were made to the documentation, including:
245 - Common options (such as -rand/-writerand, TLS version control, etc)
246 were refactored and point to newly-enhanced descriptions in openssl.pod.
247 - Added style conformance for all options (with help from Richard Levitte),
248 documented all reported missing options, added a CI build to check
249 that all options are documented and that no unimplemented options
251 - Documented some internals, such as all use of environment variables.
252 - Addressed all internal broken L<> references.
256 * All of the low level CMAC functions have been deprecated including:
258 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
259 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
261 Use of these low level functions has been informally discouraged for a long
262 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
263 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
264 and L<EVP_MAC_final(3)>.
268 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
269 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
272 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
273 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
274 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
275 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
276 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
277 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
278 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
279 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
280 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
281 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
283 Use of these low level functions has been informally discouraged
284 for a long time. Applications should use the EVP_DigestInit_ex(3),
285 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
289 * Corrected the documentation of the return values from the EVP_DigestSign*
290 set of functions. The documentation mentioned negative values for some
291 errors, but this was never the case, so the mention of negative values
294 Code that followed the documentation and thereby check with something
295 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
299 * All of the low level cipher functions have been deprecated including:
301 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
302 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
303 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
304 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
305 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
306 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
307 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
308 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
309 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
310 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
311 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
312 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
313 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
314 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
315 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
316 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
317 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
318 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
319 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
320 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
321 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
322 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
323 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
324 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
325 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
326 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
327 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
328 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
329 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
331 Use of these low level functions has been informally discouraged for
332 a long time. Applications should use the high level EVP APIs, e.g.
333 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
334 equivalently named decrypt functions instead.
336 *Matt Caswell and Paul Dale*
338 * Removed include/openssl/opensslconf.h.in and replaced it with
339 include/openssl/configuration.h.in, which differs in not including
340 <openssl/macros.h>. A short header include/openssl/opensslconf.h
341 was added to include both.
343 This allows internal hacks where one might need to modify the set
344 of configured macros, for example this if deprecated symbols are
345 still supposed to be available internally:
347 #include <openssl/configuration.h>
349 #undef OPENSSL_NO_DEPRECATED
350 #define OPENSSL_SUPPRESS_DEPRECATED
352 #include <openssl/macros.h>
354 This should not be used by applications that use the exported
355 symbols, as that will lead to linking errors.
359 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
360 used in exponentiation with 512-bit moduli. No EC algorithms are
361 affected. Analysis suggests that attacks against 2-prime RSA1024,
362 3-prime RSA1536, and DSA1024 as a result of this defect would be very
363 difficult to perform and are not believed likely. Attacks against DH512
364 are considered just feasible. However, for an attack the target would
365 have to re-use the DH512 private key, which is not recommended anyway.
366 Also applications directly using the low level API BN_mod_exp may be
367 affected if they use BN_FLG_CONSTTIME.
372 * Most memory-debug features have been deprecated, and the functionality
373 replaced with no-ops.
377 * Introduced a new method type and API, OSSL_SERIALIZER, to
378 represent generic serializers. An implementation is expected to
379 be able to serialize an object associated with a given name (such
380 as an algorithm name for an asymmetric key) into forms given by
381 implementation properties.
383 Serializers are primarily used from inside libcrypto, through
384 calls to functions like EVP_PKEY_print_private(),
385 PEM_write_bio_PrivateKey() and similar.
387 Serializers are specified in such a way that they can be made to
388 directly handle the provider side portion of an object, if this
389 provider side part comes from the same provider as the serializer
390 itself, but can also be made to handle objects in parametrized
391 form (as an OSSL_PARAM array of data). This allows a provider to
392 offer generic serializers as a service for any other provider.
396 * Added a .pragma directive to the syntax of configuration files, to
397 allow varying behavior in a supported and predictable manner.
398 Currently added pragma:
402 This allows dollar signs to be a keyword character unless it's
403 followed by a opening brace or parenthesis. This is useful for
404 platforms where dollar signs are commonly used in names, such as
405 volume names and system directory names on VMS.
409 * Added functionality to create an EVP_PKEY from user data. This
410 is effectively the same as creating a RSA, DH or DSA object and
411 then assigning them to an EVP_PKEY, but directly using algorithm
412 agnostic EVP functions. A benefit is that this should be future
413 proof for public key algorithms to come.
417 * Change the interpretation of the '--api' configuration option to
418 mean that this is a desired API compatibility level with no
419 further meaning. The previous interpretation, that this would
420 also mean to remove all deprecated symbols up to and including
421 the given version, no requires that 'no-deprecated' is also used
422 in the configuration.
424 When building applications, the desired API compatibility level
425 can be set with the OPENSSL_API_COMPAT macro like before. For
426 API compatibility version below 3.0, the old style numerical
427 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
428 For version 3.0 and on, the value is expected to be the decimal
429 value calculated from the major and minor version like this:
431 MAJOR * 10000 + MINOR * 100
435 -DOPENSSL_API_COMPAT=30000 For 3.0
436 -DOPENSSL_API_COMPAT=30200 For 3.2
438 To hide declarations that are deprecated up to and including the
439 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
440 given when building the application as well.
444 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
445 access to certificate and CRL stores via URIs and OSSL_STORE
448 This adds the following functions:
450 - X509_LOOKUP_store()
451 - X509_STORE_load_file()
452 - X509_STORE_load_path()
453 - X509_STORE_load_store()
454 - SSL_add_store_cert_subjects_to_stack()
455 - SSL_CTX_set_default_verify_store()
456 - SSL_CTX_load_verify_file()
457 - SSL_CTX_load_verify_dir()
458 - SSL_CTX_load_verify_store()
460 Also, the following functions are now deprecated:
462 - X509_STORE_load_locations() (use X509_STORE_load_file(),
463 X509_STORE_load_path() or X509_STORE_load_store() instead)
464 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
465 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
469 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
470 The presence of this system service is determined at run-time.
474 * Added functionality to create an EVP_PKEY context based on data
475 for methods from providers. This takes an algorithm name and a
476 property query string and simply stores them, with the intent
477 that any operation that uses this context will use those strings
478 to fetch the needed methods implicitly, thereby making the port
479 of application written for pre-3.0 OpenSSL easier.
483 * The undocumented function NCONF_WIN32() has been deprecated; for
484 conversion details see the HISTORY section of doc/man5/config.pod
488 * Introduced the new functions EVP_DigestSignInit_ex() and
489 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
490 EVP_DigestVerifyUpdate() have been converted to functions. See the man
491 pages for further details.
495 * Over two thousand fixes were made to the documentation, including:
496 adding missing command flags, better style conformance, documentation
499 *Rich Salz, Richard Levitte*
501 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
502 X25519, X448, Ed25519 and Ed448.
506 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
511 * Deprecated the public definition of ERR_STATE as well as the function
512 ERR_get_state(). This is done in preparation of making ERR_STATE an
517 * Added ERR functionality to give callers access to the stored function
518 names that have replaced the older function code based functions.
520 New functions are ERR_get_error_func(), ERR_peek_error_func(),
521 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
522 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
523 and ERR_peek_last_error_all().
525 These functions have become deprecated: ERR_get_error_line_data(),
526 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
527 ERR_func_error_string().
531 * Extended testing to be verbose for failing tests only. The make variables
532 VERBOSE_FAILURE or VF can be used to enable this:
534 $ make VF=1 test # Unix
535 $ mms /macro=(VF=1) test ! OpenVMS
536 $ nmake VF=1 test # Windows
541 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
542 used even when parsing explicit parameters, when loading a serialized key
543 or calling `EC_GROUP_new_from_ecpkparameters()`/
544 `EC_GROUP_new_from_ecparameters()`.
545 This prevents bypass of security hardening and performance gains,
546 especially for curves with specialized EC_METHODs.
547 By default, if a key encoded with explicit parameters is loaded and later
548 serialized, the output is still encoded with explicit parameters, even if
549 internally a "named" EC_GROUP is used for computation.
553 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
554 this change, EC_GROUP_set_generator would accept order and/or cofactor as
555 NULL. After this change, only the cofactor parameter can be NULL. It also
556 does some minimal sanity checks on the passed order.
561 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
562 An attack is simple, if the first CMS_recipientInfo is valid but the
563 second CMS_recipientInfo is chosen ciphertext. If the second
564 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
565 encryption key will be replaced by garbage, and the message cannot be
566 decoded, but if the RSA decryption fails, the correct encryption key is
567 used and the recipient will not notice the attack.
568 As a work around for this potential attack the length of the decrypted
569 key must be equal to the cipher default key length, in case the
570 certifiate is not given and all recipientInfo are tried out.
571 The old behaviour can be re-enabled in the CMS code by setting the
572 CMS_DEBUG_DECRYPT flag.
576 * Early start up entropy quality from the DEVRANDOM seed source has been
577 improved for older Linux systems. The RAND subsystem will wait for
578 /dev/random to be producing output before seeding from /dev/urandom.
579 The seeded state is stored for future library initialisations using
580 a system global shared memory segment. The shared memory identifier
581 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
582 the desired value. The default identifier is 114.
586 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
587 when primes for RSA keys are computed.
588 Since we previously always generated primes == 2 (mod 3) for RSA keys,
589 the 2-prime and 3-prime RSA modules were easy to distinguish, since
590 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
591 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
592 This avoids possible fingerprinting of newly generated RSA modules.
596 * Correct the extended master secret constant on EBCDIC systems. Without this
597 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
598 negotiate EMS will fail. Unfortunately this also means that TLS connections
599 between EBCDIC systems with this fix, and EBCDIC systems without this
600 fix will fail if they negotiate EMS.
604 * Changed the library initialisation so that the config file is now loaded
605 by default. This was already the case for libssl. It now occurs for both
606 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
607 OPENSSL_init_crypto() to suppress automatic loading of a config file.
611 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
612 where the former acts as a replacement for ERR_put_error(), and the
613 latter replaces the combination ERR_put_error()+ERR_add_error_data().
614 ERR_raise_data() adds more flexibility by taking a format string and
615 an arbitrary number of arguments following it, to be processed with
620 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
621 to check if a named provider is loaded and available. When called, it
622 will also activate all fallback providers if such are still present.
626 * Enforce a minimum DH modulus size of 512 bits.
630 * Changed DH parameters to generate the order q subgroup instead of 2q.
631 Previously generated DH parameters are still accepted by DH_check
632 but DH_generate_key works around that by clearing bit 0 of the
633 private key for those. This avoids leaking bit 0 of the private key.
637 * Significantly reduce secure memory usage by the randomness pools.
641 * {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
646 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
647 algorithms. An implementation of a key exchange algorithm can be obtained
648 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
649 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
650 the older EVP_PKEY_derive_init() function. See the man pages for the new
651 functions for further details.
655 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
659 * Removed the function names from error messages and deprecated the
662 * Removed NextStep support and the macro OPENSSL_UNISTD
666 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
667 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
668 Also removed "export var as function" capability; we do not export
669 variables, only functions.
673 * RC5_32_set_key has been changed to return an int type, with 0 indicating
674 an error and 1 indicating success. In previous versions of OpenSSL this
675 was a void type. If a key was set longer than the maximum possible this
680 * Support SM2 signing and verification schemes with X509 certificate.
684 * Use SHA256 as the default digest for TS query in the ts app.
688 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
689 This checks that the salt length is at least 128 bits, the derived key
690 length is at least 112 bits, and that the iteration count is at least 1000.
691 For backwards compatibility these checks are disabled by default in the
692 default provider, but are enabled by default in the fips provider.
693 To enable or disable these checks use the control
694 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
698 * Default cipher lists/suites are now available via a function, the
699 #defines are deprecated.
703 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
704 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
705 for Windows Store apps easier. Also, the "no-uplink" option has been added.
709 * Join the directories crypto/x509 and crypto/x509v3
713 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
714 This changes the size when using the genpkey app when no size is given. It
715 fixes an omission in earlier changes that changed all RSA, DSA and DH
716 generation apps to use 2048 bits by default.
720 * Added command 'openssl kdf' that uses the EVP_KDF API.
724 * Added command 'openssl mac' that uses the EVP_MAC API.
728 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
729 as default directories. Also added the command 'openssl info'
730 for scripting purposes.
734 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
735 deprecated. These undocumented functions were never integrated into the EVP
736 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
737 Bi-directional IGE mode. These modes were never formally standardised and
738 usage of these functions is believed to be very small. In particular
739 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
740 is ever used. The security implications are believed to be minimal, but
741 this issue was never fixed for backwards compatibility reasons. New code
742 should not use these modes.
746 * Add prediction resistance to the DRBG reseeding process.
750 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
751 mandated by IEEE Std 1619-2018.
755 * Added newline escaping functionality to a filename when using openssl dgst.
756 This output format is to replicate the output format found in the '*sum'
757 checksum programs. This aims to preserve backward compatibility.
759 *Matt Eaton, Richard Levitte, and Paul Dale*
761 * Removed the heartbeat message in DTLS feature, as it has very
762 little usage and doesn't seem to fulfill a valuable purpose.
763 The configuration option is now deprecated.
767 * Changed the output of 'openssl {digestname} < file' to display the
768 digest name in its output.
772 * Added a new generic trace API which provides support for enabling
773 instrumentation through trace output. This feature is mainly intended
774 as an aid for developers and is disabled by default. To utilize it,
775 OpenSSL needs to be configured with the `enable-trace` option.
777 If the tracing API is enabled, the application can activate trace output
778 by registering BIOs as trace channels for a number of tracing and debugging
781 The 'openssl' application has been expanded to enable any of the types
782 available via environment variables defined by the user, and serves as
783 one possible example on how to use this functionality.
785 *Richard Levitte & Matthias St. Pierre*
787 * Added build tests for C++. These are generated files that only do one
788 thing, to include one public OpenSSL head file each. This tests that
789 the public header files can be usefully included in a C++ application.
791 This test isn't enabled by default. It can be enabled with the option
792 'enable-buildtest-c++'.
796 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
800 * Add KMAC to EVP_MAC.
804 * Added property based algorithm implementation selection framework to
809 * Added SCA hardening for modular field inversion in EC_GROUP through
810 a new dedicated field_inv() pointer in EC_METHOD.
811 This also addresses a leakage affecting conversions from projective
812 to affine coordinates.
814 *Billy Bob Brumley, Nicola Tuveri*
816 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
817 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
818 those algorithms that were already supported through the EVP_PKEY API
819 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
820 and scrypt are now wrappers that call EVP_KDF.
824 * Build devcrypto engine as a dynamic engine.
828 * Add keyed BLAKE2 to EVP_MAC.
832 * Fix a bug in the computation of the endpoint-pair shared secret used
833 by DTLS over SCTP. This breaks interoperability with older versions
834 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
835 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
836 interoperability with such broken implementations. However, enabling
837 this switch breaks interoperability with correct implementations.
839 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
840 re-used X509_PUBKEY object if the second PUBKEY is malformed.
844 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
848 * Change the license to the Apache License v2.0.
852 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
854 - Major releases (indicated by incrementing the MAJOR release number)
855 may introduce incompatible API/ABI changes.
856 - Minor releases (indicated by incrementing the MINOR release number)
857 may introduce new features but retain API/ABI compatibility.
858 - Patch releases (indicated by incrementing the PATCH number)
859 are intended for bug fixes and other improvements of existing
860 features only (like improving performance or adding documentation)
861 and retain API/ABI compatibility.
865 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
869 * Remove the 'dist' target and add a tarball building script. The
870 'dist' target has fallen out of use, and it shouldn't be
871 necessary to configure just to create a source distribution.
875 * Recreate the OS390-Unix config target. It no longer relies on a
876 special script like it did for OpenSSL pre-1.1.0.
880 * Instead of having the source directories listed in Configure, add
881 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
886 * Add GMAC to EVP_MAC.
890 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
894 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
895 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
896 to facilitate the continued use of MACs through raw private keys in
897 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
901 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
902 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
906 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
907 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
908 are retained for backwards compatibility.
912 * AES-XTS mode now enforces that its two keys are different to mitigate
913 the attacked described in "Efficient Instantiations of Tweakable
914 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
915 Details of this attack can be obtained from:
916 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
920 * Rename the object files, i.e. give them other names than in previous
921 versions. Their names now include the name of the final product, as
922 well as its type mnemonic (bin, lib, shlib).
926 * Added new option for 'openssl list', '-objects', which will display the
927 list of built in objects, i.e. OIDs with names.
931 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
932 improves application performance by removing data copies and providing
933 applications with zero-copy system calls such as sendfile and splice.
941 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx] ###
944 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020] ###
946 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
947 while reading in libssl then we would report an error back to the
948 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
949 an error to the stack (which means we instead return SSL_ERROR_SSL) and
950 therefore give a hint as to what went wrong.
954 * Check that ed25519 and ed448 are allowed by the security level. Previously
955 signature algorithms not using an MD were not being checked that they were
956 allowed by the security level.
960 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
961 was not quite right. The behaviour was not consistent between resumption
962 and normal handshakes, and also not quite consistent with historical
963 behaviour. The behaviour in various scenarios has been clarified and
964 it has been updated to make it match historical behaviour as closely as
969 * *[VMS only]* The header files that the VMS compilers include automatically,
970 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
971 that the C++ compiler doesn't understand. This is a shortcoming in the
972 compiler, but can be worked around with `__cplusplus` guards.
974 C++ applications that use OpenSSL libraries must be compiled using the
975 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
976 functions. Otherwise, only functions with symbols of less than 31
977 characters can be used, as the linker will not be able to successfully
978 resolve symbols with longer names.
982 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
983 The presence of this system service is determined at run-time.
987 * Added newline escaping functionality to a filename when using openssl dgst.
988 This output format is to replicate the output format found in the '*sum'
989 checksum programs. This aims to preserve backward compatibility.
991 *Matt Eaton, Richard Levitte, and Paul Dale*
993 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
998 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019] ###
1000 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1001 number generator (RNG). This was intended to include protection in the
1002 event of a fork() system call in order to ensure that the parent and child
1003 processes did not share the same RNG state. However this protection was not
1004 being used in the default case.
1006 A partial mitigation for this issue is that the output from a high
1007 precision timer is mixed into the RNG state so the likelihood of a parent
1008 and child process sharing state is significantly reduced.
1010 If an application already calls OPENSSL_init_crypto() explicitly using
1011 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1014 *Matthias St. Pierre*
1016 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1017 used even when parsing explicit parameters, when loading a serialized key
1018 or calling `EC_GROUP_new_from_ecpkparameters()`/
1019 `EC_GROUP_new_from_ecparameters()`.
1020 This prevents bypass of security hardening and performance gains,
1021 especially for curves with specialized EC_METHODs.
1022 By default, if a key encoded with explicit parameters is loaded and later
1023 serialized, the output is still encoded with explicit parameters, even if
1024 internally a "named" EC_GROUP is used for computation.
1028 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1029 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1030 NULL. After this change, only the cofactor parameter can be NULL. It also
1031 does some minimal sanity checks on the passed order.
1036 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1037 An attack is simple, if the first CMS_recipientInfo is valid but the
1038 second CMS_recipientInfo is chosen ciphertext. If the second
1039 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1040 encryption key will be replaced by garbage, and the message cannot be
1041 decoded, but if the RSA decryption fails, the correct encryption key is
1042 used and the recipient will not notice the attack.
1043 As a work around for this potential attack the length of the decrypted
1044 key must be equal to the cipher default key length, in case the
1045 certifiate is not given and all recipientInfo are tried out.
1046 The old behaviour can be re-enabled in the CMS code by setting the
1047 CMS_DEBUG_DECRYPT flag.
1052 * Early start up entropy quality from the DEVRANDOM seed source has been
1053 improved for older Linux systems. The RAND subsystem will wait for
1054 /dev/random to be producing output before seeding from /dev/urandom.
1055 The seeded state is stored for future library initialisations using
1056 a system global shared memory segment. The shared memory identifier
1057 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1058 the desired value. The default identifier is 114.
1062 * Correct the extended master secret constant on EBCDIC systems. Without this
1063 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1064 negotiate EMS will fail. Unfortunately this also means that TLS connections
1065 between EBCDIC systems with this fix, and EBCDIC systems without this
1066 fix will fail if they negotiate EMS.
1070 * Use Windows installation paths in the mingw builds
1072 Mingw isn't a POSIX environment per se, which means that Windows
1073 paths should be used for installation.
1078 * Changed DH_check to accept parameters with order q and 2q subgroups.
1079 With order 2q subgroups the bit 0 of the private key is not secret
1080 but DH_generate_key works around that by clearing bit 0 of the
1081 private key for those. This avoids leaking bit 0 of the private key.
1085 * Significantly reduce secure memory usage by the randomness pools.
1089 * Revert the DEVRANDOM_WAIT feature for Linux systems
1091 The DEVRANDOM_WAIT feature added a select() call to wait for the
1092 /dev/random device to become readable before reading from the
1093 /dev/urandom device.
1095 It turned out that this change had negative side effects on
1096 performance which were not acceptable. After some discussion it
1097 was decided to revert this feature and leave it up to the OS
1098 resp. the platform maintainer to ensure a proper initialization
1099 during early boot time.
1101 *Matthias St. Pierre*
1103 ### Changes between 1.1.1b and 1.1.1c [28 May 2019] ###
1105 * Add build tests for C++. These are generated files that only do one
1106 thing, to include one public OpenSSL head file each. This tests that
1107 the public header files can be usefully included in a C++ application.
1109 This test isn't enabled by default. It can be enabled with the option
1110 'enable-buildtest-c++'.
1114 * Enable SHA3 pre-hashing for ECDSA and DSA.
1118 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1119 This changes the size when using the genpkey app when no size is given. It
1120 fixes an omission in earlier changes that changed all RSA, DSA and DH
1121 generation apps to use 2048 bits by default.
1125 * Reorganize the manual pages to consistently have RETURN VALUES,
1126 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1127 util/fix-doc-nits accordingly.
1129 *Paul Yang, Joshua Lock*
1131 * Add the missing accessor EVP_PKEY_get0_engine()
1135 * Have apps like 's_client' and 's_server' output the signature scheme
1136 along with other cipher suite parameters when debugging.
1140 * Make OPENSSL_config() error agnostic again.
1144 * Do the error handling in RSA decryption constant time.
1148 * Prevent over long nonces in ChaCha20-Poly1305.
1150 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1151 for every encryption operation. RFC 7539 specifies that the nonce value
1152 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1153 and front pads the nonce with 0 bytes if it is less than 12
1154 bytes. However it also incorrectly allows a nonce to be set of up to 16
1155 bytes. In this case only the last 12 bytes are significant and any
1156 additional leading bytes are ignored.
1158 It is a requirement of using this cipher that nonce values are
1159 unique. Messages encrypted using a reused nonce value are susceptible to
1160 serious confidentiality and integrity attacks. If an application changes
1161 the default nonce length to be longer than 12 bytes and then makes a
1162 change to the leading bytes of the nonce expecting the new value to be a
1163 new unique nonce then such an application could inadvertently encrypt
1164 messages with a reused nonce.
1166 Additionally the ignored bytes in a long nonce are not covered by the
1167 integrity guarantee of this cipher. Any application that relies on the
1168 integrity of these ignored leading bytes of a long nonce may be further
1169 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1170 is safe because no such use sets such a long nonce value. However user
1171 applications that use this cipher directly and set a non-default nonce
1172 length to be longer than 12 bytes may be vulnerable.
1174 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1180 * Add DEVRANDOM_WAIT feature for Linux systems
1182 On older Linux systems where the getrandom() system call is not available,
1183 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1184 Contrary to getrandom(), the /dev/urandom device will not block during
1185 early boot when the kernel CSPRNG has not been seeded yet.
1187 To mitigate this known weakness, use select() to wait for /dev/random to
1188 become readable before reading from /dev/urandom.
1190 * Ensure that SM2 only uses SM3 as digest algorithm
1194 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019] ###
1196 * Change the info callback signals for the start and end of a post-handshake
1197 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1198 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1199 confused by this and assume that a TLSv1.2 renegotiation has started. This
1200 can break KeyUpdate handling. Instead we no longer signal the start and end
1201 of a post handshake message exchange (although the messages themselves are
1202 still signalled). This could break some applications that were expecting
1203 the old signals. However without this KeyUpdate is not usable for many
1208 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018] ###
1210 * Timing vulnerability in DSA signature generation
1212 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1213 timing side channel attack. An attacker could use variations in the signing
1214 algorithm to recover the private key.
1216 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1221 * Timing vulnerability in ECDSA signature generation
1223 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1224 timing side channel attack. An attacker could use variations in the signing
1225 algorithm to recover the private key.
1227 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1232 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1233 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1234 of two gigabytes and the error handling improved.
1236 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1237 categorized as a normal bug, not a security issue, because the DRBG reseeds
1238 automatically and is fully functional even without additional randomness
1239 provided by the application.
1241 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018] ###
1243 * Add a new ClientHello callback. Provides a callback interface that gives
1244 the application the ability to adjust the nascent SSL object at the
1245 earliest stage of ClientHello processing, immediately after extensions have
1246 been collected but before they have been processed. In particular, this
1247 callback can adjust the supported TLS versions in response to the contents
1252 * Add SM2 base algorithm support.
1256 * s390x assembly pack: add (improved) hardware-support for the following
1257 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1258 aes-cfb/cfb8, aes-ecb.
1262 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1263 parameter is no longer accepted, as it leads to a corrupt table. NULL
1264 pem_str is reserved for alias entries only.
1268 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1269 step for prime curves. The new implementation is based on formulae from
1270 differential addition-and-doubling in homogeneous projective coordinates
1271 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1272 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1273 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1274 to work in projective coordinates.
1276 *Billy Bob Brumley, Nicola Tuveri*
1278 * Change generating and checking of primes so that the error rate of not
1279 being prime depends on the intended use based on the size of the input.
1280 For larger primes this will result in more rounds of Miller-Rabin.
1281 The maximal error rate for primes with more than 1080 bits is lowered
1284 *Kurt Roeckx, Annie Yousar*
1286 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1290 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1291 moving between systems, and to avoid confusion when a Windows build is
1292 done with mingw vs with MSVC. For POSIX installs, there's still a
1293 symlink or copy named 'tsget' to avoid that confusion as well.
1297 * Revert blinding in ECDSA sign and instead make problematic addition
1298 length-invariant. Switch even to fixed-length Montgomery multiplication.
1302 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1303 step for binary curves. The new implementation is based on formulae from
1304 differential addition-and-doubling in mixed Lopez-Dahab projective
1305 coordinates, modified to independently blind the operands.
1307 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1309 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1310 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1311 EC_METHODs to implement their own specialized "ladder step", to take
1312 advantage of more favorable coordinate systems or more efficient
1313 differential addition-and-doubling algorithms.
1315 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1317 * Modified the random device based seed sources to keep the relevant
1318 file descriptors open rather than reopening them on each access.
1319 This allows such sources to operate in a chroot() jail without
1320 the associated device nodes being available. This behaviour can be
1321 controlled using RAND_keep_random_devices_open().
1325 * Numerous side-channel attack mitigations have been applied. This may have
1326 performance impacts for some algorithms for the benefit of improved
1327 security. Specific changes are noted in this change log by their respective
1332 * AIX shared library support overhaul. Switch to AIX "natural" way of
1333 handling shared libraries, which means collecting shared objects of
1334 different versions and bitnesses in one common archive. This allows to
1335 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1336 doesn't affect the way 3rd party applications are linked, only how
1337 multi-version installation is managed.
1341 * Make ec_group_do_inverse_ord() more robust and available to other
1342 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1343 mitigations are applied to the fallback BN_mod_inverse().
1344 When using this function rather than BN_mod_inverse() directly, new
1345 EC cryptosystem implementations are then safer-by-default.
1349 * Add coordinate blinding for EC_POINT and implement projective
1350 coordinate blinding for generic prime curves as a countermeasure to
1351 chosen point SCA attacks.
1353 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1355 * Add blinding to ECDSA and DSA signatures to protect against side channel
1356 attacks discovered by Keegan Ryan (NCC Group).
1360 * Enforce checking in the pkeyutl command line app to ensure that the input
1361 length does not exceed the maximum supported digest length when performing
1362 a sign, verify or verifyrecover operation.
1366 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1367 I/O in combination with something like select() or poll() will hang. This
1368 can be turned off again using SSL_CTX_clear_mode().
1369 Many applications do not properly handle non-application data records, and
1370 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1371 around the problems in those applications, but can also break some.
1372 It's recommended to read the manpages about SSL_read(), SSL_write(),
1373 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1374 SSL_CTX_set_read_ahead() again.
1378 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1379 now allow empty (zero character) pass phrases.
1383 * Apply blinding to binary field modular inversion and remove patent
1384 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1388 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1389 binary and prime elliptic curves.
1393 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1394 constant time fixed point multiplication.
1398 * Revise elliptic curve scalar multiplication with timing attack
1399 defenses: ec_wNAF_mul redirects to a constant time implementation
1400 when computing fixed point and variable point multiplication (which
1401 in OpenSSL are mostly used with secret scalars in keygen, sign,
1402 ECDH derive operations).
1403 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1406 * Updated CONTRIBUTING
1410 * Updated DRBG / RAND to request nonce and additional low entropy
1411 randomness from the system.
1413 *Matthias St. Pierre*
1415 * Updated 'openssl rehash' to use OpenSSL consistent default.
1419 * Moved the load of the ssl_conf module to libcrypto, which helps
1420 loading engines that libssl uses before libssl is initialised.
1424 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1428 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1430 *Ingo Schwarze, Rich Salz*
1432 * Added output of accepting IP address and port for 'openssl s_server'
1436 * Added a new API for TLSv1.3 ciphersuites:
1437 SSL_CTX_set_ciphersuites()
1438 SSL_set_ciphersuites()
1442 * Memory allocation failures consistently add an error to the error
1447 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1448 in libcrypto when run as setuid/setgid.
1452 * Load any config file by default when libssl is used.
1456 * Added new public header file <openssl/rand_drbg.h> and documentation
1457 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1459 *Matthias St. Pierre*
1461 * QNX support removed (cannot find contributors to get their approval
1462 for the license change).
1466 * TLSv1.3 replay protection for early data has been implemented. See the
1467 SSL_read_early_data() man page for further details.
1471 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1472 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1473 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1474 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1475 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1476 configuration has been separated out. See the ciphers man page or the
1477 SSL_CTX_set_ciphersuites() man page for more information.
1481 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1482 in responder mode now supports the new "-multi" option, which
1483 spawns the specified number of child processes to handle OCSP
1484 requests. The "-timeout" option now also limits the OCSP
1485 responder's patience to wait to receive the full client request
1486 on a newly accepted connection. Child processes are respawned
1487 as needed, and the CA index file is automatically reloaded
1488 when changed. This makes it possible to run the "ocsp" responder
1489 as a long-running service, making the OpenSSL CA somewhat more
1490 feature-complete. In this mode, most diagnostic messages logged
1491 after entering the event loop are logged via syslog(3) rather than
1496 * Added support for X448 and Ed448. Heavily based on original work by
1501 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1502 objects loaded. This adds the functions OSSL_STORE_expect() and
1503 OSSL_STORE_find() as well as needed tools to construct searches and
1504 get the search data out of them.
1508 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1509 version of OpenSSL should review their configuration settings to ensure
1510 that they are still appropriate for TLSv1.3. For further information see:
1511 https://wiki.openssl.org/index.php/TLS1.3
1515 * Grand redesign of the OpenSSL random generator
1517 The default RAND method now utilizes an AES-CTR DRBG according to
1518 NIST standard SP 800-90Ar1. The new random generator is essentially
1519 a port of the default random generator from the OpenSSL FIPS 2.0
1520 object module. It is a hybrid deterministic random bit generator
1521 using an AES-CTR bit stream and which seeds and reseeds itself
1522 automatically using trusted system entropy sources.
1524 Some of its new features are:
1525 - Support for multiple DRBG instances with seed chaining.
1526 - The default RAND method makes use of a DRBG.
1527 - There is a public and private DRBG instance.
1528 - The DRBG instances are fork-safe.
1529 - Keep all global DRBG instances on the secure heap if it is enabled.
1530 - The public and private DRBG instance are per thread for lock free
1533 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1535 * Changed Configure so it only says what it does and doesn't dump
1536 so much data. Instead, ./configdata.pm should be used as a script
1537 to display all sorts of configuration data.
1541 * Added processing of "make variables" to Configure.
1545 * Added SHA512/224 and SHA512/256 algorithm support.
1549 * The last traces of Netware support, first removed in 1.1.0, have
1554 * Get rid of Makefile.shared, and in the process, make the processing
1555 of certain files (rc.obj, or the .def/.map/.opt files produced from
1556 the ordinal files) more visible and hopefully easier to trace and
1557 debug (or make silent).
1561 * Make it possible to have environment variable assignments as
1562 arguments to config / Configure.
1566 * Add multi-prime RSA (RFC 8017) support.
1570 * Add SM3 implemented according to GB/T 32905-2016
1571 * Jack Lloyd <jack.lloyd@ribose.com>,
1572 Ronald Tse <ronald.tse@ribose.com>,
1573 Erick Borsboom <erick.borsboom@ribose.com> *
1575 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1576 as documented in RFC6066.
1577 Based on a patch from Tomasz Moń
1579 *Filipe Raimundo da Silva*
1581 * Add SM4 implemented according to GB/T 32907-2016.
1582 * Jack Lloyd <jack.lloyd@ribose.com>,
1583 Ronald Tse <ronald.tse@ribose.com>,
1584 Erick Borsboom <erick.borsboom@ribose.com> *
1586 * Reimplement -newreq-nodes and ERR_error_string_n; the
1587 original author does not agree with the license change.
1591 * Add ARIA AEAD TLS support.
1595 * Some macro definitions to support VS6 have been removed. Visual
1596 Studio 6 has not worked since 1.1.0
1600 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1601 without clearing the errors.
1605 * Add "atfork" functions. If building on a system that without
1606 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1607 requirements. The RAND facility now uses/requires this.
1615 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1616 not possible to disable entirely. However, it's still possible to
1617 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1620 To disable, configure with 'no-ui-console'. 'no-ui' is still
1621 possible to use as an alias. Check at compile time with the
1622 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1623 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1627 * Add a STORE module, which implements a uniform and URI based reader of
1628 stores that can contain keys, certificates, CRLs and numerous other
1629 objects. The main API is loosely based on a few stdio functions,
1630 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1631 OSSL_STORE_error and OSSL_STORE_close.
1632 The implementation uses backends called "loaders" to implement arbitrary
1633 URI schemes. There is one built in "loader" for the 'file' scheme.
1637 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1638 then adjusted to work on FreeBSD 8.4 as well.
1639 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1640 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1644 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1645 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1646 error code calls like this:
1648 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1650 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1651 that can be encoded in C. For the foreseeable future, this will only
1654 *Richard Levitte and Tim Hudson*
1656 * Removed BSD cryptodev engine.
1660 * Add a build target 'build_all_generated', to build all generated files
1661 and only that. This can be used to prepare everything that requires
1662 things like perl for a system that lacks perl and then move everything
1663 to that system and do the rest of the build there.
1667 * In the UI interface, make it possible to duplicate the user data. This
1668 can be used by engines that need to retain the data for a longer time
1669 than just the call where this user data is passed.
1673 * Ignore the '-named_curve auto' value for compatibility of applications
1676 *Tomas Mraz <tmraz@fedoraproject.org>*
1678 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1679 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1680 alerts across multiple records (some of which could be empty). In practice
1681 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1682 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1683 support this at all. Supporting it adds significant complexity to the
1684 record layer, and its removal is unlikely to cause interoperability
1689 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1690 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1691 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1696 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1697 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1699 *Richard Levitte, Andy Polyakov*
1701 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1706 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1707 platform rather than 'mingw'.
1711 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1712 success if they are asked to add an object which already exists
1713 in the store. This change cascades to other functions which load
1714 certificates and CRLs.
1718 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1719 facilitate stack unwinding even from assembly subroutines.
1723 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1724 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1728 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1729 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1730 which is the minimum version we support.
1734 * Certificate time validation (X509_cmp_time) enforces stricter
1735 compliance with RFC 5280. Fractional seconds and timezone offsets
1736 are no longer allowed.
1740 * Add support for ARIA
1744 * s_client will now send the Server Name Indication (SNI) extension by
1745 default unless the new "-noservername" option is used. The server name is
1746 based on the host provided to the "-connect" option unless overridden by
1747 using "-servername".
1751 * Add support for SipHash
1755 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1756 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1757 prevent issues where no progress is being made and the peer continually
1758 sends unrecognised record types, using up resources processing them.
1762 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1763 using the algorithm defined in
1764 https://www.akkadia.org/drepper/SHA-crypt.txt
1768 * Heartbeat support has been removed; the ABI is changed for now.
1770 *Richard Levitte, Rich Salz*
1772 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1776 * The RSA "null" method, which was partially supported to avoid patent
1777 issues, has been replaced to always returns NULL.
1785 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019] ###
1787 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1788 used even when parsing explicit parameters, when loading a serialized key
1789 or calling `EC_GROUP_new_from_ecpkparameters()`/
1790 `EC_GROUP_new_from_ecparameters()`.
1791 This prevents bypass of security hardening and performance gains,
1792 especially for curves with specialized EC_METHODs.
1793 By default, if a key encoded with explicit parameters is loaded and later
1794 serialized, the output is still encoded with explicit parameters, even if
1795 internally a "named" EC_GROUP is used for computation.
1799 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1800 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1801 NULL. After this change, only the cofactor parameter can be NULL. It also
1802 does some minimal sanity checks on the passed order.
1807 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1808 An attack is simple, if the first CMS_recipientInfo is valid but the
1809 second CMS_recipientInfo is chosen ciphertext. If the second
1810 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1811 encryption key will be replaced by garbage, and the message cannot be
1812 decoded, but if the RSA decryption fails, the correct encryption key is
1813 used and the recipient will not notice the attack.
1814 As a work around for this potential attack the length of the decrypted
1815 key must be equal to the cipher default key length, in case the
1816 certifiate is not given and all recipientInfo are tried out.
1817 The old behaviour can be re-enabled in the CMS code by setting the
1818 CMS_DEBUG_DECRYPT flag.
1823 * Use Windows installation paths in the mingw builds
1825 Mingw isn't a POSIX environment per se, which means that Windows
1826 paths should be used for installation.
1831 ### Changes between 1.1.0j and 1.1.0k [28 May 2019] ###
1833 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1834 This changes the size when using the genpkey app when no size is given. It
1835 fixes an omission in earlier changes that changed all RSA, DSA and DH
1836 generation apps to use 2048 bits by default.
1840 * Prevent over long nonces in ChaCha20-Poly1305.
1842 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1843 for every encryption operation. RFC 7539 specifies that the nonce value
1844 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1845 and front pads the nonce with 0 bytes if it is less than 12
1846 bytes. However it also incorrectly allows a nonce to be set of up to 16
1847 bytes. In this case only the last 12 bytes are significant and any
1848 additional leading bytes are ignored.
1850 It is a requirement of using this cipher that nonce values are
1851 unique. Messages encrypted using a reused nonce value are susceptible to
1852 serious confidentiality and integrity attacks. If an application changes
1853 the default nonce length to be longer than 12 bytes and then makes a
1854 change to the leading bytes of the nonce expecting the new value to be a
1855 new unique nonce then such an application could inadvertently encrypt
1856 messages with a reused nonce.
1858 Additionally the ignored bytes in a long nonce are not covered by the
1859 integrity guarantee of this cipher. Any application that relies on the
1860 integrity of these ignored leading bytes of a long nonce may be further
1861 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1862 is safe because no such use sets such a long nonce value. However user
1863 applications that use this cipher directly and set a non-default nonce
1864 length to be longer than 12 bytes may be vulnerable.
1866 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1872 * Added SCA hardening for modular field inversion in EC_GROUP through
1873 a new dedicated field_inv() pointer in EC_METHOD.
1874 This also addresses a leakage affecting conversions from projective
1875 to affine coordinates.
1877 *Billy Bob Brumley, Nicola Tuveri*
1879 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1880 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1884 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1888 * Remove the 'dist' target and add a tarball building script. The
1889 'dist' target has fallen out of use, and it shouldn't be
1890 necessary to configure just to create a source distribution.
1894 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018] ###
1896 * Timing vulnerability in DSA signature generation
1898 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1899 timing side channel attack. An attacker could use variations in the signing
1900 algorithm to recover the private key.
1902 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1907 * Timing vulnerability in ECDSA signature generation
1909 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1910 timing side channel attack. An attacker could use variations in the signing
1911 algorithm to recover the private key.
1913 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1918 * Add coordinate blinding for EC_POINT and implement projective
1919 coordinate blinding for generic prime curves as a countermeasure to
1920 chosen point SCA attacks.
1922 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1924 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018] ###
1926 * Client DoS due to large DH parameter
1928 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1929 malicious server can send a very large prime value to the client. This will
1930 cause the client to spend an unreasonably long period of time generating a
1931 key for this prime resulting in a hang until the client has finished. This
1932 could be exploited in a Denial Of Service attack.
1934 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1939 * Cache timing vulnerability in RSA Key Generation
1941 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1942 a cache timing side channel attack. An attacker with sufficient access to
1943 mount cache timing attacks during the RSA key generation process could
1944 recover the private key.
1946 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1947 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1952 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1953 parameter is no longer accepted, as it leads to a corrupt table. NULL
1954 pem_str is reserved for alias entries only.
1958 * Revert blinding in ECDSA sign and instead make problematic addition
1959 length-invariant. Switch even to fixed-length Montgomery multiplication.
1963 * Change generating and checking of primes so that the error rate of not
1964 being prime depends on the intended use based on the size of the input.
1965 For larger primes this will result in more rounds of Miller-Rabin.
1966 The maximal error rate for primes with more than 1080 bits is lowered
1969 *Kurt Roeckx, Annie Yousar*
1971 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1975 * Add blinding to ECDSA and DSA signatures to protect against side channel
1976 attacks discovered by Keegan Ryan (NCC Group).
1980 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1981 now allow empty (zero character) pass phrases.
1985 * Certificate time validation (X509_cmp_time) enforces stricter
1986 compliance with RFC 5280. Fractional seconds and timezone offsets
1987 are no longer allowed.
1991 * Fixed a text canonicalisation bug in CMS
1993 Where a CMS detached signature is used with text content the text goes
1994 through a canonicalisation process first prior to signing or verifying a
1995 signature. This process strips trailing space at the end of lines, converts
1996 line terminators to CRLF and removes additional trailing line terminators
1997 at the end of a file. A bug in the canonicalisation process meant that
1998 some characters, such as form-feed, were incorrectly treated as whitespace
1999 and removed. This is contrary to the specification (RFC5485). This fix
2000 could mean that detached text data signed with an earlier version of
2001 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2002 signed with a fixed OpenSSL may fail to verify with an earlier version of
2003 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2004 and use the "-binary" flag (for the "cms" command line application) or set
2005 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2009 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018] ###
2011 * Constructed ASN.1 types with a recursive definition could exceed the stack
2013 Constructed ASN.1 types with a recursive definition (such as can be found
2014 in PKCS7) could eventually exceed the stack given malicious input with
2015 excessive recursion. This could result in a Denial Of Service attack. There
2016 are no such structures used within SSL/TLS that come from untrusted sources
2017 so this is considered safe.
2019 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2025 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2027 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2028 effectively reduced to only comparing the least significant bit of each
2029 byte. This allows an attacker to forge messages that would be considered as
2030 authenticated in an amount of tries lower than that guaranteed by the
2031 security claims of the scheme. The module can only be compiled by the
2032 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2034 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2040 * Add a build target 'build_all_generated', to build all generated files
2041 and only that. This can be used to prepare everything that requires
2042 things like perl for a system that lacks perl and then move everything
2043 to that system and do the rest of the build there.
2047 * Backport SSL_OP_NO_RENGOTIATION
2049 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2050 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2051 changes this is no longer possible in 1.1.0. Therefore the new
2052 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2053 1.1.0 to provide equivalent functionality.
2055 Note that if an application built against 1.1.0h headers (or above) is run
2056 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2057 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2061 * Removed the OS390-Unix config target. It relied on a script that doesn't
2066 * rsaz_1024_mul_avx2 overflow bug on x86_64
2068 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2069 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2070 Analysis suggests that attacks against RSA and DSA as a result of this
2071 defect would be very difficult to perform and are not believed likely.
2072 Attacks against DH1024 are considered just feasible, because most of the
2073 work necessary to deduce information about a private key may be performed
2074 offline. The amount of resources required for such an attack would be
2075 significant. However, for an attack on TLS to be meaningful, the server
2076 would have to share the DH1024 private key among multiple clients, which is
2077 no longer an option since CVE-2016-0701.
2079 This only affects processors that support the AVX2 but not ADX extensions
2080 like Intel Haswell (4th generation).
2082 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2083 was originally found via the OSS-Fuzz project.
2088 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017] ###
2090 * bn_sqrx8x_internal carry bug on x86_64
2092 There is a carry propagating bug in the x86_64 Montgomery squaring
2093 procedure. No EC algorithms are affected. Analysis suggests that attacks
2094 against RSA and DSA as a result of this defect would be very difficult to
2095 perform and are not believed likely. Attacks against DH are considered just
2096 feasible (although very difficult) because most of the work necessary to
2097 deduce information about a private key may be performed offline. The amount
2098 of resources required for such an attack would be very significant and
2099 likely only accessible to a limited number of attackers. An attacker would
2100 additionally need online access to an unpatched system using the target
2101 private key in a scenario with persistent DH parameters and a private
2102 key that is shared between multiple clients.
2104 This only affects processors that support the BMI1, BMI2 and ADX extensions
2105 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2107 This issue was reported to OpenSSL by the OSS-Fuzz project.
2112 * Malformed X.509 IPAddressFamily could cause OOB read
2114 If an X.509 certificate has a malformed IPAddressFamily extension,
2115 OpenSSL could do a one-byte buffer overread. The most likely result
2116 would be an erroneous display of the certificate in text format.
2118 This issue was reported to OpenSSL by the OSS-Fuzz project.
2123 ### Changes between 1.1.0e and 1.1.0f [25 May 2017] ###
2125 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2126 platform rather than 'mingw'.
2130 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2131 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2132 which is the minimum version we support.
2136 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017] ###
2138 * Encrypt-Then-Mac renegotiation crash
2140 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2141 negotiated where it was not in the original handshake (or vice-versa) then
2142 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2143 and servers are affected.
2145 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2150 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017] ###
2152 * Truncated packet could crash via OOB read
2154 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2155 cipher is being used, then a truncated packet can cause that host to
2156 perform an out-of-bounds read, usually resulting in a crash.
2158 This issue was reported to OpenSSL by Robert Święcki of Google.
2163 * Bad (EC)DHE parameters cause a client crash
2165 If a malicious server supplies bad parameters for a DHE or ECDHE key
2166 exchange then this can result in the client attempting to dereference a
2167 NULL pointer leading to a client crash. This could be exploited in a Denial
2170 This issue was reported to OpenSSL by Guido Vranken.
2175 * BN_mod_exp may produce incorrect results on x86_64
2177 There is a carry propagating bug in the x86_64 Montgomery squaring
2178 procedure. No EC algorithms are affected. Analysis suggests that attacks
2179 against RSA and DSA as a result of this defect would be very difficult to
2180 perform and are not believed likely. Attacks against DH are considered just
2181 feasible (although very difficult) because most of the work necessary to
2182 deduce information about a private key may be performed offline. The amount
2183 of resources required for such an attack would be very significant and
2184 likely only accessible to a limited number of attackers. An attacker would
2185 additionally need online access to an unpatched system using the target
2186 private key in a scenario with persistent DH parameters and a private
2187 key that is shared between multiple clients. For example this can occur by
2188 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2189 similar to CVE-2015-3193 but must be treated as a separate problem.
2191 This issue was reported to OpenSSL by the OSS-Fuzz project.
2196 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016] ###
2198 * ChaCha20/Poly1305 heap-buffer-overflow
2200 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
2201 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2202 crash. This issue is not considered to be exploitable beyond a DoS.
2204 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2209 * CMS Null dereference
2211 Applications parsing invalid CMS structures can crash with a NULL pointer
2212 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2213 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2214 structure callback if an attempt is made to free certain invalid encodings.
2215 Only CHOICE structures using a callback which do not handle NULL value are
2218 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2223 * Montgomery multiplication may produce incorrect results
2225 There is a carry propagating bug in the Broadwell-specific Montgomery
2226 multiplication procedure that handles input lengths divisible by, but
2227 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2228 and DH private keys are impossible. This is because the subroutine in
2229 question is not used in operations with the private key itself and an input
2230 of the attacker's direct choice. Otherwise the bug can manifest itself as
2231 transient authentication and key negotiation failures or reproducible
2232 erroneous outcome of public-key operations with specially crafted input.
2233 Among EC algorithms only Brainpool P-512 curves are affected and one
2234 presumably can attack ECDH key negotiation. Impact was not analyzed in
2235 detail, because pre-requisites for attack are considered unlikely. Namely
2236 multiple clients have to choose the curve in question and the server has to
2237 share the private key among them, neither of which is default behaviour.
2238 Even then only clients that chose the curve will be affected.
2240 This issue was publicly reported as transient failures and was not
2241 initially recognized as a security issue. Thanks to Richard Morgan for
2242 providing reproducible case.
2247 * Removed automatic addition of RPATH in shared libraries and executables,
2248 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2252 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016] ###
2254 * Fix Use After Free for large message sizes
2256 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2257 message larger than approx 16k is received then the underlying buffer to
2258 store the incoming message is reallocated and moved. Unfortunately a
2259 dangling pointer to the old location is left which results in an attempt to
2260 write to the previously freed location. This is likely to result in a
2261 crash, however it could potentially lead to execution of arbitrary code.
2263 This issue only affects OpenSSL 1.1.0a.
2265 This issue was reported to OpenSSL by Robert Święcki.
2270 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016] ###
2272 * OCSP Status Request extension unbounded memory growth
2274 A malicious client can send an excessively large OCSP Status Request
2275 extension. If that client continually requests renegotiation, sending a
2276 large OCSP Status Request extension each time, then there will be unbounded
2277 memory growth on the server. This will eventually lead to a Denial Of
2278 Service attack through memory exhaustion. Servers with a default
2279 configuration are vulnerable even if they do not support OCSP. Builds using
2280 the "no-ocsp" build time option are not affected.
2282 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2287 * SSL_peek() hang on empty record
2289 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2290 sends an empty record. This could be exploited by a malicious peer in a
2291 Denial Of Service attack.
2293 This issue was reported to OpenSSL by Alex Gaynor.
2298 * Excessive allocation of memory in tls_get_message_header() and
2299 dtls1_preprocess_fragment()
2301 A (D)TLS message includes 3 bytes for its length in the header for the
2302 message. This would allow for messages up to 16Mb in length. Messages of
2303 this length are excessive and OpenSSL includes a check to ensure that a
2304 peer is sending reasonably sized messages in order to avoid too much memory
2305 being consumed to service a connection. A flaw in the logic of version
2306 1.1.0 means that memory for the message is allocated too early, prior to
2307 the excessive message length check. Due to way memory is allocated in
2308 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2309 to service a connection. This could lead to a Denial of Service through
2310 memory exhaustion. However, the excessive message length check still takes
2311 place, and this would cause the connection to immediately fail. Assuming
2312 that the application calls SSL_free() on the failed connection in a timely
2313 manner then the 21Mb of allocated memory will then be immediately freed
2314 again. Therefore the excessive memory allocation will be transitory in
2315 nature. This then means that there is only a security impact if:
2317 1) The application does not call SSL_free() in a timely manner in the event
2318 that the connection fails
2320 2) The application is working in a constrained environment where there is
2321 very little free memory
2323 3) The attacker initiates multiple connection attempts such that there are
2324 multiple connections in a state where memory has been allocated for the
2325 connection; SSL_free() has not yet been called; and there is insufficient
2326 memory to service the multiple requests.
2328 Except in the instance of (1) above any Denial Of Service is likely to be
2329 transitory because as soon as the connection fails the memory is
2330 subsequently freed again in the SSL_free() call. However there is an
2331 increased risk during this period of application crashes due to the lack of
2332 memory - which would then mean a more serious Denial of Service.
2334 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2335 (CVE-2016-6307 and CVE-2016-6308)
2339 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2340 had to be removed. Primary reason is that vendor assembler can't
2341 assemble our modules with -KPIC flag. As result it, assembly
2342 support, was not even available as option. But its lack means
2343 lack of side-channel resistant code, which is incompatible with
2344 security by todays standards. Fortunately gcc is readily available
2345 prepackaged option, which we firmly point at...
2349 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016] ###
2351 * Windows command-line tool supports UTF-8 opt-in option for arguments
2352 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2353 (to any value) allows Windows user to access PKCS#12 file generated
2354 with Windows CryptoAPI and protected with non-ASCII password, as well
2355 as files generated under UTF-8 locale on Linux also protected with
2360 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2361 have been disabled by default and removed from DEFAULT, just like RC4.
2362 See the RC4 item below to re-enable both.
2366 * The method for finding the storage location for the Windows RAND seed file
2367 has changed. First we check %RANDFILE%. If that is not set then we check
2368 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2369 all else fails we fall back to C:\.
2373 * The EVP_EncryptUpdate() function has had its return type changed from void
2374 to int. A return of 0 indicates and error while a return of 1 indicates
2379 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2380 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2381 off the constant time implementation for RSA, DSA and DH have been made
2382 no-ops and deprecated.
2386 * Windows RAND implementation was simplified to only get entropy by
2387 calling CryptGenRandom(). Various other RAND-related tickets
2390 *Joseph Wylie Yandle, Rich Salz*
2392 * The stack and lhash API's were renamed to start with OPENSSL_SK_
2393 and OPENSSL_LH_, respectively. The old names are available
2394 with API compatibility. They new names are now completely documented.
2398 * Unify TYPE_up_ref(obj) methods signature.
2399 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2400 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2401 int (instead of void) like all others TYPE_up_ref() methods.
2402 So now these methods also check the return value of CRYPTO_atomic_add(),
2403 and the validity of object reference counter.
2405 *fdasilvayy@gmail.com*
2407 * With Windows Visual Studio builds, the .pdb files are installed
2408 alongside the installed libraries and executables. For a static
2409 library installation, ossl_static.pdb is the associate compiler
2410 generated .pdb file to be used when linking programs.
2414 * Remove openssl.spec. Packaging files belong with the packagers.
2418 * Automatic Darwin/OSX configuration has had a refresh, it will now
2419 recognise x86_64 architectures automatically. You can still decide
2420 to build for a different bitness with the environment variable
2421 KERNEL_BITS (can be 32 or 64), for example:
2423 KERNEL_BITS=32 ./config
2427 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2428 256 bit AES and HMAC with SHA256.
2432 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2436 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2440 * To enable users to have their own config files and build file templates,
2441 Configure looks in the directory indicated by the environment variable
2442 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2443 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2444 name and is used as is.
2448 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2449 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2450 X509_CERT_FILE_CTX was removed.
2454 * "shared" builds are now the default. To create only static libraries use
2455 the "no-shared" Configure option.
2459 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2460 All of these option have not worked for some while and are fundamental
2465 * Make various cleanup routines no-ops and mark them as deprecated. Most
2466 global cleanup functions are no longer required because they are handled
2467 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2468 Explicitly de-initing can cause problems (e.g. where a library that uses
2469 OpenSSL de-inits, but an application is still using it). The affected
2470 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2471 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2472 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2473 COMP_zlib_cleanup().
2477 * --strict-warnings no longer enables runtime debugging options
2478 such as REF_DEBUG. Instead, debug options are automatically
2479 enabled with '--debug' builds.
2481 *Andy Polyakov, Emilia Käsper*
2483 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2484 have been moved out of the public header files. New functions for managing
2485 these have been added.
2489 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2490 objects have been moved out of the public header files. New
2491 functions for managing these have been added.
2495 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2496 have been moved out of the public header files. New functions for managing
2497 these have been added.
2501 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2502 moved out of the public header files. New functions for managing these
2507 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2511 * Removed the mk1mf build scripts.
2515 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2516 it is always safe to #include a header now.
2520 * Removed the aged BC-32 config and all its supporting scripts
2524 * Removed support for Ultrix, Netware, and OS/2.
2528 * Add support for HKDF.
2530 *Alessandro Ghedini*
2532 * Add support for blake2b and blake2s
2536 * Added support for "pipelining". Ciphers that have the
2537 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2538 encryptions/decryptions simultaneously. There are currently no built-in
2539 ciphers with this property but the expectation is that engines will be able
2540 to offer it to significantly improve throughput. Support has been extended
2541 into libssl so that multiple records for a single connection can be
2542 processed in one go (for >=TLS 1.1).
2546 * Added the AFALG engine. This is an async capable engine which is able to
2547 offload work to the Linux kernel. In this initial version it only supports
2548 AES128-CBC. The kernel must be version 4.1.0 or greater.
2552 * OpenSSL now uses a new threading API. It is no longer necessary to
2553 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2554 are two supported threading models: pthreads and windows threads. It is
2555 also possible to configure OpenSSL at compile time for "no-threads". The
2556 old threading API should no longer be used. The functions have been
2557 replaced with "no-op" compatibility macros.
2559 *Alessandro Ghedini, Matt Caswell*
2561 * Modify behavior of ALPN to invoke callback after SNI/servername
2562 callback, such that updates to the SSL_CTX affect ALPN.
2566 * Add SSL_CIPHER queries for authentication and key-exchange.
2570 * Changes to the DEFAULT cipherlist:
2571 - Prefer (EC)DHE handshakes over plain RSA.
2572 - Prefer AEAD ciphers over legacy ciphers.
2573 - Prefer ECDSA over RSA when both certificates are available.
2574 - Prefer TLSv1.2 ciphers/PRF.
2575 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2580 * Change the ECC default curve list to be this, in order: x25519,
2581 secp256r1, secp521r1, secp384r1.
2585 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2586 disabled by default. They can be re-enabled using the
2587 enable-weak-ssl-ciphers option to Configure.
2591 * If the server has ALPN configured, but supports no protocols that the
2592 client advertises, send a fatal "no_application_protocol" alert.
2593 This behaviour is SHALL in RFC 7301, though it isn't universally
2594 implemented by other servers.
2598 * Add X25519 support.
2599 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2600 for public and private key encoding using the format documented in
2601 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2602 key generation and key derivation.
2604 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2609 * Deprecate SRP_VBASE_get_by_user.
2610 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2611 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2612 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2613 seed, even if the seed is configured.
2615 Users should use SRP_VBASE_get1_by_user instead. Note that in
2616 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2617 also that even though configuring the SRP seed attempts to hide
2618 invalid usernames by continuing the handshake with fake
2619 credentials, this behaviour is not constant time and no strong
2620 guarantees are made that the handshake is indistinguishable from
2621 that of a valid user.
2625 * Configuration change; it's now possible to build dynamic engines
2626 without having to build shared libraries and vice versa. This
2627 only applies to the engines in engines/, those in crypto/engine/
2628 will always be built into libcrypto (i.e. "static").
2630 Building dynamic engines is enabled by default; to disable, use
2631 the configuration option "disable-dynamic-engine".
2633 The only requirements for building dynamic engines are the
2634 presence of the DSO module and building with position independent
2635 code, so they will also automatically be disabled if configuring
2636 with "disable-dso" or "disable-pic".
2638 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2639 are also taken away from openssl/opensslconf.h, as they are
2644 * Configuration change; if there is a known flag to compile
2645 position independent code, it will always be applied on the
2646 libcrypto and libssl object files, and never on the application
2647 object files. This means other libraries that use routines from
2648 libcrypto / libssl can be made into shared libraries regardless
2649 of how OpenSSL was configured.
2651 If this isn't desirable, the configuration options "disable-pic"
2652 or "no-pic" can be used to disable the use of PIC. This will
2653 also disable building shared libraries and dynamic engines.
2657 * Removed JPAKE code. It was experimental and has no wide use.
2661 * The INSTALL_PREFIX Makefile variable has been renamed to
2662 DESTDIR. That makes for less confusion on what this variable
2663 is for. Also, the configuration option --install_prefix is
2668 * Heartbeat for TLS has been removed and is disabled by default
2669 for DTLS; configure with enable-heartbeats. Code that uses the
2670 old #define's might need to be updated.
2672 *Emilia Käsper, Rich Salz*
2674 * Rename REF_CHECK to REF_DEBUG.
2678 * New "unified" build system
2680 The "unified" build system is aimed to be a common system for all
2681 platforms we support. With it comes new support for VMS.
2683 This system builds supports building in a different directory tree
2684 than the source tree. It produces one Makefile (for unix family
2685 or lookalikes), or one descrip.mms (for VMS).
2687 The source of information to make the Makefile / descrip.mms is
2688 small files called 'build.info', holding the necessary
2689 information for each directory with source to compile, and a
2690 template in Configurations, like unix-Makefile.tmpl or
2693 With this change, the library names were also renamed on Windows
2694 and on VMS. They now have names that are closer to the standard
2695 on Unix, and include the major version number, and in certain
2696 cases, the architecture they are built for. See "Notes on shared
2697 libraries" in INSTALL.
2699 We rely heavily on the perl module Text::Template.
2703 * Added support for auto-initialisation and de-initialisation of the library.
2704 OpenSSL no longer requires explicit init or deinit routines to be called,
2705 except in certain circumstances. See the OPENSSL_init_crypto() and
2706 OPENSSL_init_ssl() man pages for further information.
2710 * The arguments to the DTLSv1_listen function have changed. Specifically the
2711 "peer" argument is now expected to be a BIO_ADDR object.
2713 * Rewrite of BIO networking library. The BIO library lacked consistent
2714 support of IPv6, and adding it required some more extensive
2715 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2716 which hold all types of addresses and chains of address information.
2717 It also introduces a new API, with functions like BIO_socket,
2718 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2719 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2720 have been adapted accordingly.
2724 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2729 * CRIME protection: disable compression by default, even if OpenSSL is
2730 compiled with zlib enabled. Applications can still enable compression
2731 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2732 using the SSL_CONF library to configure compression.
2736 * The signature of the session callback configured with
2737 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2738 was explicitly marked as 'const unsigned char*' instead of
2743 * Always DPURIFY. Remove the use of uninitialized memory in the
2744 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2748 * Removed many obsolete configuration items, including
2749 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2750 MD2_CHAR, MD2_INT, MD2_LONG
2752 IDEA_SHORT, IDEA_LONG
2753 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2755 *Rich Salz, with advice from Andy Polyakov*
2757 * Many BN internals have been moved to an internal header file.
2759 *Rich Salz with help from Andy Polyakov*
2761 * Configuration and writing out the results from it has changed.
2762 Files such as Makefile include/openssl/opensslconf.h and are now
2763 produced through general templates, such as Makefile.in and
2764 crypto/opensslconf.h.in and some help from the perl module
2767 Also, the center of configuration information is no longer
2768 Makefile. Instead, Configure produces a perl module in
2769 configdata.pm which holds most of the config data (in the hash
2770 table %config), the target data that comes from the target
2771 configuration in one of the Configurations/*.conf files (in
2776 * To clarify their intended purposes, the Configure options
2777 --prefix and --openssldir change their semantics, and become more
2778 straightforward and less interdependent.
2780 --prefix shall be used exclusively to give the location INSTALLTOP
2781 where programs, scripts, libraries, include files and manuals are
2782 going to be installed. The default is now /usr/local.
2784 --openssldir shall be used exclusively to give the default
2785 location OPENSSLDIR where certificates, private keys, CRLs are
2786 managed. This is also where the default openssl.cnf gets
2788 If the directory given with this option is a relative path, the
2789 values of both the --prefix value and the --openssldir value will
2790 be combined to become OPENSSLDIR.
2791 The default for --openssldir is INSTALLTOP/ssl.
2793 Anyone who uses --openssldir to specify where OpenSSL is to be
2794 installed MUST change to use --prefix instead.
2798 * The GOST engine was out of date and therefore it has been removed. An up
2799 to date GOST engine is now being maintained in an external repository.
2800 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2801 support for GOST ciphersuites (these are only activated if a GOST engine
2806 * EGD is no longer supported by default; use enable-egd when
2809 *Ben Kaduk and Rich Salz*
2811 * The distribution now has Makefile.in files, which are used to
2812 create Makefile's when Configure is run. *Configure must be run
2813 before trying to build now.*
2817 * The return value for SSL_CIPHER_description() for error conditions
2822 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2824 Obtaining and performing DNSSEC validation of TLSA records is
2825 the application's responsibility. The application provides
2826 the TLSA records of its choice to OpenSSL, and these are then
2827 used to authenticate the peer.
2829 The TLSA records need not even come from DNS. They can, for
2830 example, be used to implement local end-entity certificate or
2831 trust-anchor "pinning", where the "pin" data takes the form
2832 of TLSA records, which can augment or replace verification
2833 based on the usual WebPKI public certification authorities.
2837 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2838 continues to support deprecated interfaces in default builds.
2839 However, applications are strongly advised to compile their
2840 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2841 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2842 or the 1.1.0 releases.
2844 In environments in which all applications have been ported to
2845 not use any deprecated interfaces OpenSSL's Configure script
2846 should be used with the --api=1.1.0 option to entirely remove
2847 support for the deprecated features from the library and
2848 unconditionally disable them in the installed headers.
2849 Essentially the same effect can be achieved with the "no-deprecated"
2850 argument to Configure, except that this will always restrict
2851 the build to just the latest API, rather than a fixed API
2854 As applications are ported to future revisions of the API,
2855 they should update their compile-time OPENSSL_API_COMPAT define
2856 accordingly, but in most cases should be able to continue to
2857 compile with later releases.
2859 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2860 0x10000000L and 0x00908000L, respectively. However those
2861 versions did not support the OPENSSL_API_COMPAT feature, and
2862 so applications are not typically tested for explicit support
2863 of just the undeprecated features of either release.
2867 * Add support for setting the minimum and maximum supported protocol.
2868 It can bet set via the SSL_set_min_proto_version() and
2869 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2870 MaxProtocol. It's recommended to use the new APIs to disable
2871 protocols instead of disabling individual protocols using
2872 SSL_set_options() or SSL_CONF's Protocol. This change also
2873 removes support for disabling TLS 1.2 in the OpenSSL TLS
2874 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2878 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2882 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2883 and integrates ECDSA and ECDH functionality into EC. Implementations can
2884 now redirect key generation and no longer need to convert to or from
2887 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2888 include the ec.h header file instead.
2892 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2893 ciphers who are no longer supported and drops support the ephemeral RSA key
2894 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2898 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2899 opaque. For HMAC_CTX, the following constructors and destructors
2902 HMAC_CTX *HMAC_CTX_new(void);
2903 void HMAC_CTX_free(HMAC_CTX *ctx);
2905 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2906 destroy such methods has been added. See EVP_MD_meth_new(3) and
2907 EVP_CIPHER_meth_new(3) for documentation.
2910 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2911 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2912 EVP_MD_CTX_reset() should be called instead to reinitialise
2913 an already created structure.
2914 2) For consistency with the majority of our object creators and
2915 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2916 EVP_MD_CTX_(new|free). The old names are retained as macros
2917 for deprecated builds.
2921 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2922 cryptographic operations to be performed asynchronously as long as an
2923 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2924 further details. Libssl has also had this capability integrated with the
2925 introduction of the new mode SSL_MODE_ASYNC and associated error
2926 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2927 pages. This work was developed in partnership with Intel Corp.
2931 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2932 always enabled now. If you want to disable the support you should
2933 exclude it using the list of supported ciphers. This also means that the
2934 "-no_ecdhe" option has been removed from s_server.
2938 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2939 SSL_{CTX_}set1_curves() which can set a list.
2943 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2944 curve you want to support using SSL_{CTX_}set1_curves().
2948 * State machine rewrite. The state machine code has been significantly
2949 refactored in order to remove much duplication of code and solve issues
2950 with the old code (see ssl/statem/README for further details). This change
2951 does have some associated API changes. Notably the SSL_state() function
2952 has been removed and replaced by SSL_get_state which now returns an
2953 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2954 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2959 * All instances of the string "ssleay" in the public API were replaced
2960 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2961 Some error codes related to internal RSA_eay API's were renamed.
2965 * The demo files in crypto/threads were moved to demo/threads.
2969 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2972 *Matt Caswell, Rich Salz*
2974 * New ASN.1 embed macro.
2976 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2977 structure is not allocated: it is part of the parent. That is instead of
2985 This reduces memory fragmentation and make it impossible to accidentally
2986 set a mandatory field to NULL.
2988 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2989 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2990 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2995 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2999 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3000 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3001 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3002 DES and RC4 ciphersuites.
3006 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3007 This changes the decoding behaviour for some invalid messages,
3008 though the change is mostly in the more lenient direction, and
3009 legacy behaviour is preserved as much as possible.
3013 * Fix no-stdio build.
3014 * David Woodhouse <David.Woodhouse@intel.com> and also
3015 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
3017 * New testing framework
3018 The testing framework has been largely rewritten and is now using
3019 perl and the perl modules Test::Harness and an extended variant of
3020 Test::More called OpenSSL::Test to do its work. All test scripts in
3021 test/ have been rewritten into test recipes, and all direct calls to
3022 executables in test/Makefile have become individual recipes using the
3023 simplified testing OpenSSL::Test::Simple.
3025 For documentation on our testing modules, do:
3027 perldoc test/testlib/OpenSSL/Test/Simple.pm
3028 perldoc test/testlib/OpenSSL/Test.pm
3032 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3033 are used; the latter aborts on memory leaks (usually checked on exit).
3034 Some undocumented "set malloc, etc., hooks" functions were removed
3035 and others were changed. All are now documented.
3039 * In DSA_generate_parameters_ex, if the provided seed is too short,
3042 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3044 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3045 from RFC4279, RFC4785, RFC5487, RFC5489.
3047 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3048 original RSA_PSK patch.
3052 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3053 era flag was never set throughout the codebase (only read). Also removed
3054 SSL3_FLAGS_POP_BUFFER which was only used if
3055 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3059 * Changed the default name options in the "ca", "crl", "req" and "x509"
3060 to be "oneline" instead of "compat".
3064 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3065 not aware of clients that still exhibit this bug, and the workaround
3066 hasn't been working properly for a while.
3070 * The return type of BIO_number_read() and BIO_number_written() as well as
3071 the corresponding num_read and num_write members in the BIO structure has
3072 changed from unsigned long to uint64_t. On platforms where an unsigned
3073 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3078 * Given the pervasive nature of TLS extensions it is inadvisable to run
3079 OpenSSL without support for them. It also means that maintaining
3080 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3081 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3085 * Removed support for the two export grade static DH ciphersuites
3086 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3087 were newly added (along with a number of other static DH ciphersuites) to
3088 1.0.2. However the two export ones have *never* worked since they were
3089 introduced. It seems strange in any case to be adding new export
3090 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3094 * Version negotiation has been rewritten. In particular SSLv23_method(),
3095 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3096 and turned into macros which simply call the new preferred function names
3097 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3098 should use the new names instead. Also as part of this change the ssl23.h
3099 header file has been removed.
3103 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3104 code and the associated standard is no longer considered fit-for-purpose.
3108 * RT2547 was closed. When generating a private key, try to make the
3109 output file readable only by the owner. This behavior change might
3110 be noticeable when interacting with other software.
3112 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3117 * Added HTTP GET support to the ocsp command.
3121 * Changed default digest for the dgst and enc commands from MD5 to
3126 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3130 * Added support for TLS extended master secret from
3131 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3132 initial patch which was a great help during development.
3136 * All libssl internal structures have been removed from the public header
3137 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3138 now redundant). Users should not attempt to access internal structures
3139 directly. Instead they should use the provided API functions.
3143 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3144 Access to deprecated functions can be re-enabled by running config with
3145 "enable-deprecated". In addition applications wishing to use deprecated
3146 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3147 will, by default, disable some transitive includes that previously existed
3148 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3152 * Added support for OCB mode. OpenSSL has been granted a patent license
3153 compatible with the OpenSSL license for use of OCB. Details are available
3154 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
3155 for OCB can be removed by calling config with no-ocb.
3159 * SSLv2 support has been removed. It still supports receiving a SSLv2
3160 compatible client hello.
3164 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3165 done while fixing the error code for the key-too-small case.
3167 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3169 * CA.sh has been removed; use CA.pl instead.
3173 * Removed old DES API.
3177 * Remove various unsupported platforms:
3183 Sinix/ReliantUNIX RM400
3188 16-bit platforms such as WIN16
3192 * Clean up OPENSSL_NO_xxx #define's
3193 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3194 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3195 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3196 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3197 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3198 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3199 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3200 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3201 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3202 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3206 * Cleaned up dead code
3207 Remove all but one '#ifdef undef' which is to be looked at.
3211 * Clean up calling of xxx_free routines.
3212 Just like free(), fix most of the xxx_free routines to accept
3213 NULL. Remove the non-null checks from callers. Save much code.
3217 * Add secure heap for storage of private keys (when possible).
3218 Add BIO_s_secmem(), CBIGNUM, etc.
3219 Contributed by Akamai Technologies under our Corporate CLA.
3223 * Experimental support for a new, fast, unbiased prime candidate generator,
3224 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3226 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3228 * New output format NSS in the sess_id command line tool. This allows
3229 exporting the session id and the master key in NSS keylog format.
3231 *Martin Kaiser <martin@kaiser.cx>*
3233 * Harmonize version and its documentation. -f flag is used to display
3236 *mancha <mancha1@zoho.com>*
3238 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3239 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3241 *mancha <mancha1@zoho.com>*
3243 * Fix some double frees. These are not thought to be exploitable.
3245 *mancha <mancha1@zoho.com>*
3247 * A missing bounds check in the handling of the TLS heartbeat extension
3248 can be used to reveal up to 64k of memory to a connected client or
3251 Thanks for Neel Mehta of Google Security for discovering this bug and to
3252 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3253 preparing the fix [CVE-2014-0160][]
3255 *Adam Langley, Bodo Moeller*
3257 * Fix for the attack described in the paper "Recovering OpenSSL
3258 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3259 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3260 http://eprint.iacr.org/2014/140
3262 Thanks to Yuval Yarom and Naomi Benger for discovering this
3263 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3265 *Yuval Yarom and Naomi Benger*
3267 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3268 this fixes a limitation in previous versions of OpenSSL.
3272 * Experimental encrypt-then-mac support.
3274 Experimental support for encrypt then mac from
3275 draft-gutmann-tls-encrypt-then-mac-02.txt
3277 To enable it set the appropriate extension number (0x42 for the test
3278 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3280 For non-compliant peers (i.e. just about everything) this should have no
3283 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3288 * Add EVP support for key wrapping algorithms, to avoid problems with
3289 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3290 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3291 algorithms and include tests cases.
3295 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3300 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3301 MGF1 digest and OAEP label.
3305 * Make openssl verify return errors.
3307 *Chris Palmer <palmer@google.com> and Ben Laurie*
3309 * New function ASN1_TIME_diff to calculate the difference between two
3310 ASN1_TIME structures or one structure and the current time.
3314 * Update fips_test_suite to support multiple command line options. New
3315 test to induce all self test errors in sequence and check expected
3320 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3321 sign or verify all in one operation.
3325 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3326 test programs and fips_test_suite. Includes functionality to parse
3327 the minimal script output of fipsalgest.pl directly.
3331 * Add authorisation parameter to FIPS_module_mode_set().
3335 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3339 * Use separate DRBG fields for internal and external flags. New function
3340 FIPS_drbg_health_check() to perform on demand health checking. Add
3341 generation tests to fips_test_suite with reduced health check interval to
3342 demonstrate periodic health checking. Add "nodh" option to
3343 fips_test_suite to skip very slow DH test.
3347 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3352 * More extensive health check for DRBG checking many more failure modes.
3353 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3354 combination: call this in fips_test_suite.
3358 * Add support for canonical generation of DSA parameter 'g'. See
3361 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3362 POST to handle HMAC cases.
3366 * Add functions FIPS_module_version() and FIPS_module_version_text()
3367 to return numerical and string versions of the FIPS module number.
3371 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3372 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3373 outside the validated module in the FIPS capable OpenSSL.
3377 * Minor change to DRBG entropy callback semantics. In some cases
3378 there is no multiple of the block length between min_len and
3379 max_len. Allow the callback to return more than max_len bytes
3380 of entropy but discard any extra: it is the callback's responsibility
3381 to ensure that the extra data discarded does not impact the
3382 requested amount of entropy.
3386 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3387 information in FIPS186-3, SP800-57 and SP800-131A.
3391 * CCM support via EVP. Interface is very similar to GCM case except we
3392 must supply all data in one chunk (i.e. no update, final) and the
3393 message length must be supplied if AAD is used. Add algorithm test
3398 * Initial version of POST overhaul. Add POST callback to allow the status
3399 of POST to be monitored and/or failures induced. Modify fips_test_suite
3400 to use callback. Always run all selftests even if one fails.
3404 * XTS support including algorithm test driver in the fips_gcmtest program.
3405 Note: this does increase the maximum key length from 32 to 64 bytes but
3406 there should be no binary compatibility issues as existing applications
3407 will never use XTS mode.
3411 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3412 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3413 performs algorithm blocking for unapproved PRNG types. Also do not
3414 set PRNG type in FIPS_mode_set(): leave this to the application.
3415 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3416 the standard OpenSSL PRNG: set additional data to a date time vector.
3420 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3421 This shouldn't present any incompatibility problems because applications
3422 shouldn't be using these directly and any that are will need to rethink
3423 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3427 * Extensive self tests and health checking required by SP800-90 DRBG.
3428 Remove strength parameter from FIPS_drbg_instantiate and always
3429 instantiate at maximum supported strength.
3433 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3437 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3441 * New function DH_compute_key_padded() to compute a DH key and pad with
3442 leading zeroes if needed: this complies with SP800-56A et al.
3446 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3447 anything, incomplete, subject to change and largely untested at present.
3451 * Modify fipscanisteronly build option to only build the necessary object
3452 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3456 * Add experimental option FIPSSYMS to give all symbols in
3457 fipscanister.o and FIPS or fips prefix. This will avoid
3458 conflicts with future versions of OpenSSL. Add perl script
3459 util/fipsas.pl to preprocess assembly language source files
3460 and rename any affected symbols.
3464 * Add selftest checks and algorithm block of non-fips algorithms in
3465 FIPS mode. Remove DES2 from selftests.
3469 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3470 return internal method without any ENGINE dependencies. Add new
3471 tiny fips sign and verify functions.
3475 * New build option no-ec2m to disable characteristic 2 code.
3479 * New build option "fipscanisteronly". This only builds fipscanister.o
3480 and (currently) associated fips utilities. Uses the file Makefile.fips
3481 instead of Makefile.org as the prototype.
3485 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3486 Update fips_gcmtest to use IV generator.
3490 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3491 setting output buffer to NULL. The *Final function must be
3492 called although it will not retrieve any additional data. The tag
3493 can be set or retrieved with a ctrl. The IV length is by default 12
3494 bytes (96 bits) but can be set to an alternative value. If the IV
3495 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3500 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3501 underlying do_cipher function handles all cipher semantics itself
3502 including padding and finalisation. This is useful if (for example)
3503 an ENGINE cipher handles block padding itself. The behaviour of
3504 do_cipher is subtly changed if this flag is set: the return value
3505 is the number of characters written to the output buffer (zero is
3506 no longer an error code) or a negative error code. Also if the
3507 input buffer is NULL and length 0 finalisation should be performed.
3511 * If a candidate issuer certificate is already part of the constructed
3512 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3516 * Improve forward-security support: add functions
3518 void SSL_CTX_set_not_resumable_session_callback(
3519 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3520 void SSL_set_not_resumable_session_callback(
3521 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3523 for use by SSL/TLS servers; the callback function will be called whenever a
3524 new session is created, and gets to decide whether the session may be
3525 cached to make it resumable (return 0) or not (return 1). (As by the
3526 SSL/TLS protocol specifications, the session_id sent by the server will be
3527 empty to indicate that the session is not resumable; also, the server will
3528 not generate RFC 4507 (RFC 5077) session tickets.)
3530 A simple reasonable callback implementation is to return is_forward_secure.
3531 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3532 by the SSL/TLS server library, indicating whether it can provide forward
3535 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3537 * New -verify_name option in command line utilities to set verification
3542 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3543 Add CMAC pkey methods.
3547 * Experimental renegotiation in s_server -www mode. If the client
3548 browses /reneg connection is renegotiated. If /renegcert it is
3549 renegotiated requesting a certificate.
3553 * Add an "external" session cache for debugging purposes to s_server. This
3554 should help trace issues which normally are only apparent in deployed
3555 multi-process servers.
3559 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3560 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3561 BIO_set_cipher() and some obscure PEM functions were changed so they
3562 can now return an error. The RAND changes required a change to the
3563 RAND_METHOD structure.
3567 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3568 a gcc attribute to warn if the result of a function is ignored. This
3569 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3570 whose return value is often ignored.
3574 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3575 These allow SCTs (signed certificate timestamps) to be requested and
3576 validated when establishing a connection.
3578 *Rob Percival <robpercival@google.com>*
3583 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019] ###
3585 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3586 used even when parsing explicit parameters, when loading a serialized key
3587 or calling `EC_GROUP_new_from_ecpkparameters()`/
3588 `EC_GROUP_new_from_ecparameters()`.
3589 This prevents bypass of security hardening and performance gains,
3590 especially for curves with specialized EC_METHODs.
3591 By default, if a key encoded with explicit parameters is loaded and later
3592 serialized, the output is still encoded with explicit parameters, even if
3593 internally a "named" EC_GROUP is used for computation.
3597 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3598 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3599 NULL. After this change, only the cofactor parameter can be NULL. It also
3600 does some minimal sanity checks on the passed order.
3605 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3606 An attack is simple, if the first CMS_recipientInfo is valid but the
3607 second CMS_recipientInfo is chosen ciphertext. If the second
3608 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3609 encryption key will be replaced by garbage, and the message cannot be
3610 decoded, but if the RSA decryption fails, the correct encryption key is
3611 used and the recipient will not notice the attack.
3612 As a work around for this potential attack the length of the decrypted
3613 key must be equal to the cipher default key length, in case the
3614 certifiate is not given and all recipientInfo are tried out.
3615 The old behaviour can be re-enabled in the CMS code by setting the
3616 CMS_DEBUG_DECRYPT flag.
3621 * Document issue with installation paths in diverse Windows builds
3623 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3624 binaries and run-time config file.
3629 ### Changes between 1.0.2r and 1.0.2s [28 May 2019] ###
3631 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3632 This changes the size when using the genpkey app when no size is given. It
3633 fixes an omission in earlier changes that changed all RSA, DSA and DH
3634 generation apps to use 2048 bits by default.
3638 * Add FIPS support for Android Arm 64-bit
3640 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3641 Module in Version 2.0.10. For some reason, the corresponding target
3642 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3643 built with FIPS support on Android Arm 64-bit. This omission has been
3646 *Matthias St. Pierre*
3648 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019] ###
3650 * 0-byte record padding oracle
3652 If an application encounters a fatal protocol error and then calls
3653 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3654 then OpenSSL can respond differently to the calling application if a 0 byte
3655 record is received with invalid padding compared to if a 0 byte record is
3656 received with an invalid MAC. If the application then behaves differently
3657 based on that in a way that is detectable to the remote peer, then this
3658 amounts to a padding oracle that could be used to decrypt data.
3660 In order for this to be exploitable "non-stitched" ciphersuites must be in
3661 use. Stitched ciphersuites are optimised implementations of certain
3662 commonly used ciphersuites. Also the application must call SSL_shutdown()
3663 twice even if a protocol error has occurred (applications should not do
3664 this but some do anyway).
3666 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3667 Aviram, with additional investigation by Steven Collison and Andrew
3668 Hourselt. It was reported to OpenSSL on 10th December 2018.
3673 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3677 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018] ###
3679 * Microarchitecture timing vulnerability in ECC scalar multiplication
3681 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3682 shown to be vulnerable to a microarchitecture timing side channel attack.
3683 An attacker with sufficient access to mount local timing attacks during
3684 ECDSA signature generation could recover the private key.
3686 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3687 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3693 * Timing vulnerability in DSA signature generation
3695 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3696 timing side channel attack. An attacker could use variations in the signing
3697 algorithm to recover the private key.
3699 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3704 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3705 Module, accidentally introduced while backporting security fixes from the
3706 development branch and hindering the use of ECC in FIPS mode.
3710 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018] ###
3712 * Client DoS due to large DH parameter
3714 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3715 malicious server can send a very large prime value to the client. This will
3716 cause the client to spend an unreasonably long period of time generating a
3717 key for this prime resulting in a hang until the client has finished. This
3718 could be exploited in a Denial Of Service attack.
3720 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3725 * Cache timing vulnerability in RSA Key Generation
3727 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3728 a cache timing side channel attack. An attacker with sufficient access to
3729 mount cache timing attacks during the RSA key generation process could
3730 recover the private key.
3732 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3733 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3738 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3739 parameter is no longer accepted, as it leads to a corrupt table. NULL
3740 pem_str is reserved for alias entries only.
3744 * Revert blinding in ECDSA sign and instead make problematic addition
3745 length-invariant. Switch even to fixed-length Montgomery multiplication.
3749 * Change generating and checking of primes so that the error rate of not
3750 being prime depends on the intended use based on the size of the input.
3751 For larger primes this will result in more rounds of Miller-Rabin.
3752 The maximal error rate for primes with more than 1080 bits is lowered
3755 *Kurt Roeckx, Annie Yousar*
3757 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3761 * Add blinding to ECDSA and DSA signatures to protect against side channel
3762 attacks discovered by Keegan Ryan (NCC Group).
3766 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3767 now allow empty (zero character) pass phrases.
3771 * Certificate time validation (X509_cmp_time) enforces stricter
3772 compliance with RFC 5280. Fractional seconds and timezone offsets
3773 are no longer allowed.
3777 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018] ###
3779 * Constructed ASN.1 types with a recursive definition could exceed the stack
3781 Constructed ASN.1 types with a recursive definition (such as can be found
3782 in PKCS7) could eventually exceed the stack given malicious input with
3783 excessive recursion. This could result in a Denial Of Service attack. There
3784 are no such structures used within SSL/TLS that come from untrusted sources
3785 so this is considered safe.
3787 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3793 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017] ###
3795 * Read/write after SSL object in error state
3797 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3798 mechanism. The intent was that if a fatal error occurred during a handshake
3799 then OpenSSL would move into the error state and would immediately fail if
3800 you attempted to continue the handshake. This works as designed for the
3801 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3802 SSL_connect()), however due to a bug it does not work correctly if
3803 SSL_read() or SSL_write() is called directly. In that scenario, if the
3804 handshake fails then a fatal error will be returned in the initial function
3805 call. If SSL_read()/SSL_write() is subsequently called by the application
3806 for the same SSL object then it will succeed and the data is passed without
3807 being decrypted/encrypted directly from the SSL/TLS record layer.
3809 In order to exploit this issue an application bug would have to be present
3810 that resulted in a call to SSL_read()/SSL_write() being issued after having
3811 already received a fatal error.
3813 This issue was reported to OpenSSL by David Benjamin (Google).
3818 * rsaz_1024_mul_avx2 overflow bug on x86_64
3820 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3821 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3822 Analysis suggests that attacks against RSA and DSA as a result of this
3823 defect would be very difficult to perform and are not believed likely.
3824 Attacks against DH1024 are considered just feasible, because most of the
3825 work necessary to deduce information about a private key may be performed
3826 offline. The amount of resources required for such an attack would be
3827 significant. However, for an attack on TLS to be meaningful, the server
3828 would have to share the DH1024 private key among multiple clients, which is
3829 no longer an option since CVE-2016-0701.
3831 This only affects processors that support the AVX2 but not ADX extensions
3832 like Intel Haswell (4th generation).
3834 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3835 was originally found via the OSS-Fuzz project.
3840 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017] ###
3842 * bn_sqrx8x_internal carry bug on x86_64
3844 There is a carry propagating bug in the x86_64 Montgomery squaring
3845 procedure. No EC algorithms are affected. Analysis suggests that attacks
3846 against RSA and DSA as a result of this defect would be very difficult to
3847 perform and are not believed likely. Attacks against DH are considered just
3848 feasible (although very difficult) because most of the work necessary to
3849 deduce information about a private key may be performed offline. The amount
3850 of resources required for such an attack would be very significant and
3851 likely only accessible to a limited number of attackers. An attacker would
3852 additionally need online access to an unpatched system using the target
3853 private key in a scenario with persistent DH parameters and a private
3854 key that is shared between multiple clients.
3856 This only affects processors that support the BMI1, BMI2 and ADX extensions
3857 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3859 This issue was reported to OpenSSL by the OSS-Fuzz project.
3864 * Malformed X.509 IPAddressFamily could cause OOB read
3866 If an X.509 certificate has a malformed IPAddressFamily extension,
3867 OpenSSL could do a one-byte buffer overread. The most likely result
3868 would be an erroneous display of the certificate in text format.
3870 This issue was reported to OpenSSL by the OSS-Fuzz project.
3875 ### Changes between 1.0.2k and 1.0.2l [25 May 2017] ###
3877 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3878 platform rather than 'mingw'.
3882 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017] ###
3884 * Truncated packet could crash via OOB read
3886 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3887 cipher is being used, then a truncated packet can cause that host to
3888 perform an out-of-bounds read, usually resulting in a crash.
3890 This issue was reported to OpenSSL by Robert Święcki of Google.
3895 * BN_mod_exp may produce incorrect results on x86_64
3897 There is a carry propagating bug in the x86_64 Montgomery squaring
3898 procedure. No EC algorithms are affected. Analysis suggests that attacks
3899 against RSA and DSA as a result of this defect would be very difficult to
3900 perform and are not believed likely. Attacks against DH are considered just
3901 feasible (although very difficult) because most of the work necessary to
3902 deduce information about a private key may be performed offline. The amount
3903 of resources required for such an attack would be very significant and
3904 likely only accessible to a limited number of attackers. An attacker would
3905 additionally need online access to an unpatched system using the target
3906 private key in a scenario with persistent DH parameters and a private
3907 key that is shared between multiple clients. For example this can occur by
3908 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3909 similar to CVE-2015-3193 but must be treated as a separate problem.
3911 This issue was reported to OpenSSL by the OSS-Fuzz project.
3916 * Montgomery multiplication may produce incorrect results
3918 There is a carry propagating bug in the Broadwell-specific Montgomery
3919 multiplication procedure that handles input lengths divisible by, but
3920 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3921 and DH private keys are impossible. This is because the subroutine in
3922 question is not used in operations with the private key itself and an input
3923 of the attacker's direct choice. Otherwise the bug can manifest itself as
3924 transient authentication and key negotiation failures or reproducible
3925 erroneous outcome of public-key operations with specially crafted input.
3926 Among EC algorithms only Brainpool P-512 curves are affected and one
3927 presumably can attack ECDH key negotiation. Impact was not analyzed in
3928 detail, because pre-requisites for attack are considered unlikely. Namely
3929 multiple clients have to choose the curve in question and the server has to
3930 share the private key among them, neither of which is default behaviour.
3931 Even then only clients that chose the curve will be affected.
3933 This issue was publicly reported as transient failures and was not
3934 initially recognized as a security issue. Thanks to Richard Morgan for
3935 providing reproducible case.
3940 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3941 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3942 prevent issues where no progress is being made and the peer continually
3943 sends unrecognised record types, using up resources processing them.
3947 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016] ###
3949 * Missing CRL sanity check
3951 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
3952 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
3953 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
3955 This issue only affects the OpenSSL 1.0.2i
3960 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016] ###
3962 * OCSP Status Request extension unbounded memory growth
3964 A malicious client can send an excessively large OCSP Status Request
3965 extension. If that client continually requests renegotiation, sending a
3966 large OCSP Status Request extension each time, then there will be unbounded
3967 memory growth on the server. This will eventually lead to a Denial Of
3968 Service attack through memory exhaustion. Servers with a default
3969 configuration are vulnerable even if they do not support OCSP. Builds using
3970 the "no-ocsp" build time option are not affected.
3972 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3977 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
3980 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
3986 * OOB write in MDC2_Update()
3988 An overflow can occur in MDC2_Update() either if called directly or
3989 through the EVP_DigestUpdate() function using MDC2. If an attacker
3990 is able to supply very large amounts of input data after a previous
3991 call to EVP_EncryptUpdate() with a partial block then a length check
3992 can overflow resulting in a heap corruption.
3994 The amount of data needed is comparable to SIZE_MAX which is impractical
3997 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4002 * Malformed SHA512 ticket DoS
4004 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4005 DoS attack where a malformed ticket will result in an OOB read which will
4008 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4009 a custom server callback and ticket lookup mechanism.
4011 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4016 * OOB write in BN_bn2dec()
4018 The function BN_bn2dec() does not check the return value of BN_div_word().
4019 This can cause an OOB write if an application uses this function with an
4020 overly large BIGNUM. This could be a problem if an overly large certificate
4021 or CRL is printed out from an untrusted source. TLS is not affected because
4022 record limits will reject an oversized certificate before it is parsed.
4024 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4029 * OOB read in TS_OBJ_print_bio()
4031 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4032 the total length the OID text representation would use and not the amount
4033 of data written. This will result in OOB reads when large OIDs are
4036 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4041 * Pointer arithmetic undefined behaviour
4043 Avoid some undefined pointer arithmetic
4045 A common idiom in the codebase is to check limits in the following manner:
4048 Where "p" points to some malloc'd data of SIZE bytes and
4051 "len" here could be from some externally supplied data (e.g. from a TLS
4054 The rules of C pointer arithmetic are such that "p + len" is only well
4055 defined where len <= SIZE. Therefore the above idiom is actually
4056 undefined behaviour.
4058 For example this could cause problems if some malloc implementation
4059 provides an address for "p" such that "p + len" actually overflows for
4060 values of len that are too big and therefore p + len < limit.
4062 This issue was reported to OpenSSL by Guido Vranken
4067 * Constant time flag not preserved in DSA signing
4069 Operations in the DSA signing algorithm should run in constant time in
4070 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4071 implementation means that a non-constant time codepath is followed for
4072 certain operations. This has been demonstrated through a cache-timing
4073 attack to be sufficient for an attacker to recover the private DSA key.
4075 This issue was reported by César Pereida (Aalto University), Billy Brumley
4076 (Tampere University of Technology), and Yuval Yarom (The University of
4077 Adelaide and NICTA).
4082 * DTLS buffered message DoS
4084 In a DTLS connection where handshake messages are delivered out-of-order
4085 those messages that OpenSSL is not yet ready to process will be buffered
4086 for later use. Under certain circumstances, a flaw in the logic means that
4087 those messages do not get removed from the buffer even though the handshake
4088 has been completed. An attacker could force up to approx. 15 messages to
4089 remain in the buffer when they are no longer required. These messages will
4090 be cleared when the DTLS connection is closed. The default maximum size for
4091 a message is 100k. Therefore the attacker could force an additional 1500k
4092 to be consumed per connection. By opening many simulataneous connections an
4093 attacker could cause a DoS attack through memory exhaustion.
4095 This issue was reported to OpenSSL by Quan Luo.
4100 * DTLS replay protection DoS
4102 A flaw in the DTLS replay attack protection mechanism means that records
4103 that arrive for future epochs update the replay protection "window" before
4104 the MAC for the record has been validated. This could be exploited by an
4105 attacker by sending a record for the next epoch (which does not have to
4106 decrypt or have a valid MAC), with a very large sequence number. This means
4107 that all subsequent legitimate packets are dropped causing a denial of
4108 service for a specific DTLS connection.
4110 This issue was reported to OpenSSL by the OCAP audit team.
4115 * Certificate message OOB reads
4117 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4118 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4119 theoretical DoS risk but this has not been observed in practice on common
4122 The messages affected are client certificate, client certificate request
4123 and server certificate. As a result the attack can only be performed
4124 against a client or a server which enables client authentication.
4126 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4131 ### Changes between 1.0.2g and 1.0.2h [3 May 2016] ###
4133 * Prevent padding oracle in AES-NI CBC MAC check
4135 A MITM attacker can use a padding oracle attack to decrypt traffic
4136 when the connection uses an AES CBC cipher and the server support
4139 This issue was introduced as part of the fix for Lucky 13 padding
4140 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4141 constant time by making sure that always the same bytes are read and
4142 compared against either the MAC or padding bytes. But it no longer
4143 checked that there was enough data to have both the MAC and padding
4146 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4151 * Fix EVP_EncodeUpdate overflow
4153 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4154 Base64 encoding of binary data. If an attacker is able to supply very large
4155 amounts of input data then a length check can overflow resulting in a heap
4158 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4159 the PEM_write_bio* family of functions. These are mainly used within the
4160 OpenSSL command line applications, so any application which processes data
4161 from an untrusted source and outputs it as a PEM file should be considered
4162 vulnerable to this issue. User applications that call these APIs directly
4163 with large amounts of untrusted data may also be vulnerable.
4165 This issue was reported by Guido Vranken.
4170 * Fix EVP_EncryptUpdate overflow
4172 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4173 is able to supply very large amounts of input data after a previous call to
4174 EVP_EncryptUpdate() with a partial block then a length check can overflow
4175 resulting in a heap corruption. Following an analysis of all OpenSSL
4176 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4177 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4178 the first called function after an EVP_EncryptInit(), and therefore that
4179 specific call must be safe. The second form is where the length passed to
4180 EVP_EncryptUpdate() can be seen from the code to be some small value and
4181 therefore there is no possibility of an overflow. Since all instances are
4182 one of these two forms, it is believed that there can be no overflows in
4183 internal code due to this problem. It should be noted that
4184 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4185 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4186 of these calls have also been analysed too and it is believed there are no
4187 instances in internal usage where an overflow could occur.
4189 This issue was reported by Guido Vranken.
4194 * Prevent ASN.1 BIO excessive memory allocation
4196 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4197 a short invalid encoding can cause allocation of large amounts of memory
4198 potentially consuming excessive resources or exhausting memory.
4200 Any application parsing untrusted data through d2i BIO functions is
4201 affected. The memory based functions such as d2i_X509() are *not* affected.
4202 Since the memory based functions are used by the TLS library, TLS
4203 applications are not affected.
4205 This issue was reported by Brian Carpenter.
4212 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4213 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4214 in arbitrary stack data being returned in the buffer.
4216 This issue was reported by Guido Vranken.
4221 * Modify behavior of ALPN to invoke callback after SNI/servername
4222 callback, such that updates to the SSL_CTX affect ALPN.
4226 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4231 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4232 methods are enabled and ssl2 is disabled the methods return NULL.
4236 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016] ###
4238 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4239 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4240 provide any "EXPORT" or "LOW" strength ciphers.
4244 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4245 is by default disabled at build-time. Builds that are not configured with
4246 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4247 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4248 will need to explicitly call either of:
4250 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4252 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4254 as appropriate. Even if either of those is used, or the application
4255 explicitly uses the version-specific SSLv2_method() or its client and
4256 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4257 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4258 ciphers, and SSLv2 56-bit DES are no longer available.
4263 * Fix a double-free in DSA code
4265 A double free bug was discovered when OpenSSL parses malformed DSA private
4266 keys and could lead to a DoS attack or memory corruption for applications
4267 that receive DSA private keys from untrusted sources. This scenario is
4270 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4276 * Disable SRP fake user seed to address a server memory leak.
4278 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4280 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4281 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4282 was changed to ignore the "fake user" SRP seed, even if the seed
4285 Users should use SRP_VBASE_get1_by_user instead. Note that in
4286 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4287 also that even though configuring the SRP seed attempts to hide
4288 invalid usernames by continuing the handshake with fake
4289 credentials, this behaviour is not constant time and no strong
4290 guarantees are made that the handshake is indistinguishable from
4291 that of a valid user.
4296 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4298 In the BN_hex2bn function the number of hex digits is calculated using an
4299 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4300 large values of |i| this can result in |bn_expand| not allocating any
4301 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4302 field as NULL leading to a subsequent NULL ptr deref. For very large values
4303 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4304 In this case memory is allocated to the internal BIGNUM data field, but it
4305 is insufficiently sized leading to heap corruption. A similar issue exists
4306 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4307 is ever called by user applications with very large untrusted hex/dec data.
4308 This is anticipated to be a rare occurrence.
4310 All OpenSSL internal usage of these functions use data that is not expected
4311 to be untrusted, e.g. config file data or application command line
4312 arguments. If user developed applications generate config file data based
4313 on untrusted data then it is possible that this could also lead to security
4314 consequences. This is also anticipated to be rare.
4316 This issue was reported to OpenSSL by Guido Vranken.
4321 * Fix memory issues in BIO_*printf functions
4323 The internal |fmtstr| function used in processing a "%s" format string in
4324 the BIO_*printf functions could overflow while calculating the length of a
4325 string and cause an OOB read when printing very long strings.
4327 Additionally the internal |doapr_outch| function can attempt to write to an
4328 OOB memory location (at an offset from the NULL pointer) in the event of a
4329 memory allocation failure. In 1.0.2 and below this could be caused where
4330 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4331 could be in processing a very long "%s" format string. Memory leaks can
4334 The first issue may mask the second issue dependent on compiler behaviour.
4335 These problems could enable attacks where large amounts of untrusted data
4336 is passed to the BIO_*printf functions. If applications use these functions
4337 in this way then they could be vulnerable. OpenSSL itself uses these
4338 functions when printing out human-readable dumps of ASN.1 data. Therefore
4339 applications that print this data could be vulnerable if the data is from
4340 untrusted sources. OpenSSL command line applications could also be
4341 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4342 as command line arguments.
4344 Libssl is not considered directly vulnerable. Additionally certificates etc
4345 received via remote connections via libssl are also unlikely to be able to
4346 trigger these issues because of message size limits enforced within libssl.
4348 This issue was reported to OpenSSL Guido Vranken.
4353 * Side channel attack on modular exponentiation
4355 A side-channel attack was found which makes use of cache-bank conflicts on
4356 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4357 of RSA keys. The ability to exploit this issue is limited as it relies on
4358 an attacker who has control of code in a thread running on the same
4359 hyper-threaded core as the victim thread which is performing decryptions.
4361 This issue was reported to OpenSSL by Yuval Yarom, The University of
4362 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4363 Nadia Heninger, University of Pennsylvania with more information at
4364 http://cachebleed.info.
4369 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4370 if no keysize is specified with default_bits. This fixes an
4371 omission in an earlier change that changed all RSA/DSA key generation
4372 apps to use 2048 bits by default.
4376 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016] ###
4377 * DH small subgroups
4379 Historically OpenSSL only ever generated DH parameters based on "safe"
4380 primes. More recently (in version 1.0.2) support was provided for
4381 generating X9.42 style parameter files such as those required for RFC 5114
4382 support. The primes used in such files may not be "safe". Where an
4383 application is using DH configured with parameters based on primes that are
4384 not "safe" then an attacker could use this fact to find a peer's private
4385 DH exponent. This attack requires that the attacker complete multiple
4386 handshakes in which the peer uses the same private DH exponent. For example
4387 this could be used to discover a TLS server's private DH exponent if it's
4388 reusing the private DH exponent or it's using a static DH ciphersuite.
4390 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4391 TLS. It is not on by default. If the option is not set then the server
4392 reuses the same private DH exponent for the life of the server process and
4393 would be vulnerable to this attack. It is believed that many popular
4394 applications do set this option and would therefore not be at risk.
4396 The fix for this issue adds an additional check where a "q" parameter is
4397 available (as is the case in X9.42 based parameters). This detects the
4398 only known attack, and is the only possible defense for static DH
4399 ciphersuites. This could have some performance impact.
4401 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4402 default and cannot be disabled. This could have some performance impact.
4404 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4409 * SSLv2 doesn't block disabled ciphers
4411 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4412 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4413 been disabled, provided that the SSLv2 protocol was not also disabled via
4416 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4417 and Sebastian Schinzel.
4422 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015] ###
4424 * BN_mod_exp may produce incorrect results on x86_64
4426 There is a carry propagating bug in the x86_64 Montgomery squaring
4427 procedure. No EC algorithms are affected. Analysis suggests that attacks
4428 against RSA and DSA as a result of this defect would be very difficult to
4429 perform and are not believed likely. Attacks against DH are considered just
4430 feasible (although very difficult) because most of the work necessary to
4431 deduce information about a private key may be performed offline. The amount
4432 of resources required for such an attack would be very significant and
4433 likely only accessible to a limited number of attackers. An attacker would
4434 additionally need online access to an unpatched system using the target
4435 private key in a scenario with persistent DH parameters and a private
4436 key that is shared between multiple clients. For example this can occur by
4437 default in OpenSSL DHE based SSL/TLS ciphersuites.
4439 This issue was reported to OpenSSL by Hanno Böck.
4444 * Certificate verify crash with missing PSS parameter
4446 The signature verification routines will crash with a NULL pointer
4447 dereference if presented with an ASN.1 signature using the RSA PSS
4448 algorithm and absent mask generation function parameter. Since these
4449 routines are used to verify certificate signature algorithms this can be
4450 used to crash any certificate verification operation and exploited in a
4451 DoS attack. Any application which performs certificate verification is
4452 vulnerable including OpenSSL clients and servers which enable client
4455 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4460 * X509_ATTRIBUTE memory leak
4462 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4463 memory. This structure is used by the PKCS#7 and CMS routines so any
4464 application which reads PKCS#7 or CMS data from untrusted sources is
4465 affected. SSL/TLS is not affected.
4467 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4473 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4474 This changes the decoding behaviour for some invalid messages,
4475 though the change is mostly in the more lenient direction, and
4476 legacy behaviour is preserved as much as possible.
4480 * In DSA_generate_parameters_ex, if the provided seed is too short,
4483 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4485 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015] ###
4487 * Alternate chains certificate forgery
4489 During certificate verification, OpenSSL will attempt to find an
4490 alternative certificate chain if the first attempt to build such a chain
4491 fails. An error in the implementation of this logic can mean that an
4492 attacker could cause certain checks on untrusted certificates to be
4493 bypassed, such as the CA flag, enabling them to use a valid leaf
4494 certificate to act as a CA and "issue" an invalid certificate.
4496 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4501 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015] ###
4503 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4504 incompatibility in the handling of HMAC. The previous ABI has now been
4509 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015] ###
4511 * Malformed ECParameters causes infinite loop
4513 When processing an ECParameters structure OpenSSL enters an infinite loop
4514 if the curve specified is over a specially malformed binary polynomial
4517 This can be used to perform denial of service against any
4518 system which processes public keys, certificate requests or
4519 certificates. This includes TLS clients and TLS servers with
4520 client authentication enabled.
4522 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4527 * Exploitable out-of-bounds read in X509_cmp_time
4529 X509_cmp_time does not properly check the length of the ASN1_TIME
4530 string and can read a few bytes out of bounds. In addition,
4531 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4534 An attacker can use this to craft malformed certificates and CRLs of
4535 various sizes and potentially cause a segmentation fault, resulting in
4536 a DoS on applications that verify certificates or CRLs. TLS clients
4537 that verify CRLs are affected. TLS clients and servers with client
4538 authentication enabled may be affected if they use custom verification
4541 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4542 independently by Hanno Böck.
4547 * PKCS7 crash with missing EnvelopedContent
4549 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4550 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4551 with missing content and trigger a NULL pointer dereference on parsing.
4553 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4554 structures from untrusted sources are affected. OpenSSL clients and
4555 servers are not affected.
4557 This issue was reported to OpenSSL by Michal Zalewski (Google).
4562 * CMS verify infinite loop with unknown hash function
4564 When verifying a signedData message the CMS code can enter an infinite loop
4565 if presented with an unknown hash function OID. This can be used to perform
4566 denial of service against any system which verifies signedData messages using
4568 This issue was reported to OpenSSL by Johannes Bauer.
4573 * Race condition handling NewSessionTicket
4575 If a NewSessionTicket is received by a multi-threaded client when attempting to
4576 reuse a previous ticket then a race condition can occur potentially leading to
4577 a double free of the ticket data.
4582 * Only support 256-bit or stronger elliptic curves with the
4583 'ecdh_auto' setting (server) or by default (client). Of supported
4584 curves, prefer P-256 (both).
4588 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015] ###
4590 * ClientHello sigalgs DoS fix
4592 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4593 invalid signature algorithms extension a NULL pointer dereference will
4594 occur. This can be exploited in a DoS attack against the server.
4596 This issue was was reported to OpenSSL by David Ramos of Stanford
4600 *Stephen Henson and Matt Caswell*
4602 * Multiblock corrupted pointer fix
4604 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4605 feature only applies on 64 bit x86 architecture platforms that support AES
4606 NI instructions. A defect in the implementation of "multiblock" can cause
4607 OpenSSL's internal write buffer to become incorrectly set to NULL when
4608 using non-blocking IO. Typically, when the user application is using a
4609 socket BIO for writing, this will only result in a failed connection.
4610 However if some other BIO is used then it is likely that a segmentation
4611 fault will be triggered, thus enabling a potential DoS attack.
4613 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4618 * Segmentation fault in DTLSv1_listen fix
4620 The DTLSv1_listen function is intended to be stateless and processes the
4621 initial ClientHello from many peers. It is common for user code to loop
4622 over the call to DTLSv1_listen until a valid ClientHello is received with
4623 an associated cookie. A defect in the implementation of DTLSv1_listen means
4624 that state is preserved in the SSL object from one invocation to the next
4625 that can lead to a segmentation fault. Errors processing the initial
4626 ClientHello can trigger this scenario. An example of such an error could be
4627 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4630 This issue was reported to OpenSSL by Per Allansson.
4635 * Segmentation fault in ASN1_TYPE_cmp fix
4637 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4638 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4639 certificate signature algorithm consistency this can be used to crash any
4640 certificate verification operation and exploited in a DoS attack. Any
4641 application which performs certificate verification is vulnerable including
4642 OpenSSL clients and servers which enable client authentication.
4647 * Segmentation fault for invalid PSS parameters fix
4649 The signature verification routines will crash with a NULL pointer
4650 dereference if presented with an ASN.1 signature using the RSA PSS
4651 algorithm and invalid parameters. Since these routines are used to verify
4652 certificate signature algorithms this can be used to crash any
4653 certificate verification operation and exploited in a DoS attack. Any
4654 application which performs certificate verification is vulnerable including
4655 OpenSSL clients and servers which enable client authentication.
4657 This issue was was reported to OpenSSL by Brian Carpenter.
4662 * ASN.1 structure reuse memory corruption fix
4664 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4665 memory corruption via an invalid write. Such reuse is and has been
4666 strongly discouraged and is believed to be rare.
4668 Applications that parse structures containing CHOICE or ANY DEFINED BY
4669 components may be affected. Certificate parsing (d2i_X509 and related
4670 functions) are however not affected. OpenSSL clients and servers are
4676 * PKCS7 NULL pointer dereferences fix
4678 The PKCS#7 parsing code does not handle missing outer ContentInfo
4679 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4680 missing content and trigger a NULL pointer dereference on parsing.
4682 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4683 otherwise parse PKCS#7 structures from untrusted sources are
4684 affected. OpenSSL clients and servers are not affected.
4686 This issue was reported to OpenSSL by Michal Zalewski (Google).
4691 * DoS via reachable assert in SSLv2 servers fix
4693 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4694 servers that both support SSLv2 and enable export cipher suites by sending
4695 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4697 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4698 (OpenSSL development team).
4703 * Empty CKE with client auth and DHE fix
4705 If client auth is used then a server can seg fault in the event of a DHE
4706 ciphersuite being selected and a zero length ClientKeyExchange message
4707 being sent by the client. This could be exploited in a DoS attack.
4712 * Handshake with unseeded PRNG fix
4714 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4715 with an unseeded PRNG. The conditions are:
4716 - The client is on a platform where the PRNG has not been seeded
4717 automatically, and the user has not seeded manually
4718 - A protocol specific client method version has been used (i.e. not
4719 SSL_client_methodv23)
4720 - A ciphersuite is used that does not require additional random data from
4721 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4723 If the handshake succeeds then the client random that has been used will
4724 have been generated from a PRNG with insufficient entropy and therefore the
4725 output may be predictable.
4727 For example using the following command with an unseeded openssl will
4728 succeed on an unpatched platform:
4730 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4735 * Use After Free following d2i_ECPrivatekey error fix
4737 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4738 could cause a use after free condition. This, in turn, could cause a double
4739 free in several private key parsing functions (such as d2i_PrivateKey
4740 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4741 for applications that receive EC private keys from untrusted
4742 sources. This scenario is considered rare.
4744 This issue was discovered by the BoringSSL project and fixed in their
4750 * X509_to_X509_REQ NULL pointer deref fix
4752 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4753 the certificate key is invalid. This function is rarely used in practice.
4755 This issue was discovered by Brian Carpenter.
4760 * Removed the export ciphers from the DEFAULT ciphers
4764 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015] ###
4766 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4767 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4768 So far those who have to target multiple platforms would compromise
4769 and argue that binary targeting say ARMv5 would still execute on
4770 ARMv8. "Universal" build resolves this compromise by providing
4771 near-optimal performance even on newer platforms.
4775 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4776 (other platforms pending).
4778 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4780 * Add support for the SignedCertificateTimestampList certificate and
4781 OCSP response extensions from RFC6962.
4785 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4786 for corner cases. (Certain input points at infinity could lead to
4787 bogus results, with non-infinity inputs mapped to infinity too.)
4791 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4792 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4793 common cases are optimized and there still is room for further
4794 improvements. Vector Permutation AES for Altivec is also added.
4798 * Add support for little-endian ppc64 Linux target.
4800 *Marcelo Cerri (IBM)*
4802 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4803 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4804 are optimized and there still is room for further improvements.
4805 Both 32- and 64-bit modes are supported.
4807 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4809 * Improved ARMv7 NEON support.
4813 * Support for SPARC Architecture 2011 crypto extensions, first
4814 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4815 SHA256/512, MD5, GHASH and modular exponentiation.
4817 *Andy Polyakov, David Miller*
4819 * Accelerated modular exponentiation for Intel processors, a.k.a.
4822 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4824 * Support for new and upcoming Intel processors, including AVX2,
4825 BMI and SHA ISA extensions. This includes additional "stitched"
4826 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4829 This work was sponsored by Intel Corp.
4833 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4834 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4835 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4839 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4840 this fixes a limitation in previous versions of OpenSSL.
4844 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4845 MGF1 digest and OAEP label.
4849 * Add EVP support for key wrapping algorithms, to avoid problems with
4850 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4851 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4852 algorithms and include tests cases.
4856 * Add functions to allocate and set the fields of an ECDSA_METHOD
4859 *Douglas E. Engert, Steve Henson*
4861 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4862 difference in days and seconds between two tm or ASN1_TIME structures.
4866 * Add -rev test option to s_server to just reverse order of characters
4867 received by client and send back to server. Also prints an abbreviated
4868 summary of the connection parameters.
4872 * New option -brief for s_client and s_server to print out a brief summary
4873 of connection parameters.
4877 * Add callbacks for arbitrary TLS extensions.
4879 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4881 * New option -crl_download in several openssl utilities to download CRLs
4882 from CRLDP extension in certificates.
4886 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4890 * New function X509_CRL_diff to generate a delta CRL from the difference
4891 of two full CRLs. Add support to "crl" utility.
4895 * New functions to set lookup_crls function and to retrieve
4896 X509_STORE from X509_STORE_CTX.
4900 * Print out deprecated issuer and subject unique ID fields in
4905 * Extend OCSP I/O functions so they can be used for simple general purpose
4906 HTTP as well as OCSP. New wrapper function which can be used to download
4907 CRLs using the OCSP API.
4911 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
4915 * SSL_CONF* functions. These provide a common framework for application
4916 configuration using configuration files or command lines.
4920 * SSL/TLS tracing code. This parses out SSL/TLS records using the
4921 message callback and prints the results. Needs compile time option
4922 "enable-ssl-trace". New options to s_client and s_server to enable
4927 * New ctrl and macro to retrieve supported points extensions.
4928 Print out extension in s_server and s_client.
4932 * New functions to retrieve certificate signature and signature
4937 * Add functions to retrieve and manipulate the raw cipherlist sent by a
4942 * New Suite B modes for TLS code. These use and enforce the requirements
4943 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
4944 only use Suite B curves. The Suite B modes can be set by using the
4945 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
4949 * New chain verification flags for Suite B levels of security. Check
4950 algorithms are acceptable when flags are set in X509_verify_cert.
4954 * Make tls1_check_chain return a set of flags indicating checks passed
4955 by a certificate chain. Add additional tests to handle client
4956 certificates: checks for matching certificate type and issuer name
4961 * If an attempt is made to use a signature algorithm not in the peer
4962 preference list abort the handshake. If client has no suitable
4963 signature algorithms in response to a certificate request do not
4964 use the certificate.
4968 * If server EC tmp key is not in client preference list abort handshake.
4972 * Add support for certificate stores in CERT structure. This makes it
4973 possible to have different stores per SSL structure or one store in
4974 the parent SSL_CTX. Include distinct stores for certificate chain
4975 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
4976 to build and store a certificate chain in CERT structure: returning
4977 an error if the chain cannot be built: this will allow applications
4978 to test if a chain is correctly configured.
4980 Note: if the CERT based stores are not set then the parent SSL_CTX
4981 store is used to retain compatibility with existing behaviour.
4986 * New function ssl_set_client_disabled to set a ciphersuite disabled
4987 mask based on the current session, check mask when sending client
4988 hello and checking the requested ciphersuite.
4992 * New ctrls to retrieve and set certificate types in a certificate
4993 request message. Print out received values in s_client. If certificate
4994 types is not set with custom values set sensible values based on
4995 supported signature algorithms.
4999 * Support for distinct client and server supported signature algorithms.
5003 * Add certificate callback. If set this is called whenever a certificate
5004 is required by client or server. An application can decide which
5005 certificate chain to present based on arbitrary criteria: for example
5006 supported signature algorithms. Add very simple example to s_server.
5007 This fixes many of the problems and restrictions of the existing client
5008 certificate callback: for example you can now clear an existing
5009 certificate and specify the whole chain.
5013 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5014 the certificate can be used for (if anything). Set valid_flags field
5015 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5016 to have similar checks in it.
5018 Add new "cert_flags" field to CERT structure and include a "strict mode".
5019 This enforces some TLS certificate requirements (such as only permitting
5020 certificate signature algorithms contained in the supported algorithms
5021 extension) which some implementations ignore: this option should be used
5022 with caution as it could cause interoperability issues.
5026 * Update and tidy signature algorithm extension processing. Work out
5027 shared signature algorithms based on preferences and peer algorithms
5028 and print them out in s_client and s_server. Abort handshake if no
5029 shared signature algorithms.
5033 * Add new functions to allow customised supported signature algorithms
5034 for SSL and SSL_CTX structures. Add options to s_client and s_server
5039 * New function SSL_certs_clear() to delete all references to certificates
5040 from an SSL structure. Before this once a certificate had been added
5041 it couldn't be removed.
5045 * Integrate hostname, email address and IP address checking with certificate
5046 verification. New verify options supporting checking in openssl utility.
5050 * Fixes and wildcard matching support to hostname and email checking
5051 functions. Add manual page.
5053 *Florian Weimer (Red Hat Product Security Team)*
5055 * New functions to check a hostname email or IP address against a
5056 certificate. Add options x509 utility to print results of checks against
5061 * Fix OCSP checking.
5063 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5065 * Initial experimental support for explicitly trusted non-root CAs.
5066 OpenSSL still tries to build a complete chain to a root but if an
5067 intermediate CA has a trust setting included that is used. The first
5068 setting is used: whether to trust (e.g., -addtrust option to the x509
5073 * Add -trusted_first option which attempts to find certificates in the
5074 trusted store even if an untrusted chain is also supplied.
5078 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5079 platform support for Linux and Android.
5083 * Support for linux-x32, ILP32 environment in x86_64 framework.
5087 * Experimental multi-implementation support for FIPS capable OpenSSL.
5088 When in FIPS mode the approved implementations are used as normal,
5089 when not in FIPS mode the internal unapproved versions are used instead.
5090 This means that the FIPS capable OpenSSL isn't forced to use the
5091 (often lower performance) FIPS implementations outside FIPS mode.
5095 * Transparently support X9.42 DH parameters when calling
5096 PEM_read_bio_DHparameters. This means existing applications can handle
5097 the new parameter format automatically.
5101 * Initial experimental support for X9.42 DH parameter format: mainly
5102 to support use of 'q' parameter for RFC5114 parameters.
5106 * Add DH parameters from RFC5114 including test data to dhtest.
5110 * Support for automatic EC temporary key parameter selection. If enabled
5111 the most preferred EC parameters are automatically used instead of
5112 hardcoded fixed parameters. Now a server just has to call:
5113 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5114 support ECDH and use the most appropriate parameters.
5118 * Enhance and tidy EC curve and point format TLS extension code. Use
5119 static structures instead of allocation if default values are used.
5120 New ctrls to set curves we wish to support and to retrieve shared curves.
5121 Print out shared curves in s_server. New options to s_server and s_client
5122 to set list of supported curves.
5126 * New ctrls to retrieve supported signature algorithms and
5127 supported curve values as an array of NIDs. Extend openssl utility
5128 to print out received values.
5132 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5133 between NIDs and the more common NIST names such as "P-256". Enhance
5134 ecparam utility and ECC method to recognise the NIST names for curves.
5138 * Enhance SSL/TLS certificate chain handling to support different
5139 chains for each certificate instead of one chain in the parent SSL_CTX.
5143 * Support for fixed DH ciphersuite client authentication: where both
5144 server and client use DH certificates with common parameters.
5148 * Support for fixed DH ciphersuites: those requiring DH server
5153 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5155 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5156 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5157 X509_CINF_get_signature were reverted post internal team review.
5163 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016] ###
5165 * OCSP Status Request extension unbounded memory growth
5167 A malicious client can send an excessively large OCSP Status Request
5168 extension. If that client continually requests renegotiation, sending a
5169 large OCSP Status Request extension each time, then there will be unbounded
5170 memory growth on the server. This will eventually lead to a Denial Of
5171 Service attack through memory exhaustion. Servers with a default
5172 configuration are vulnerable even if they do not support OCSP. Builds using
5173 the "no-ocsp" build time option are not affected.
5175 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5180 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5183 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5189 * OOB write in MDC2_Update()
5191 An overflow can occur in MDC2_Update() either if called directly or
5192 through the EVP_DigestUpdate() function using MDC2. If an attacker
5193 is able to supply very large amounts of input data after a previous
5194 call to EVP_EncryptUpdate() with a partial block then a length check
5195 can overflow resulting in a heap corruption.
5197 The amount of data needed is comparable to SIZE_MAX which is impractical
5200 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5205 * Malformed SHA512 ticket DoS
5207 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5208 DoS attack where a malformed ticket will result in an OOB read which will
5211 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5212 a custom server callback and ticket lookup mechanism.
5214 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5219 * OOB write in BN_bn2dec()
5221 The function BN_bn2dec() does not check the return value of BN_div_word().
5222 This can cause an OOB write if an application uses this function with an
5223 overly large BIGNUM. This could be a problem if an overly large certificate
5224 or CRL is printed out from an untrusted source. TLS is not affected because
5225 record limits will reject an oversized certificate before it is parsed.
5227 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5232 * OOB read in TS_OBJ_print_bio()
5234 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5235 the total length the OID text representation would use and not the amount
5236 of data written. This will result in OOB reads when large OIDs are
5239 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5244 * Pointer arithmetic undefined behaviour
5246 Avoid some undefined pointer arithmetic
5248 A common idiom in the codebase is to check limits in the following manner:
5251 Where "p" points to some malloc'd data of SIZE bytes and
5254 "len" here could be from some externally supplied data (e.g. from a TLS
5257 The rules of C pointer arithmetic are such that "p + len" is only well
5258 defined where len <= SIZE. Therefore the above idiom is actually
5259 undefined behaviour.
5261 For example this could cause problems if some malloc implementation
5262 provides an address for "p" such that "p + len" actually overflows for
5263 values of len that are too big and therefore p + len < limit.
5265 This issue was reported to OpenSSL by Guido Vranken
5270 * Constant time flag not preserved in DSA signing
5272 Operations in the DSA signing algorithm should run in constant time in
5273 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5274 implementation means that a non-constant time codepath is followed for
5275 certain operations. This has been demonstrated through a cache-timing
5276 attack to be sufficient for an attacker to recover the private DSA key.
5278 This issue was reported by César Pereida (Aalto University), Billy Brumley
5279 (Tampere University of Technology), and Yuval Yarom (The University of
5280 Adelaide and NICTA).
5285 * DTLS buffered message DoS
5287 In a DTLS connection where handshake messages are delivered out-of-order
5288 those messages that OpenSSL is not yet ready to process will be buffered
5289 for later use. Under certain circumstances, a flaw in the logic means that
5290 those messages do not get removed from the buffer even though the handshake
5291 has been completed. An attacker could force up to approx. 15 messages to
5292 remain in the buffer when they are no longer required. These messages will
5293 be cleared when the DTLS connection is closed. The default maximum size for
5294 a message is 100k. Therefore the attacker could force an additional 1500k
5295 to be consumed per connection. By opening many simulataneous connections an
5296 attacker could cause a DoS attack through memory exhaustion.
5298 This issue was reported to OpenSSL by Quan Luo.
5303 * DTLS replay protection DoS
5305 A flaw in the DTLS replay attack protection mechanism means that records
5306 that arrive for future epochs update the replay protection "window" before
5307 the MAC for the record has been validated. This could be exploited by an
5308 attacker by sending a record for the next epoch (which does not have to
5309 decrypt or have a valid MAC), with a very large sequence number. This means
5310 that all subsequent legitimate packets are dropped causing a denial of
5311 service for a specific DTLS connection.
5313 This issue was reported to OpenSSL by the OCAP audit team.
5318 * Certificate message OOB reads
5320 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5321 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5322 theoretical DoS risk but this has not been observed in practice on common
5325 The messages affected are client certificate, client certificate request
5326 and server certificate. As a result the attack can only be performed
5327 against a client or a server which enables client authentication.
5329 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5334 ### Changes between 1.0.1s and 1.0.1t [3 May 2016] ###
5336 * Prevent padding oracle in AES-NI CBC MAC check
5338 A MITM attacker can use a padding oracle attack to decrypt traffic
5339 when the connection uses an AES CBC cipher and the server support
5342 This issue was introduced as part of the fix for Lucky 13 padding
5343 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5344 constant time by making sure that always the same bytes are read and
5345 compared against either the MAC or padding bytes. But it no longer
5346 checked that there was enough data to have both the MAC and padding
5349 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5354 * Fix EVP_EncodeUpdate overflow
5356 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5357 Base64 encoding of binary data. If an attacker is able to supply very large
5358 amounts of input data then a length check can overflow resulting in a heap
5361 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5362 the PEM_write_bio* family of functions. These are mainly used within the
5363 OpenSSL command line applications, so any application which processes data
5364 from an untrusted source and outputs it as a PEM file should be considered
5365 vulnerable to this issue. User applications that call these APIs directly
5366 with large amounts of untrusted data may also be vulnerable.
5368 This issue was reported by Guido Vranken.
5373 * Fix EVP_EncryptUpdate overflow
5375 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5376 is able to supply very large amounts of input data after a previous call to
5377 EVP_EncryptUpdate() with a partial block then a length check can overflow
5378 resulting in a heap corruption. Following an analysis of all OpenSSL
5379 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5380 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5381 the first called function after an EVP_EncryptInit(), and therefore that
5382 specific call must be safe. The second form is where the length passed to
5383 EVP_EncryptUpdate() can be seen from the code to be some small value and
5384 therefore there is no possibility of an overflow. Since all instances are
5385 one of these two forms, it is believed that there can be no overflows in
5386 internal code due to this problem. It should be noted that
5387 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5388 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5389 of these calls have also been analysed too and it is believed there are no
5390 instances in internal usage where an overflow could occur.
5392 This issue was reported by Guido Vranken.
5397 * Prevent ASN.1 BIO excessive memory allocation
5399 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5400 a short invalid encoding can casuse allocation of large amounts of memory
5401 potentially consuming excessive resources or exhausting memory.
5403 Any application parsing untrusted data through d2i BIO functions is
5404 affected. The memory based functions such as d2i_X509() are *not* affected.
5405 Since the memory based functions are used by the TLS library, TLS
5406 applications are not affected.
5408 This issue was reported by Brian Carpenter.
5415 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5416 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5417 in arbitrary stack data being returned in the buffer.
5419 This issue was reported by Guido Vranken.
5424 * Modify behavior of ALPN to invoke callback after SNI/servername
5425 callback, such that updates to the SSL_CTX affect ALPN.
5429 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5434 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5435 methods are enabled and ssl2 is disabled the methods return NULL.
5439 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016] ###
5441 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5442 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5443 provide any "EXPORT" or "LOW" strength ciphers.
5447 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5448 is by default disabled at build-time. Builds that are not configured with
5449 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5450 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5451 will need to explicitly call either of:
5453 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5455 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5457 as appropriate. Even if either of those is used, or the application
5458 explicitly uses the version-specific SSLv2_method() or its client and
5459 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5460 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5461 ciphers, and SSLv2 56-bit DES are no longer available.
5466 * Fix a double-free in DSA code
5468 A double free bug was discovered when OpenSSL parses malformed DSA private
5469 keys and could lead to a DoS attack or memory corruption for applications
5470 that receive DSA private keys from untrusted sources. This scenario is
5473 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5479 * Disable SRP fake user seed to address a server memory leak.
5481 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5483 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5484 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5485 was changed to ignore the "fake user" SRP seed, even if the seed
5488 Users should use SRP_VBASE_get1_by_user instead. Note that in
5489 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5490 also that even though configuring the SRP seed attempts to hide
5491 invalid usernames by continuing the handshake with fake
5492 credentials, this behaviour is not constant time and no strong
5493 guarantees are made that the handshake is indistinguishable from
5494 that of a valid user.
5499 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5501 In the BN_hex2bn function the number of hex digits is calculated using an
5502 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5503 large values of |i| this can result in |bn_expand| not allocating any
5504 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5505 field as NULL leading to a subsequent NULL ptr deref. For very large values
5506 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5507 In this case memory is allocated to the internal BIGNUM data field, but it
5508 is insufficiently sized leading to heap corruption. A similar issue exists
5509 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5510 is ever called by user applications with very large untrusted hex/dec data.
5511 This is anticipated to be a rare occurrence.
5513 All OpenSSL internal usage of these functions use data that is not expected
5514 to be untrusted, e.g. config file data or application command line
5515 arguments. If user developed applications generate config file data based
5516 on untrusted data then it is possible that this could also lead to security
5517 consequences. This is also anticipated to be rare.
5519 This issue was reported to OpenSSL by Guido Vranken.
5524 * Fix memory issues in BIO_*printf functions
5526 The internal |fmtstr| function used in processing a "%s" format string in
5527 the BIO_*printf functions could overflow while calculating the length of a
5528 string and cause an OOB read when printing very long strings.
5530 Additionally the internal |doapr_outch| function can attempt to write to an
5531 OOB memory location (at an offset from the NULL pointer) in the event of a
5532 memory allocation failure. In 1.0.2 and below this could be caused where
5533 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5534 could be in processing a very long "%s" format string. Memory leaks can
5537 The first issue may mask the second issue dependent on compiler behaviour.
5538 These problems could enable attacks where large amounts of untrusted data
5539 is passed to the BIO_*printf functions. If applications use these functions
5540 in this way then they could be vulnerable. OpenSSL itself uses these
5541 functions when printing out human-readable dumps of ASN.1 data. Therefore
5542 applications that print this data could be vulnerable if the data is from
5543 untrusted sources. OpenSSL command line applications could also be
5544 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5545 as command line arguments.
5547 Libssl is not considered directly vulnerable. Additionally certificates etc
5548 received via remote connections via libssl are also unlikely to be able to
5549 trigger these issues because of message size limits enforced within libssl.
5551 This issue was reported to OpenSSL Guido Vranken.
5556 * Side channel attack on modular exponentiation
5558 A side-channel attack was found which makes use of cache-bank conflicts on
5559 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5560 of RSA keys. The ability to exploit this issue is limited as it relies on
5561 an attacker who has control of code in a thread running on the same
5562 hyper-threaded core as the victim thread which is performing decryptions.
5564 This issue was reported to OpenSSL by Yuval Yarom, The University of
5565 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5566 Nadia Heninger, University of Pennsylvania with more information at
5567 http://cachebleed.info.
5572 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5573 if no keysize is specified with default_bits. This fixes an
5574 omission in an earlier change that changed all RSA/DSA key generation
5575 apps to use 2048 bits by default.
5579 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016] ###
5581 * Protection for DH small subgroup attacks
5583 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5584 switched on by default and cannot be disabled. This could have some
5589 * SSLv2 doesn't block disabled ciphers
5591 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5592 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5593 been disabled, provided that the SSLv2 protocol was not also disabled via
5596 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5597 and Sebastian Schinzel.
5602 * Reject DH handshakes with parameters shorter than 1024 bits.
5606 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015] ###
5608 * Certificate verify crash with missing PSS parameter
5610 The signature verification routines will crash with a NULL pointer
5611 dereference if presented with an ASN.1 signature using the RSA PSS
5612 algorithm and absent mask generation function parameter. Since these
5613 routines are used to verify certificate signature algorithms this can be
5614 used to crash any certificate verification operation and exploited in a
5615 DoS attack. Any application which performs certificate verification is
5616 vulnerable including OpenSSL clients and servers which enable client
5619 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5624 * X509_ATTRIBUTE memory leak
5626 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5627 memory. This structure is used by the PKCS#7 and CMS routines so any
5628 application which reads PKCS#7 or CMS data from untrusted sources is
5629 affected. SSL/TLS is not affected.
5631 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5637 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5638 This changes the decoding behaviour for some invalid messages,
5639 though the change is mostly in the more lenient direction, and
5640 legacy behaviour is preserved as much as possible.
5644 * In DSA_generate_parameters_ex, if the provided seed is too short,
5645 use a random seed, as already documented.
5647 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5649 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015] ###
5651 * Alternate chains certificate forgery
5653 During certificate verfification, OpenSSL will attempt to find an
5654 alternative certificate chain if the first attempt to build such a chain
5655 fails. An error in the implementation of this logic can mean that an
5656 attacker could cause certain checks on untrusted certificates to be
5657 bypassed, such as the CA flag, enabling them to use a valid leaf
5658 certificate to act as a CA and "issue" an invalid certificate.
5660 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5666 * Race condition handling PSK identify hint
5668 If PSK identity hints are received by a multi-threaded client then
5669 the values are wrongly updated in the parent SSL_CTX structure. This can
5670 result in a race condition potentially leading to a double free of the
5676 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015] ###
5677 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5678 incompatibility in the handling of HMAC. The previous ABI has now been
5681 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015] ###
5683 * Malformed ECParameters causes infinite loop
5685 When processing an ECParameters structure OpenSSL enters an infinite loop
5686 if the curve specified is over a specially malformed binary polynomial
5689 This can be used to perform denial of service against any
5690 system which processes public keys, certificate requests or
5691 certificates. This includes TLS clients and TLS servers with
5692 client authentication enabled.
5694 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5699 * Exploitable out-of-bounds read in X509_cmp_time
5701 X509_cmp_time does not properly check the length of the ASN1_TIME
5702 string and can read a few bytes out of bounds. In addition,
5703 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5706 An attacker can use this to craft malformed certificates and CRLs of
5707 various sizes and potentially cause a segmentation fault, resulting in
5708 a DoS on applications that verify certificates or CRLs. TLS clients
5709 that verify CRLs are affected. TLS clients and servers with client
5710 authentication enabled may be affected if they use custom verification
5713 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5714 independently by Hanno Böck.
5719 * PKCS7 crash with missing EnvelopedContent
5721 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5722 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5723 with missing content and trigger a NULL pointer dereference on parsing.
5725 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5726 structures from untrusted sources are affected. OpenSSL clients and
5727 servers are not affected.
5729 This issue was reported to OpenSSL by Michal Zalewski (Google).
5734 * CMS verify infinite loop with unknown hash function
5736 When verifying a signedData message the CMS code can enter an infinite loop
5737 if presented with an unknown hash function OID. This can be used to perform
5738 denial of service against any system which verifies signedData messages using
5740 This issue was reported to OpenSSL by Johannes Bauer.
5745 * Race condition handling NewSessionTicket
5747 If a NewSessionTicket is received by a multi-threaded client when attempting to
5748 reuse a previous ticket then a race condition can occur potentially leading to
5749 a double free of the ticket data.
5754 * Reject DH handshakes with parameters shorter than 768 bits.
5756 *Kurt Roeckx and Emilia Kasper*
5758 * dhparam: generate 2048-bit parameters by default.
5760 *Kurt Roeckx and Emilia Kasper*
5762 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015] ###
5764 * Segmentation fault in ASN1_TYPE_cmp fix
5766 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5767 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5768 certificate signature algorithm consistency this can be used to crash any
5769 certificate verification operation and exploited in a DoS attack. Any
5770 application which performs certificate verification is vulnerable including
5771 OpenSSL clients and servers which enable client authentication.
5776 * ASN.1 structure reuse memory corruption fix
5778 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5779 memory corruption via an invalid write. Such reuse is and has been
5780 strongly discouraged and is believed to be rare.
5782 Applications that parse structures containing CHOICE or ANY DEFINED BY
5783 components may be affected. Certificate parsing (d2i_X509 and related
5784 functions) are however not affected. OpenSSL clients and servers are
5790 * PKCS7 NULL pointer dereferences fix
5792 The PKCS#7 parsing code does not handle missing outer ContentInfo
5793 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5794 missing content and trigger a NULL pointer dereference on parsing.
5796 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5797 otherwise parse PKCS#7 structures from untrusted sources are
5798 affected. OpenSSL clients and servers are not affected.
5800 This issue was reported to OpenSSL by Michal Zalewski (Google).
5805 * DoS via reachable assert in SSLv2 servers fix
5807 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5808 servers that both support SSLv2 and enable export cipher suites by sending
5809 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5811 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5812 (OpenSSL development team).
5817 * Use After Free following d2i_ECPrivatekey error fix
5819 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5820 could cause a use after free condition. This, in turn, could cause a double
5821 free in several private key parsing functions (such as d2i_PrivateKey
5822 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5823 for applications that receive EC private keys from untrusted
5824 sources. This scenario is considered rare.
5826 This issue was discovered by the BoringSSL project and fixed in their
5832 * X509_to_X509_REQ NULL pointer deref fix
5834 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5835 the certificate key is invalid. This function is rarely used in practice.
5837 This issue was discovered by Brian Carpenter.
5842 * Removed the export ciphers from the DEFAULT ciphers
5846 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015] ###
5848 * Build fixes for the Windows and OpenVMS platforms
5850 *Matt Caswell and Richard Levitte*
5852 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015] ###
5854 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5855 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5856 dereference. This could lead to a Denial Of Service attack. Thanks to
5857 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5862 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5863 dtls1_buffer_record function under certain conditions. In particular this
5864 could occur if an attacker sent repeated DTLS records with the same
5865 sequence number but for the next epoch. The memory leak could be exploited
5866 by an attacker in a Denial of Service attack through memory exhaustion.
5867 Thanks to Chris Mueller for reporting this issue.
5872 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5873 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5874 method would be set to NULL which could later result in a NULL pointer
5875 dereference. Thanks to Frank Schmirler for reporting this issue.
5880 * Abort handshake if server key exchange message is omitted for ephemeral
5883 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5884 reporting this issue.
5889 * Remove non-export ephemeral RSA code on client and server. This code
5890 violated the TLS standard by allowing the use of temporary RSA keys in
5891 non-export ciphersuites and could be used by a server to effectively
5892 downgrade the RSA key length used to a value smaller than the server
5893 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
5894 INRIA or reporting this issue.
5899 * Fixed issue where DH client certificates are accepted without verification.
5900 An OpenSSL server will accept a DH certificate for client authentication
5901 without the certificate verify message. This effectively allows a client to
5902 authenticate without the use of a private key. This only affects servers
5903 which trust a client certificate authority which issues certificates
5904 containing DH keys: these are extremely rare and hardly ever encountered.
5905 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
5911 * Ensure that the session ID context of an SSL is updated when its
5912 SSL_CTX is updated via SSL_set_SSL_CTX.
5914 The session ID context is typically set from the parent SSL_CTX,
5915 and can vary with the CTX.
5919 * Fix various certificate fingerprint issues.
5921 By using non-DER or invalid encodings outside the signed portion of a
5922 certificate the fingerprint can be changed without breaking the signature.
5923 Although no details of the signed portion of the certificate can be changed
5924 this can cause problems with some applications: e.g. those using the
5925 certificate fingerprint for blacklists.
5927 1. Reject signatures with non zero unused bits.
5929 If the BIT STRING containing the signature has non zero unused bits reject
5930 the signature. All current signature algorithms require zero unused bits.
5932 2. Check certificate algorithm consistency.
5934 Check the AlgorithmIdentifier inside TBS matches the one in the
5935 certificate signature. NB: this will result in signature failure
5936 errors for some broken certificates.
5938 Thanks to Konrad Kraszewski from Google for reporting this issue.
5940 3. Check DSA/ECDSA signatures use DER.
5942 Re-encode DSA/ECDSA signatures and compare with the original received
5943 signature. Return an error if there is a mismatch.
5945 This will reject various cases including garbage after signature
5946 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
5947 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
5948 (negative or with leading zeroes).
5950 Further analysis was conducted and fixes were developed by Stephen Henson
5951 of the OpenSSL core team.
5957 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
5958 results on some platforms, including x86_64. This bug occurs at random
5959 with a very low probability, and is not known to be exploitable in any
5960 way, though its exact impact is difficult to determine. Thanks to Pieter
5961 Wuille (Blockstream) who reported this issue and also suggested an initial
5962 fix. Further analysis was conducted by the OpenSSL development team and
5963 Adam Langley of Google. The final fix was developed by Andy Polyakov of
5964 the OpenSSL core team.
5969 *) Do not resume sessions on the server if the negotiated protocol
5970 version does not match the session's version. Resuming with a different
5971 version, while not strictly forbidden by the RFC, is of questionable
5972 sanity and breaks all known clients.
5974 *David Benjamin, Emilia Käsper*
5976 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
5977 early CCS messages during renegotiation. (Note that because
5978 renegotiation is encrypted, this early CCS was not exploitable.)
5982 *) Tighten client-side session ticket handling during renegotiation:
5983 ensure that the client only accepts a session ticket if the server sends
5984 the extension anew in the ServerHello. Previously, a TLS client would
5985 reuse the old extension state and thus accept a session ticket if one was
5986 announced in the initial ServerHello.
5988 Similarly, ensure that the client requires a session ticket if one
5989 was advertised in the ServerHello. Previously, a TLS client would
5990 ignore a missing NewSessionTicket message.
5994 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014] ###
5998 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
5999 sends a carefully crafted handshake message, to cause OpenSSL to fail
6000 to free up to 64k of memory causing a memory leak. This could be
6001 exploited in a Denial Of Service attack. This issue affects OpenSSL
6002 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6003 whether SRTP is used or configured. Implementations of OpenSSL that
6004 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6006 The fix was developed by the OpenSSL team.
6011 * Session Ticket Memory Leak.
6013 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6014 integrity of that ticket is first verified. In the event of a session
6015 ticket integrity check failing, OpenSSL will fail to free memory
6016 causing a memory leak. By sending a large number of invalid session
6017 tickets an attacker could exploit this issue in a Denial Of Service
6023 * Build option no-ssl3 is incomplete.
6025 When OpenSSL is configured with "no-ssl3" as a build option, servers
6026 could accept and complete a SSL 3.0 handshake, and clients could be
6027 configured to send them.
6030 *Akamai and the OpenSSL team*
6032 * Add support for TLS_FALLBACK_SCSV.
6033 Client applications doing fallback retries should call
6034 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6037 *Adam Langley, Bodo Moeller*
6039 * Add additional DigestInfo checks.
6041 Re-encode DigestInto in DER and check against the original when
6042 verifying RSA signature: this will reject any improperly encoded
6043 DigestInfo structures.
6045 Note: this is a precautionary measure and no attacks are currently known.
6050 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014] ###
6052 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6053 SRP code can be overrun an internal buffer. Add sanity check that
6054 g, A, B < N to SRP code.
6056 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6057 Group for discovering this issue.
6062 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6063 TLS 1.0 instead of higher protocol versions when the ClientHello message
6064 is badly fragmented. This allows a man-in-the-middle attacker to force a
6065 downgrade to TLS 1.0 even if both the server and the client support a
6066 higher protocol version, by modifying the client's TLS records.
6068 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6069 researching this issue.
6074 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6075 to a denial of service attack. A malicious server can crash the client
6076 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6077 ciphersuite and sending carefully crafted handshake messages.
6079 Thanks to Felix Gröbert (Google) for discovering and researching this
6085 * By sending carefully crafted DTLS packets an attacker could cause openssl
6086 to leak memory. This can be exploited through a Denial of Service attack.
6087 Thanks to Adam Langley for discovering and researching this issue.
6092 * An attacker can force openssl to consume large amounts of memory whilst
6093 processing DTLS handshake messages. This can be exploited through a
6094 Denial of Service attack.
6095 Thanks to Adam Langley for discovering and researching this issue.
6100 * An attacker can force an error condition which causes openssl to crash
6101 whilst processing DTLS packets due to memory being freed twice. This
6102 can be exploited through a Denial of Service attack.
6103 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6109 * If a multithreaded client connects to a malicious server using a resumed
6110 session and the server sends an ec point format extension it could write
6111 up to 255 bytes to freed memory.
6113 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6119 * A malicious server can crash an OpenSSL client with a null pointer
6120 dereference (read) by specifying an SRP ciphersuite even though it was not
6121 properly negotiated with the client. This can be exploited through a
6122 Denial of Service attack.
6124 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6125 discovering and researching this issue.
6130 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6131 X509_name_oneline, X509_name_print_ex et al. to leak some information
6132 from the stack. Applications may be affected if they echo pretty printing
6133 output to the attacker.
6135 Thanks to Ivan Fratric (Google) for discovering this issue.
6138 *Emilia Käsper, and Steve Henson*
6140 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6141 for corner cases. (Certain input points at infinity could lead to
6142 bogus results, with non-infinity inputs mapped to infinity too.)
6146 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014] ###
6148 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6149 handshake can force the use of weak keying material in OpenSSL
6150 SSL/TLS clients and servers.
6152 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6153 researching this issue. [CVE-2014-0224][]
6155 *KIKUCHI Masashi, Steve Henson*
6157 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6158 OpenSSL DTLS client the code can be made to recurse eventually crashing
6161 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6164 *Imre Rad, Steve Henson*
6166 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6167 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6168 client or server. This is potentially exploitable to run arbitrary
6169 code on a vulnerable client or server.
6171 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6173 *Jüri Aedla, Steve Henson*
6175 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6176 are subject to a denial of service attack.
6178 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6179 this issue. [CVE-2014-3470][]
6181 *Felix Gröbert, Ivan Fratric, Steve Henson*
6183 * Harmonize version and its documentation. -f flag is used to display
6186 *mancha <mancha1@zoho.com>*
6188 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6189 in i2d_ECPrivateKey.
6191 *mancha <mancha1@zoho.com>*
6193 * Fix some double frees. These are not thought to be exploitable.
6195 *mancha <mancha1@zoho.com>*
6197 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014] ###
6199 * A missing bounds check in the handling of the TLS heartbeat extension
6200 can be used to reveal up to 64k of memory to a connected client or
6203 Thanks for Neel Mehta of Google Security for discovering this bug and to
6204 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6205 preparing the fix [CVE-2014-0160][]
6207 *Adam Langley, Bodo Moeller*
6209 * Fix for the attack described in the paper "Recovering OpenSSL
6210 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6211 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6212 http://eprint.iacr.org/2014/140
6214 Thanks to Yuval Yarom and Naomi Benger for discovering this
6215 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6217 *Yuval Yarom and Naomi Benger*
6219 * TLS pad extension: draft-agl-tls-padding-03
6221 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6222 TLS client Hello record length value would otherwise be > 255 and
6223 less that 512 pad with a dummy extension containing zeroes so it
6224 is at least 512 bytes long.
6227 *Adam Langley, Steve Henson*
6229 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014] ###
6231 * Fix for TLS record tampering bug. A carefully crafted invalid
6232 handshake could crash OpenSSL with a NULL pointer exception.
6233 Thanks to Anton Johansson for reporting this issues.
6236 * Keep original DTLS digest and encryption contexts in retransmission
6237 structures so we can use the previous session parameters if they need
6238 to be resent. [CVE-2013-6450][]
6242 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6243 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6244 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6245 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6246 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6247 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6249 *Rob Stradling, Adam Langley*
6251 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013] ###
6253 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6254 supporting platforms or when small records were transferred.
6256 *Andy Polyakov, Steve Henson*
6258 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013] ###
6260 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6262 This addresses the flaw in CBC record processing discovered by
6263 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6264 at: http://www.isg.rhul.ac.uk/tls/
6266 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6267 Security Group at Royal Holloway, University of London
6268 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6269 Emilia Käsper for the initial patch.
6272 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6274 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6275 ciphersuites which can be exploited in a denial of service attack.
6276 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6277 and detecting this bug and to Wolfgang Ettlinger
6278 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6283 * Return an error when checking OCSP signatures when key is NULL.
6284 This fixes a DoS attack. [CVE-2013-0166][]
6288 * Make openssl verify return errors.
6290 *Chris Palmer <palmer@google.com> and Ben Laurie*
6292 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6293 the right response is stapled. Also change SSL_get_certificate()
6294 so it returns the certificate actually sent.
6295 See http://rt.openssl.org/Ticket/Display.html?id=2836.
6297 *Rob Stradling <rob.stradling@comodo.com>*
6299 * Fix possible deadlock when decoding public keys.
6303 * Don't use TLS 1.0 record version number in initial client hello
6308 ### Changes between 1.0.1b and 1.0.1c [10 May 2012] ###
6310 * Sanity check record length before skipping explicit IV in TLS
6311 1.2, 1.1 and DTLS to fix DoS attack.
6313 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6314 fuzzing as a service testing platform.
6319 * Initialise tkeylen properly when encrypting CMS messages.
6320 Thanks to Solar Designer of Openwall for reporting this issue.
6324 * In FIPS mode don't try to use composite ciphers as they are not
6329 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012] ###
6331 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6332 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6333 mean any application compiled against OpenSSL 1.0.0 headers setting
6334 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6335 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6336 0x10000000L Any application which was previously compiled against
6337 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6338 will need to be recompiled as a result. Letting be results in
6339 inability to disable specifically TLS 1.1 and in client context,
6340 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6344 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6345 disable just protocol X, but all protocols above X *if* there are
6346 protocols *below* X still enabled. In more practical terms it means
6347 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6348 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6349 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6354 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012] ###
6356 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6357 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6358 in CRYPTO_realloc_clean.
6360 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6361 issue and to Adam Langley <agl@chromium.org> for fixing it.
6364 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6366 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6370 * Workarounds for some broken servers that "hang" if a client hello
6371 record length exceeds 255 bytes.
6373 1. Do not use record version number > TLS 1.0 in initial client
6374 hello: some (but not all) hanging servers will now work.
6375 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6376 the number of ciphers sent in the client hello. This should be
6377 set to an even number, such as 50, for example by passing:
6378 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6379 Most broken servers should now work.
6380 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6381 TLS 1.2 client support entirely.
6385 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6389 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012] ###
6391 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6392 STRING form instead of a DigestInfo.
6396 * The format used for MDC2 RSA signatures is inconsistent between EVP
6397 and the RSA_sign/RSA_verify functions. This was made more apparent when
6398 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6399 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6400 the correct format in RSA_verify so both forms transparently work.
6404 * Some servers which support TLS 1.0 can choke if we initially indicate
6405 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6406 encrypted premaster secret. As a workaround use the maximum permitted
6407 client version in client hello, this should keep such servers happy
6408 and still work with previous versions of OpenSSL.
6412 * Add support for TLS/DTLS heartbeats.
6414 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6416 * Add support for SCTP.
6418 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6420 * Improved PRNG seeding for VOS.
6422 *Paul Green <Paul.Green@stratus.com>*
6424 * Extensive assembler packs updates, most notably:
6426 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6427 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6428 - x86_64: bit-sliced AES implementation;
6429 - ARM: NEON support, contemporary platforms optimizations;
6430 - s390x: z196 support;
6431 - *: GHASH and GF(2^m) multiplication implementations;
6435 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6436 (removal of unnecessary code)
6438 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6440 * Add TLS key material exporter from RFC 5705.
6444 * Add DTLS-SRTP negotiation from RFC 5764.
6448 * Add Next Protocol Negotiation,
6449 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
6450 disabled with a no-npn flag to config or Configure. Code donated
6453 *Adam Langley <agl@google.com> and Ben Laurie*
6455 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6456 NIST-P256, NIST-P521, with constant-time single point multiplication on
6457 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6458 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6459 Code made available under Apache License version 2.0.
6461 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6462 line to include this in your build of OpenSSL, and run "make depend" (or
6463 "make update"). This enables the following EC_METHODs:
6465 EC_GFp_nistp224_method()
6466 EC_GFp_nistp256_method()
6467 EC_GFp_nistp521_method()
6469 EC_GROUP_new_by_curve_name() will automatically use these (while
6470 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6473 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6475 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6476 all platforms. Move ssize_t definition from e_os.h to the public
6477 header file e_os2.h as it now appears in public header file cms.h
6481 * New -sigopt option to the ca, req and x509 utilities. Additional
6482 signature parameters can be passed using this option and in
6487 * Add RSA PSS signing function. This will generate and set the
6488 appropriate AlgorithmIdentifiers for PSS based on those in the
6489 corresponding EVP_MD_CTX structure. No application support yet.
6493 * Support for companion algorithm specific ASN1 signing routines.
6494 New function ASN1_item_sign_ctx() signs a pre-initialised
6495 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6496 the appropriate parameters.
6500 * Add new algorithm specific ASN1 verification initialisation function
6501 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6502 handling will be the same no matter what EVP_PKEY_METHOD is used.
6503 Add a PSS handler to support verification of PSS signatures: checked
6504 against a number of sample certificates.
6508 * Add signature printing for PSS. Add PSS OIDs.
6510 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6512 * Add algorithm specific signature printing. An individual ASN1 method
6513 can now print out signatures instead of the standard hex dump.
6515 More complex signatures (e.g. PSS) can print out more meaningful
6516 information. Include DSA version that prints out the signature
6521 * Password based recipient info support for CMS library: implementing
6526 * Split password based encryption into PBES2 and PBKDF2 functions. This
6527 neatly separates the code into cipher and PBE sections and is required
6528 for some algorithms that split PBES2 into separate pieces (such as
6529 password based CMS).
6533 * Session-handling fixes:
6534 - Fix handling of connections that are resuming with a session ID,
6535 but also support Session Tickets.
6536 - Fix a bug that suppressed issuing of a new ticket if the client
6537 presented a ticket with an expired session.
6538 - Try to set the ticket lifetime hint to something reasonable.
6539 - Make tickets shorter by excluding irrelevant information.
6540 - On the client side, don't ignore renewed tickets.
6542 *Adam Langley, Bodo Moeller (Google)*
6544 * Fix PSK session representation.
6548 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6550 This work was sponsored by Intel.
6554 * Add GCM support to TLS library. Some custom code is needed to split
6555 the IV between the fixed (from PRF) and explicit (from TLS record)
6556 portions. This adds all GCM ciphersuites supported by RFC5288 and
6557 RFC5289. Generalise some AES* cipherstrings to include GCM and
6558 add a special AESGCM string for GCM only.
6562 * Expand range of ctrls for AES GCM. Permit setting invocation
6563 field on decrypt and retrieval of invocation field only on encrypt.
6567 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6568 As required by RFC5289 these ciphersuites cannot be used if for
6569 versions of TLS earlier than 1.2.
6573 * For FIPS capable OpenSSL interpret a NULL default public key method
6574 as unset and return the appropriate default but do *not* set the default.
6575 This means we can return the appropriate method in applications that
6576 switch between FIPS and non-FIPS modes.
6580 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6581 ENGINE is used then we cannot handle that in the FIPS module so we
6582 keep original code iff non-FIPS operations are allowed.
6586 * Add -attime option to openssl utilities.
6588 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6590 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6594 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6595 FIPS EC methods unconditionally for now.
6599 * New build option no-ec2m to disable characteristic 2 code.
6603 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6604 all cases can be covered as some introduce binary incompatibilities.
6608 * Redirect RSA operations to FIPS module including keygen,
6609 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6613 * Add similar low level API blocking to ciphers.
6617 * Low level digest APIs are not approved in FIPS mode: any attempt
6618 to use these will cause a fatal error. Applications that *really* want
6619 to use them can use the private_* version instead.
6623 * Redirect cipher operations to FIPS module for FIPS builds.
6627 * Redirect digest operations to FIPS module for FIPS builds.
6631 * Update build system to add "fips" flag which will link in fipscanister.o
6632 for static and shared library builds embedding a signature if needed.
6636 * Output TLS supported curves in preference order instead of numerical
6637 order. This is currently hardcoded for the highest order curves first.
6638 This should be configurable so applications can judge speed vs strength.
6642 * Add TLS v1.2 server support for client authentication.
6646 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6651 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6652 FIPS modules versions.
6656 * Add TLS v1.2 client side support for client authentication. Keep cache
6657 of handshake records longer as we don't know the hash algorithm to use
6658 until after the certificate request message is received.
6662 * Initial TLS v1.2 client support. Add a default signature algorithms
6663 extension including all the algorithms we support. Parse new signature
6664 format in client key exchange. Relax some ECC signing restrictions for
6665 TLS v1.2 as indicated in RFC5246.
6669 * Add server support for TLS v1.2 signature algorithms extension. Switch
6670 to new signature format when needed using client digest preference.
6671 All server ciphersuites should now work correctly in TLS v1.2. No client
6672 support yet and no support for client certificates.
6676 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6677 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6678 ciphersuites. At present only RSA key exchange ciphersuites work with
6679 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6680 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6681 and version checking.
6685 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6686 with this defined it will not be affected by any changes to ssl internal
6687 structures. Add several utility functions to allow openssl application
6688 to work with OPENSSL_NO_SSL_INTERN defined.
6692 * A long standing patch to add support for SRP from EdelWeb (Peter
6693 Sylvester and Christophe Renou) was integrated.
6694 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6695 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6698 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6702 * Permit abbreviated handshakes when renegotiating using the function
6703 SSL_renegotiate_abbreviated().
6705 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6707 * Add call to ENGINE_register_all_complete() to
6708 ENGINE_load_builtin_engines(), so some implementations get used
6709 automatically instead of needing explicit application support.
6713 * Add support for TLS key exporter as described in RFC5705.
6715 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6717 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6718 a few changes are required:
6720 Add SSL_OP_NO_TLSv1_1 flag.
6721 Add TLSv1_1 methods.
6722 Update version checking logic to handle version 1.1.
6723 Add explicit IV handling (ported from DTLS code).
6724 Add command line options to s_client/s_server.
6731 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015] ###
6733 * X509_ATTRIBUTE memory leak
6735 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6736 memory. This structure is used by the PKCS#7 and CMS routines so any
6737 application which reads PKCS#7 or CMS data from untrusted sources is
6738 affected. SSL/TLS is not affected.
6740 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6746 * Race condition handling PSK identify hint
6748 If PSK identity hints are received by a multi-threaded client then
6749 the values are wrongly updated in the parent SSL_CTX structure. This can
6750 result in a race condition potentially leading to a double free of the
6756 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015] ###
6758 * Malformed ECParameters causes infinite loop
6760 When processing an ECParameters structure OpenSSL enters an infinite loop
6761 if the curve specified is over a specially malformed binary polynomial
6764 This can be used to perform denial of service against any
6765 system which processes public keys, certificate requests or
6766 certificates. This includes TLS clients and TLS servers with
6767 client authentication enabled.
6769 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6774 * Exploitable out-of-bounds read in X509_cmp_time
6776 X509_cmp_time does not properly check the length of the ASN1_TIME
6777 string and can read a few bytes out of bounds. In addition,
6778 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6781 An attacker can use this to craft malformed certificates and CRLs of
6782 various sizes and potentially cause a segmentation fault, resulting in
6783 a DoS on applications that verify certificates or CRLs. TLS clients
6784 that verify CRLs are affected. TLS clients and servers with client
6785 authentication enabled may be affected if they use custom verification
6788 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6789 independently by Hanno Böck.
6794 * PKCS7 crash with missing EnvelopedContent
6796 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6797 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6798 with missing content and trigger a NULL pointer dereference on parsing.
6800 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6801 structures from untrusted sources are affected. OpenSSL clients and
6802 servers are not affected.
6804 This issue was reported to OpenSSL by Michal Zalewski (Google).
6809 * CMS verify infinite loop with unknown hash function
6811 When verifying a signedData message the CMS code can enter an infinite loop
6812 if presented with an unknown hash function OID. This can be used to perform
6813 denial of service against any system which verifies signedData messages using
6815 This issue was reported to OpenSSL by Johannes Bauer.
6820 * Race condition handling NewSessionTicket
6822 If a NewSessionTicket is received by a multi-threaded client when attempting to
6823 reuse a previous ticket then a race condition can occur potentially leading to
6824 a double free of the ticket data.
6829 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015] ###
6831 * Segmentation fault in ASN1_TYPE_cmp fix
6833 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6834 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6835 certificate signature algorithm consistency this can be used to crash any
6836 certificate verification operation and exploited in a DoS attack. Any
6837 application which performs certificate verification is vulnerable including
6838 OpenSSL clients and servers which enable client authentication.
6843 * ASN.1 structure reuse memory corruption fix
6845 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6846 memory corruption via an invalid write. Such reuse is and has been
6847 strongly discouraged and is believed to be rare.
6849 Applications that parse structures containing CHOICE or ANY DEFINED BY
6850 components may be affected. Certificate parsing (d2i_X509 and related
6851 functions) are however not affected. OpenSSL clients and servers are
6857 * PKCS7 NULL pointer dereferences fix
6859 The PKCS#7 parsing code does not handle missing outer ContentInfo
6860 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6861 missing content and trigger a NULL pointer dereference on parsing.
6863 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6864 otherwise parse PKCS#7 structures from untrusted sources are
6865 affected. OpenSSL clients and servers are not affected.
6867 This issue was reported to OpenSSL by Michal Zalewski (Google).
6872 * DoS via reachable assert in SSLv2 servers fix
6874 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6875 servers that both support SSLv2 and enable export cipher suites by sending
6876 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6878 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6879 (OpenSSL development team).
6884 * Use After Free following d2i_ECPrivatekey error fix
6886 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6887 could cause a use after free condition. This, in turn, could cause a double
6888 free in several private key parsing functions (such as d2i_PrivateKey
6889 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6890 for applications that receive EC private keys from untrusted
6891 sources. This scenario is considered rare.
6893 This issue was discovered by the BoringSSL project and fixed in their
6899 * X509_to_X509_REQ NULL pointer deref fix
6901 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6902 the certificate key is invalid. This function is rarely used in practice.
6904 This issue was discovered by Brian Carpenter.
6909 * Removed the export ciphers from the DEFAULT ciphers
6913 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015] ###
6915 * Build fixes for the Windows and OpenVMS platforms
6917 *Matt Caswell and Richard Levitte*
6919 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015] ###
6921 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6922 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6923 dereference. This could lead to a Denial Of Service attack. Thanks to
6924 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6929 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6930 dtls1_buffer_record function under certain conditions. In particular this
6931 could occur if an attacker sent repeated DTLS records with the same
6932 sequence number but for the next epoch. The memory leak could be exploited
6933 by an attacker in a Denial of Service attack through memory exhaustion.
6934 Thanks to Chris Mueller for reporting this issue.
6939 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6940 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6941 method would be set to NULL which could later result in a NULL pointer
6942 dereference. Thanks to Frank Schmirler for reporting this issue.
6947 * Abort handshake if server key exchange message is omitted for ephemeral
6950 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6951 reporting this issue.
6956 * Remove non-export ephemeral RSA code on client and server. This code
6957 violated the TLS standard by allowing the use of temporary RSA keys in
6958 non-export ciphersuites and could be used by a server to effectively
6959 downgrade the RSA key length used to a value smaller than the server
6960 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6961 INRIA or reporting this issue.
6966 * Fixed issue where DH client certificates are accepted without verification.
6967 An OpenSSL server will accept a DH certificate for client authentication
6968 without the certificate verify message. This effectively allows a client to
6969 authenticate without the use of a private key. This only affects servers
6970 which trust a client certificate authority which issues certificates
6971 containing DH keys: these are extremely rare and hardly ever encountered.
6972 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6978 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6979 results on some platforms, including x86_64. This bug occurs at random
6980 with a very low probability, and is not known to be exploitable in any
6981 way, though its exact impact is difficult to determine. Thanks to Pieter
6982 Wuille (Blockstream) who reported this issue and also suggested an initial
6983 fix. Further analysis was conducted by the OpenSSL development team and
6984 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6985 the OpenSSL core team.
6990 * Fix various certificate fingerprint issues.
6992 By using non-DER or invalid encodings outside the signed portion of a
6993 certificate the fingerprint can be changed without breaking the signature.
6994 Although no details of the signed portion of the certificate can be changed
6995 this can cause problems with some applications: e.g. those using the
6996 certificate fingerprint for blacklists.
6998 1. Reject signatures with non zero unused bits.
7000 If the BIT STRING containing the signature has non zero unused bits reject
7001 the signature. All current signature algorithms require zero unused bits.
7003 2. Check certificate algorithm consistency.
7005 Check the AlgorithmIdentifier inside TBS matches the one in the
7006 certificate signature. NB: this will result in signature failure
7007 errors for some broken certificates.
7009 Thanks to Konrad Kraszewski from Google for reporting this issue.
7011 3. Check DSA/ECDSA signatures use DER.
7013 Reencode DSA/ECDSA signatures and compare with the original received
7014 signature. Return an error if there is a mismatch.
7016 This will reject various cases including garbage after signature
7017 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7018 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7019 (negative or with leading zeroes).
7021 Further analysis was conducted and fixes were developed by Stephen Henson
7022 of the OpenSSL core team.
7028 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014] ###
7030 * Session Ticket Memory Leak.
7032 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7033 integrity of that ticket is first verified. In the event of a session
7034 ticket integrity check failing, OpenSSL will fail to free memory
7035 causing a memory leak. By sending a large number of invalid session
7036 tickets an attacker could exploit this issue in a Denial Of Service
7042 * Build option no-ssl3 is incomplete.
7044 When OpenSSL is configured with "no-ssl3" as a build option, servers
7045 could accept and complete a SSL 3.0 handshake, and clients could be
7046 configured to send them.
7049 *Akamai and the OpenSSL team*
7051 * Add support for TLS_FALLBACK_SCSV.
7052 Client applications doing fallback retries should call
7053 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7056 *Adam Langley, Bodo Moeller*
7058 * Add additional DigestInfo checks.
7060 Reencode DigestInto in DER and check against the original when
7061 verifying RSA signature: this will reject any improperly encoded
7062 DigestInfo structures.
7064 Note: this is a precautionary measure and no attacks are currently known.
7069 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014] ###
7071 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7072 to a denial of service attack. A malicious server can crash the client
7073 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7074 ciphersuite and sending carefully crafted handshake messages.
7076 Thanks to Felix Gröbert (Google) for discovering and researching this
7082 * By sending carefully crafted DTLS packets an attacker could cause openssl
7083 to leak memory. This can be exploited through a Denial of Service attack.
7084 Thanks to Adam Langley for discovering and researching this issue.
7089 * An attacker can force openssl to consume large amounts of memory whilst
7090 processing DTLS handshake messages. This can be exploited through a
7091 Denial of Service attack.
7092 Thanks to Adam Langley for discovering and researching this issue.
7097 * An attacker can force an error condition which causes openssl to crash
7098 whilst processing DTLS packets due to memory being freed twice. This
7099 can be exploited through a Denial of Service attack.
7100 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7106 * If a multithreaded client connects to a malicious server using a resumed
7107 session and the server sends an ec point format extension it could write
7108 up to 255 bytes to freed memory.
7110 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7116 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7117 X509_name_oneline, X509_name_print_ex et al. to leak some information
7118 from the stack. Applications may be affected if they echo pretty printing
7119 output to the attacker.
7121 Thanks to Ivan Fratric (Google) for discovering this issue.
7124 *Emilia Käsper, and Steve Henson*
7126 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7127 for corner cases. (Certain input points at infinity could lead to
7128 bogus results, with non-infinity inputs mapped to infinity too.)
7132 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014] ###
7134 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7135 handshake can force the use of weak keying material in OpenSSL
7136 SSL/TLS clients and servers.
7138 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7139 researching this issue. [CVE-2014-0224][]
7141 *KIKUCHI Masashi, Steve Henson*
7143 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7144 OpenSSL DTLS client the code can be made to recurse eventually crashing
7147 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7150 *Imre Rad, Steve Henson*
7152 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7153 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7154 client or server. This is potentially exploitable to run arbitrary
7155 code on a vulnerable client or server.
7157 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7159 *Jüri Aedla, Steve Henson*
7161 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7162 are subject to a denial of service attack.
7164 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7165 this issue. [CVE-2014-3470][]
7167 *Felix Gröbert, Ivan Fratric, Steve Henson*
7169 * Harmonize version and its documentation. -f flag is used to display
7172 *mancha <mancha1@zoho.com>*
7174 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7175 in i2d_ECPrivateKey.
7177 *mancha <mancha1@zoho.com>*
7179 * Fix some double frees. These are not thought to be exploitable.
7181 *mancha <mancha1@zoho.com>*
7183 * Fix for the attack described in the paper "Recovering OpenSSL
7184 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7185 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7186 http://eprint.iacr.org/2014/140
7188 Thanks to Yuval Yarom and Naomi Benger for discovering this
7189 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7191 *Yuval Yarom and Naomi Benger*
7193 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014] ###
7195 * Keep original DTLS digest and encryption contexts in retransmission
7196 structures so we can use the previous session parameters if they need
7197 to be resent. [CVE-2013-6450][]
7201 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7202 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7203 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7204 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7205 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7206 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7208 *Rob Stradling, Adam Langley*
7210 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013] ###
7212 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7214 This addresses the flaw in CBC record processing discovered by
7215 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7216 at: http://www.isg.rhul.ac.uk/tls/
7218 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7219 Security Group at Royal Holloway, University of London
7220 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7221 Emilia Käsper for the initial patch.
7224 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7226 * Return an error when checking OCSP signatures when key is NULL.
7227 This fixes a DoS attack. [CVE-2013-0166][]
7231 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7232 the right response is stapled. Also change SSL_get_certificate()
7233 so it returns the certificate actually sent.
7234 See http://rt.openssl.org/Ticket/Display.html?id=2836.
7235 (This is a backport)
7237 *Rob Stradling <rob.stradling@comodo.com>*
7239 * Fix possible deadlock when decoding public keys.
7243 ### Changes between 1.0.0i and 1.0.0j [10 May 2012] ###
7245 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7248 * Sanity check record length before skipping explicit IV in DTLS
7251 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7252 fuzzing as a service testing platform.
7257 * Initialise tkeylen properly when encrypting CMS messages.
7258 Thanks to Solar Designer of Openwall for reporting this issue.
7262 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012] ###
7264 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7265 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7266 in CRYPTO_realloc_clean.
7268 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7269 issue and to Adam Langley <agl@chromium.org> for fixing it.
7272 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7274 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012] ###
7276 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7277 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7278 content decryption and always return the same error. Note: this attack
7279 needs on average 2^20 messages so it only affects automated senders. The
7280 old behaviour can be re-enabled in the CMS code by setting the
7281 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7282 an MMA defence is not necessary.
7283 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7284 this issue. [CVE-2012-0884][]
7288 * Fix CVE-2011-4619: make sure we really are receiving a
7289 client hello before rejecting multiple SGC restarts. Thanks to
7290 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7294 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012] ###
7296 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7297 Thanks to Antonio Martin, Enterprise Secure Access Research and
7298 Development, Cisco Systems, Inc. for discovering this bug and
7299 preparing a fix. [CVE-2012-0050][]
7303 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012] ###
7305 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7306 of the Vaudenay padding oracle attack on CBC mode encryption
7307 which enables an efficient plaintext recovery attack against
7308 the OpenSSL implementation of DTLS. Their attack exploits timing
7309 differences arising during decryption processing. A research
7310 paper describing this attack can be found at:
7311 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
7312 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7313 Security Group at Royal Holloway, University of London
7314 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7315 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7316 for preparing the fix. [CVE-2011-4108][]
7318 *Robin Seggelmann, Michael Tuexen*
7320 * Clear bytes used for block padding of SSL 3.0 records.
7323 *Adam Langley (Google)*
7325 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7326 Kadianakis <desnacked@gmail.com> for discovering this issue and
7327 Adam Langley for preparing the fix. [CVE-2011-4619][]
7329 *Adam Langley (Google)*
7331 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7333 *Andrey Kulikov <amdeich@gmail.com>*
7335 * Prevent malformed RFC3779 data triggering an assertion failure.
7336 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7337 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7339 *Rob Austein <sra@hactrn.net>*
7341 * Improved PRNG seeding for VOS.
7343 *Paul Green <Paul.Green@stratus.com>*
7345 * Fix ssl_ciph.c set-up race.
7347 *Adam Langley (Google)*
7349 * Fix spurious failures in ecdsatest.c.
7351 *Emilia Käsper (Google)*
7353 * Fix the BIO_f_buffer() implementation (which was mixing different
7354 interpretations of the '..._len' fields).
7356 *Adam Langley (Google)*
7358 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7359 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7360 threads won't reuse the same blinding coefficients.
7362 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7363 lock to call BN_BLINDING_invert_ex, and avoids one use of
7364 BN_BLINDING_update for each BN_BLINDING structure (previously,
7365 the last update always remained unused).
7367 *Emilia Käsper (Google)*
7369 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7371 *Bob Buckholz (Google)*
7373 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011] ###
7375 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7376 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7378 *Kaspar Brand <ossl@velox.ch>*
7380 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7381 for multi-threaded use of ECDH. [CVE-2011-3210][]
7383 *Adam Langley (Google)*
7385 * Fix x509_name_ex_d2i memory leak on bad inputs.
7389 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7390 signature public key algorithm by using OID xref utilities instead.
7391 Before this you could only use some ECC ciphersuites with SHA1 only.
7395 * Add protection against ECDSA timing attacks as mentioned in the paper
7396 by Billy Bob Brumley and Nicola Tuveri, see:
7398 http://eprint.iacr.org/2011/232.pdf
7401 *Billy Bob Brumley and Nicola Tuveri*
7403 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011] ###
7405 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7407 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7409 * Fix bug in string printing code: if *any* escaping is enabled we must
7410 escape the escape character (backslash) or the resulting string is
7415 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010] ###
7417 * Disable code workaround for ancient and obsolete Netscape browsers
7418 and servers: an attacker can use it in a ciphersuite downgrade attack.
7419 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7423 * Fixed J-PAKE implementation error, originally discovered by
7424 Sebastien Martini, further info and confirmation from Stefan
7425 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7429 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010] ###
7431 * Fix extension code to avoid race conditions which can result in a buffer
7432 overrun vulnerability: resumed sessions must not be modified as they can
7433 be shared by multiple threads. CVE-2010-3864
7437 * Fix WIN32 build system to correctly link an ENGINE directory into
7442 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010] ###
7444 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7447 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7449 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010] ###
7451 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7452 context. The operation can be customised via the ctrl mechanism in
7453 case ENGINEs want to include additional functionality.
7457 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7461 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7462 output hashes compatible with older versions of OpenSSL.
7464 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7466 * Fix compression algorithm handling: if resuming a session use the
7467 compression algorithm of the resumed session instead of determining
7468 it from client hello again. Don't allow server to change algorithm.
7472 * Add load_crls() function to apps tidying load_certs() too. Add option
7473 to verify utility to allow additional CRLs to be included.
7477 * Update OCSP request code to permit adding custom headers to the request:
7478 some responders need this.
7482 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7485 *Julia Lawall <julia@diku.dk>*
7487 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7488 needlessly dereferenced structures, used obsolete functions and
7489 didn't handle all updated verify codes correctly.
7493 * Disable MD2 in the default configuration.
7497 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7498 indicate the initial BIO being pushed or popped. This makes it possible
7499 to determine whether the BIO is the one explicitly called or as a result
7500 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7501 it handles reference counts correctly and doesn't zero out the I/O bio
7502 when it is not being explicitly popped. WARNING: applications which
7503 included workarounds for the old buggy behaviour will need to be modified
7504 or they could free up already freed BIOs.
7508 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7509 renaming to all platforms (within the 0.9.8 branch, this was
7510 done conditionally on Netware platforms to avoid a name clash).
7512 *Guenter <lists@gknw.net>*
7514 * Add ECDHE and PSK support to DTLS.
7516 *Michael Tuexen <tuexen@fh-muenster.de>*
7518 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7523 * Add "missing" function EVP_MD_flags() (without this the only way to
7524 retrieve a digest flags is by accessing the structure directly. Update
7525 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
7526 or cipher is registered as in the "from" argument. Print out all
7527 registered digests in the dgst usage message instead of manually
7528 attempting to work them out.
7532 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7533 this allows the use of compression and extensions. Change default cipher
7534 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7535 by default unless an application cipher string requests it.
7539 * Alter match criteria in PKCS12_parse(). It used to try to use local
7540 key ids to find matching certificates and keys but some PKCS#12 files
7541 don't follow the (somewhat unwritten) rules and this strategy fails.
7542 Now just gather all certificates together and the first private key
7543 then look for the first certificate that matches the key.
7547 * Support use of registered digest and cipher names for dgst and cipher
7548 commands instead of having to add each one as a special case. So now
7555 openssl dgst -sha256 foo
7557 and this works for ENGINE based algorithms too.
7562 * Update Gost ENGINE to support parameter files.
7564 *Victor B. Wagner <vitus@cryptocom.ru>*
7566 * Support GeneralizedTime in ca utility.
7568 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7570 * Enhance the hash format used for certificate directory links. The new
7571 form uses the canonical encoding (meaning equivalent names will work
7572 even if they aren't identical) and uses SHA1 instead of MD5. This form
7573 is incompatible with the older format and as a result c_rehash should
7574 be used to rebuild symbolic links.
7578 * Make PKCS#8 the default write format for private keys, replacing the
7579 traditional format. This form is standardised, more secure and doesn't
7580 include an implicit MD5 dependency.
7584 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7585 committed to OpenSSL should pass this lot as a minimum.
7589 * Add session ticket override functionality for use by EAP-FAST.
7591 *Jouni Malinen <j@w1.fi>*
7593 * Modify HMAC functions to return a value. Since these can be implemented
7594 in an ENGINE errors can occur.
7598 * Type-checked OBJ_bsearch_ex.
7602 * Type-checked OBJ_bsearch. Also some constification necessitated
7603 by type-checking. Still to come: TXT_DB, bsearch(?),
7604 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7609 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7610 seconds to a tm structure directly, instead of going through OS
7611 specific date routines. This avoids any issues with OS routines such
7612 as the year 2038 bug. New *_adj() functions for ASN1 time structures
7613 and X509_time_adj_ex() to cover the extended range. The existing
7614 X509_time_adj() is still usable and will no longer have any date issues.
7618 * Delta CRL support. New use deltas option which will attempt to locate
7619 and search any appropriate delta CRLs available.
7621 This work was sponsored by Google.
7625 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7626 code and add additional score elements. Validate alternate CRL paths
7627 as part of the CRL checking and indicate a new error "CRL path validation
7628 error" in this case. Applications wanting additional details can use
7629 the verify callback and check the new "parent" field. If this is not
7630 NULL CRL path validation is taking place. Existing applications won't
7631 see this because it requires extended CRL support which is off by
7634 This work was sponsored by Google.
7638 * Support for freshest CRL extension.
7640 This work was sponsored by Google.
7644 * Initial indirect CRL support. Currently only supported in the CRLs
7645 passed directly and not via lookup. Process certificate issuer
7646 CRL entry extension and lookup CRL entries by bother issuer name
7647 and serial number. Check and process CRL issuer entry in IDP extension.
7649 This work was sponsored by Google.
7653 * Add support for distinct certificate and CRL paths. The CRL issuer
7654 certificate is validated separately in this case. Only enabled if
7655 an extended CRL support flag is set: this flag will enable additional
7656 CRL functionality in future.
7658 This work was sponsored by Google.
7662 * Add support for policy mappings extension.
7664 This work was sponsored by Google.
7668 * Fixes to pathlength constraint, self issued certificate handling,
7669 policy processing to align with RFC3280 and PKITS tests.
7671 This work was sponsored by Google.
7675 * Support for name constraints certificate extension. DN, email, DNS
7676 and URI types are currently supported.
7678 This work was sponsored by Google.
7682 * To cater for systems that provide a pointer-based thread ID rather
7683 than numeric, deprecate the current numeric thread ID mechanism and
7684 replace it with a structure and associated callback type. This
7685 mechanism allows a numeric "hash" to be extracted from a thread ID in
7686 either case, and on platforms where pointers are larger than 'long',
7687 mixing is done to help ensure the numeric 'hash' is usable even if it
7688 can't be guaranteed unique. The default mechanism is to use "&errno"
7689 as a pointer-based thread ID to distinguish between threads.
7691 Applications that want to provide their own thread IDs should now use
7692 CRYPTO_THREADID_set_callback() to register a callback that will call
7693 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7695 Note that ERR_remove_state() is now deprecated, because it is tied
7696 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7697 to free the current thread's error state should be replaced by
7698 ERR_remove_thread_state(NULL).
7700 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7701 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7702 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7703 application was previously providing a numeric thread callback that
7704 was inappropriate for distinguishing threads, then uniqueness might
7705 have been obtained with &errno that happened immediately in the
7706 intermediate development versions of OpenSSL; this is no longer the
7707 case, the numeric thread callback will now override the automatic use
7710 *Geoff Thorpe, with help from Bodo Moeller*
7712 * Initial support for different CRL issuing certificates. This covers a
7713 simple case where the self issued certificates in the chain exist and
7714 the real CRL issuer is higher in the existing chain.
7716 This work was sponsored by Google.
7720 * Removed effectively defunct crypto/store from the build.
7724 * Revamp of STACK to provide stronger type-checking. Still to come:
7725 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7726 ASN1_STRING, CONF_VALUE.
7730 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7731 RAM on SSL connections. This option can save about 34k per idle SSL.
7735 * Revamp of LHASH to provide stronger type-checking. Still to come:
7736 STACK, TXT_DB, bsearch, qsort.
7740 * Initial support for Cryptographic Message Syntax (aka CMS) based
7741 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7742 support for data, signedData, compressedData, digestedData and
7743 encryptedData, envelopedData types included. Scripts to check against
7744 RFC4134 examples draft and interop and consistency checks of many
7745 content types and variants.
7749 * Add options to enc utility to support use of zlib compression BIO.
7753 * Extend mk1mf to support importing of options and assembly language
7754 files from Configure script, currently only included in VC-WIN32.
7755 The assembly language rules can now optionally generate the source
7756 files from the associated perl scripts.
7760 * Implement remaining functionality needed to support GOST ciphersuites.
7761 Interop testing has been performed using CryptoPro implementations.
7763 *Victor B. Wagner <vitus@cryptocom.ru>*
7765 * s390x assembler pack.
7769 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7774 * Implement Opaque PRF Input TLS extension as specified in
7775 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7776 official specification yet and no extension type assignment by
7777 IANA exists, this extension (for now) will have to be explicitly
7778 enabled when building OpenSSL by providing the extension number
7779 to use. For example, specify an option
7781 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7783 to the "config" or "Configure" script to enable the extension,
7784 assuming extension number 0x9527 (which is a completely arbitrary
7785 and unofficial assignment based on the MD5 hash of the Internet
7786 Draft). Note that by doing so, you potentially lose
7787 interoperability with other TLS implementations since these might
7788 be using the same extension number for other purposes.
7790 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7791 opaque PRF input value to use in the handshake. This will create
7792 an internal copy of the length-'len' string at 'src', and will
7793 return non-zero for success.
7795 To get more control and flexibility, provide a callback function
7798 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7799 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7803 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7806 Callback function 'cb' will be called in handshakes, and is
7807 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7808 Argument 'arg' is for application purposes (the value as given to
7809 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7810 be provided to the callback function). The callback function
7811 has to return non-zero to report success: usually 1 to use opaque
7812 PRF input just if possible, or 2 to enforce use of the opaque PRF
7813 input. In the latter case, the library will abort the handshake
7814 if opaque PRF input is not successfully negotiated.
7816 Arguments 'peerinput' and 'len' given to the callback function
7817 will always be NULL and 0 in the case of a client. A server will
7818 see the client's opaque PRF input through these variables if
7819 available (NULL and 0 otherwise). Note that if the server
7820 provides an opaque PRF input, the length must be the same as the
7821 length of the client's opaque PRF input.
7823 Note that the callback function will only be called when creating
7824 a new session (session resumption can resume whatever was
7825 previously negotiated), and will not be called in SSL 2.0
7826 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7827 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7828 for applications that need to enforce opaque PRF input.
7833 * Update ssl code to support digests other than SHA1+MD5 for handshake
7837 *Victor B. Wagner <vitus@cryptocom.ru>*
7839 * Add RFC4507 support to OpenSSL. This includes the corrections in
7840 RFC4507bis. The encrypted ticket format is an encrypted encoded
7841 SSL_SESSION structure, that way new session features are automatically
7844 If a client application caches session in an SSL_SESSION structure
7845 support is transparent because tickets are now stored in the encoded
7848 The SSL_CTX structure automatically generates keys for ticket
7849 protection in servers so again support should be possible
7850 with no application modification.
7852 If a client or server wishes to disable RFC4507 support then the option
7853 SSL_OP_NO_TICKET can be set.
7855 Add a TLS extension debugging callback to allow the contents of any client
7856 or server extensions to be examined.
7858 This work was sponsored by Google.
7862 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7863 OpenSSL should now compile cleanly on gcc 4.2
7865 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7867 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7868 support including streaming MAC support: this is required for GOST
7869 ciphersuite support.
7871 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7873 * Add option -stream to use PKCS#7 streaming in smime utility. New
7874 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7875 to output in BER and PEM format.
7879 * Experimental support for use of HMAC via EVP_PKEY interface. This
7880 allows HMAC to be handled via the EVP_DigestSign*() interface. The
7881 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7882 ENGINE support for HMAC keys which are unextractable. New -mac and
7883 -macopt options to dgst utility.
7887 * New option -sigopt to dgst utility. Update dgst to use
7888 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
7889 alternative signing parameters such as X9.31 or PSS in the dgst
7894 * Change ssl_cipher_apply_rule(), the internal function that does
7895 the work each time a ciphersuite string requests enabling
7896 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7897 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7898 the order of disabled ciphersuites such that those ciphersuites
7899 that most recently went from enabled to disabled not only stay
7900 in order with respect to each other, but also have higher priority
7901 than other disabled ciphersuites the next time ciphersuites are
7904 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
7905 the same ciphersuites as with "HIGH" alone, but in a specific
7906 order where the PSK ciphersuites come first (since they are the
7907 most recently disabled ciphersuites when "HIGH" is parsed).
7909 Also, change ssl_create_cipher_list() (using this new
7910 functionality) such that between otherwise identical
7911 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
7916 * Change ssl_create_cipher_list() so that it automatically
7917 arranges the ciphersuites in reasonable order before starting
7918 to process the rule string. Thus, the definition for "DEFAULT"
7919 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
7920 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
7921 This makes it much easier to arrive at a reasonable default order
7922 in applications for which anonymous ciphers are OK (meaning
7923 that you can't actually use DEFAULT).
7925 *Bodo Moeller; suggested by Victor Duchovni*
7927 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
7928 processing) into multiple integers instead of setting
7929 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
7930 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
7931 (These masks as well as the individual bit definitions are hidden
7932 away into the non-exported interface ssl/ssl_locl.h, so this
7933 change to the definition of the SSL_CIPHER structure shouldn't
7934 affect applications.) This give us more bits for each of these
7935 categories, so there is no longer a need to coagulate AES128 and
7936 AES256 into a single algorithm bit, and to coagulate Camellia128
7937 and Camellia256 into a single algorithm bit, which has led to all
7940 Thus, among other things, the kludge introduced in 0.9.7m and
7941 0.9.8e for masking out AES256 independently of AES128 or masking
7942 out Camellia256 independently of AES256 is not needed here in 0.9.9.
7944 With the change, we also introduce new ciphersuite aliases that
7945 so far were missing: "AES128", "AES256", "CAMELLIA128", and
7950 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
7951 Use the leftmost N bytes of the signature input if the input is
7952 larger than the prime q (with N being the size in bytes of q).
7956 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
7957 it yet and it is largely untested.
7961 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
7965 * Initial incomplete changes to avoid need for function casts in OpenSSL
7966 some compilers (gcc 4.2 and later) reject their use. Safestack is
7967 reimplemented. Update ASN1 to avoid use of legacy functions.
7971 * Win32/64 targets are linked with Winsock2.
7975 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
7976 to external functions. This can be used to increase CRL handling
7977 efficiency especially when CRLs are very large by (for example) storing
7978 the CRL revoked certificates in a database.
7982 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
7983 new CRLs added to a directory can be used. New command line option
7984 -verify_return_error to s_client and s_server. This causes real errors
7985 to be returned by the verify callback instead of carrying on no matter
7986 what. This reflects the way a "real world" verify callback would behave.
7990 * GOST engine, supporting several GOST algorithms and public key formats.
7991 Kindly donated by Cryptocom.
7995 * Partial support for Issuing Distribution Point CRL extension. CRLs
7996 partitioned by DP are handled but no indirect CRL or reason partitioning
7997 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
7998 selected via a scoring technique which handles IDP and AKID in CRLs.
8002 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8003 will ultimately be used for all verify operations: this will remove the
8004 X509_STORE dependency on certificate verification and allow alternative
8005 lookup methods. X509_STORE based implementations of these two callbacks.
8009 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8010 Modify get_crl() to find a valid (unexpired) CRL if possible.
8014 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8015 this would be called X509_CRL_cmp() but that name is already used by
8016 a function that just compares CRL issuer names. Cache several CRL
8017 extensions in X509_CRL structure and cache CRLDP in X509.
8021 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8022 this maps equivalent X509_NAME structures into a consistent structure.
8023 Name comparison can then be performed rapidly using memcmp().
8027 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8032 * Allow digests to supply their own micalg string for S/MIME type using
8033 the ctrl EVP_MD_CTRL_MICALG.
8037 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8038 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8039 ctrl. It can then customise the structure before and/or after signing
8044 * New function OBJ_add_sigid() to allow application defined signature OIDs
8045 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8046 to free up any added signature OIDs.
8050 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8051 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8052 digest and cipher tables. New options added to openssl utility:
8053 list-message-digest-algorithms and list-cipher-algorithms.
8057 * Change the array representation of binary polynomials: the list
8058 of degrees of non-zero coefficients is now terminated with -1.
8059 Previously it was terminated with 0, which was also part of the
8060 value; thus, the array representation was not applicable to
8061 polynomials where t^0 has coefficient zero. This change makes
8062 the array representation useful in a more general context.
8066 * Various modifications and fixes to SSL/TLS cipher string
8067 handling. For ECC, the code now distinguishes between fixed ECDH
8068 with RSA certificates on the one hand and with ECDSA certificates
8069 on the other hand, since these are separate ciphersuites. The
8070 unused code for Fortezza ciphersuites has been removed.
8072 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8073 (not "ECDHE"). For consistency with the code for DH
8074 certificates, use of ECDH certificates is now considered ECDH
8075 authentication, not RSA or ECDSA authentication (the latter is
8076 merely the CA's signing algorithm and not actively used in the
8079 The temporary ciphersuite alias "ECCdraft" is no longer
8080 available, and ECC ciphersuites are no longer excluded from "ALL"
8081 and "DEFAULT". The following aliases now exist for RFC 4492
8082 ciphersuites, most of these by analogy with the DH case:
8084 kECDHr - ECDH cert, signed with RSA
8085 kECDHe - ECDH cert, signed with ECDSA
8086 kECDH - ECDH cert (signed with either RSA or ECDSA)
8087 kEECDH - ephemeral ECDH
8088 ECDH - ECDH cert or ephemeral ECDH
8094 AECDH - anonymous ECDH
8095 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8100 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8101 Use correct micalg parameters depending on digest(s) in signed message.
8105 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8106 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8110 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8111 an engine to register a method. Add ENGINE lookups for methods and
8112 functional reference processing.
8116 * New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
8117 EVP_{Sign,Verify}* which allow an application to customise the signature
8122 * New -resign option to smime utility. This adds one or more signers
8123 to an existing PKCS#7 signedData structure. Also -md option to use an
8124 alternative message digest algorithm for signing.
8128 * Tidy up PKCS#7 routines and add new functions to make it easier to
8129 create PKCS7 structures containing multiple signers. Update smime
8130 application to support multiple signers.
8134 * New -macalg option to pkcs12 utility to allow setting of an alternative
8139 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8140 Reorganize PBE internals to lookup from a static table using NIDs,
8141 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8142 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8143 PRF which will be automatically used with PBES2.
8147 * Replace the algorithm specific calls to generate keys in "req" with the
8152 * Update PKCS#7 enveloped data routines to use new API. This is now
8153 supported by any public key method supporting the encrypt operation. A
8154 ctrl is added to allow the public key algorithm to examine or modify
8155 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8160 * Add a ctrl to asn1 method to allow a public key algorithm to express
8161 a default digest type to use. In most cases this will be SHA1 but some
8162 algorithms (such as GOST) need to specify an alternative digest. The
8163 return value indicates how strong the preference is 1 means optional and
8164 2 is mandatory (that is it is the only supported type). Modify
8165 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8166 use the default md. Update openssl utilities to use the default digest
8167 type for signing if it is not explicitly indicated.
8171 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8172 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8173 signing method from the key type. This effectively removes the link
8174 between digests and public key types.
8178 * Add an OID cross reference table and utility functions. Its purpose is to
8179 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8180 rsaEncryption. This will allow some of the algorithm specific hackery
8181 needed to use the correct OID to be removed.
8185 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8186 structures for PKCS7_sign(). They are now set up by the relevant public
8191 * Add provisional EC pkey method with support for ECDSA and ECDH.
8195 * Add support for key derivation (agreement) in the API, DH method and
8200 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8201 public and private key formats. As a side effect these add additional
8202 command line functionality not previously available: DSA signatures can be
8203 generated and verified using pkeyutl and DH key support and generation in
8210 *Oliver Tappe <zooey@hirschkaefer.de>*
8212 * New make target "install_html_docs" installs HTML renditions of the
8215 *Oliver Tappe <zooey@hirschkaefer.de>*
8217 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8218 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8219 support key and parameter generation and add initial key generation
8220 functionality for RSA.
8224 * Add functions for main EVP_PKEY_method operations. The undocumented
8225 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
8226 EVP_PKEY_{encrypt,decrypt}_old.
8230 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8231 key API, doesn't do much yet.
8235 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8236 public key algorithms. New option to openssl utility:
8237 "list-public-key-algorithms" to print out info.
8241 * Implement the Supported Elliptic Curves Extension for
8242 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8246 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8247 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8251 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8252 utilities such as rsa, dsa, dsaparam etc except they process any key
8257 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8258 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8259 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8264 * Initial support for pluggable public key ASN1.
8265 De-spaghettify the public key ASN1 handling. Move public and private
8266 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8267 algorithm specific handling to a single module within the relevant
8268 algorithm directory. Add functions to allow (near) opaque processing
8269 of public and private key structures.
8273 * Implement the Supported Point Formats Extension for
8274 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8278 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8279 for the psk identity [hint] and the psk callback functions to the
8280 SSL_SESSION, SSL and SSL_CTX structure.
8283 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8287 SSL_CTX_use_psk_identity_hint
8288 SSL_get_psk_identity_hint
8289 SSL_get_psk_identity
8290 SSL_use_psk_identity_hint
8293 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8295 * Add RFC 3161 compliant time stamp request creation, response generation
8296 and response verification functionality.
8298 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8300 * Add initial support for TLS extensions, specifically for the server_name
8301 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8302 have new members for a host name. The SSL data structure has an
8303 additional member SSL_CTX *initial_ctx so that new sessions can be
8304 stored in that context to allow for session resumption, even after the
8305 SSL has been switched to a new SSL_CTX in reaction to a client's
8306 server_name extension.
8308 New functions (subject to change):
8310 SSL_get_servername()
8311 SSL_get_servername_type()
8314 New CTRL codes and macros (subject to change):
8316 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8317 - SSL_CTX_set_tlsext_servername_callback()
8318 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8319 - SSL_CTX_set_tlsext_servername_arg()
8320 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8322 openssl s_client has a new '-servername ...' option.
8324 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8325 '-key2 ...', '-servername_fatal' (subject to change). This allows
8326 testing the HostName extension for a specific single host name ('-cert'
8327 and '-key' remain fallbacks for handshakes without HostName
8328 negotiation). If the unrecognized_name alert has to be sent, this by
8329 default is a warning; it becomes fatal with the '-servername_fatal'
8333 *Peter Sylvester, Remy Allais, Christophe Renou*
8335 * Whirlpool hash implementation is added.
8339 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8340 bn(64,32). Because of instruction set limitations it doesn't have
8341 any negative impact on performance. This was done mostly in order
8342 to make it possible to share assembler modules, such as bn_mul_mont
8343 implementations, between 32- and 64-bit builds without hassle.
8347 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8348 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8353 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8354 dedicated Montgomery multiplication procedure, is introduced.
8355 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8356 "64-bit" performance on certain 32-bit targets.
8360 * New option SSL_OP_NO_COMP to disable use of compression selectively
8361 in SSL structures. New SSL ctrl to set maximum send fragment size.
8362 Save memory by setting the I/O buffer sizes dynamically instead of
8363 using the maximum available value.
8367 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8368 in addition to the text details.
8372 * Very, very preliminary EXPERIMENTAL support for printing of general
8373 ASN1 structures. This currently produces rather ugly output and doesn't
8374 handle several customised structures at all.
8378 * Integrated support for PVK file format and some related formats such
8379 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8380 these in the 'rsa' and 'dsa' utilities.
8384 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8388 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8389 place for the (very old) "NETSCAPE" format certificates which are now
8390 handled using new ASN1 code equivalents.
8394 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8395 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8396 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8400 * Modify CRL distribution points extension code to print out previously
8401 unsupported fields. Enhance extension setting code to allow setting of
8406 * Add print and set support for Issuing Distribution Point CRL extension.
8410 * Change 'Configure' script to enable Camellia by default.
8417 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010] ###
8419 * When rejecting SSL/TLS records due to an incorrect version number, never
8420 update s->server with a new major version number. As of
8421 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8422 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8423 the previous behavior could result in a read attempt at NULL when
8424 receiving specific incorrect SSL/TLS records once record payload
8425 protection is active. [CVE-2010-0740][]
8427 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8429 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8430 could be crashed if the relevant tables were not present (e.g. chrooted).
8432 *Tomas Hoger <thoger@redhat.com>*
8434 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010] ###
8436 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8438 *Martin Olsson, Neel Mehta*
8440 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8441 accommodate for stack sorting, always a write lock!).
8445 * On some versions of WIN32 Heap32Next is very slow. This can cause
8446 excessive delays in the RAND_poll(): over a minute. As a workaround
8447 include a time check in the inner Heap32Next loop too.
8451 * The code that handled flushing of data in SSL/TLS originally used the
8452 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8453 the problem outlined in PR#1949. The fix suggested there however can
8454 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8455 of Apache). So instead simplify the code to flush unconditionally.
8456 This should be fine since flushing with no data to flush is a no op.
8460 * Handle TLS versions 2.0 and later properly and correctly use the
8461 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8462 off ancient servers have a habit of sticking around for a while...
8466 * Modify compression code so it frees up structures without using the
8467 ex_data callbacks. This works around a problem where some applications
8468 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8469 restarting) then use compression (e.g. SSL with compression) later.
8470 This results in significant per-connection memory leaks and
8471 has caused some security issues including CVE-2008-1678 and
8476 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8477 change when encrypting or decrypting.
8481 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8482 connect and renegotiate with servers which do not support RI.
8483 Until RI is more widely deployed this option is enabled by default.
8487 * Add "missing" ssl ctrls to clear options and mode.
8491 * If client attempts to renegotiate and doesn't support RI respond with
8492 a no_renegotiation alert as required by RFC5746. Some renegotiating
8493 TLS clients will continue a connection gracefully when they receive
8494 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8495 waiting for a server hello which it will never receive. Now we treat a
8496 received no_renegotiation alert as a fatal error. This is because
8497 applications requesting a renegotiation might well expect it to succeed
8498 and would have no code in place to handle the server denying it so the
8499 only safe thing to do is to terminate the connection.
8503 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8504 peer supports secure renegotiation and 0 otherwise. Print out peer
8505 renegotiation support in s_client/s_server.
8509 * Replace the highly broken and deprecated SPKAC certification method with
8510 the updated NID creation version. This should correctly handle UTF8.
8514 * Implement RFC5746. Re-enable renegotiation but require the extension
8515 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8516 turns out to be a bad idea. It has been replaced by
8517 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8518 SSL_CTX_set_options(). This is really not recommended unless you
8519 know what you are doing.
8521 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8523 * Fixes to stateless session resumption handling. Use initial_ctx when
8524 issuing and attempting to decrypt tickets in case it has changed during
8525 servername handling. Use a non-zero length session ID when attempting
8526 stateless session resumption: this makes it possible to determine if
8527 a resumption has occurred immediately after receiving server hello
8528 (several places in OpenSSL subtly assume this) instead of later in
8533 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8534 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8535 fixes for a few places where the return code is not checked
8538 *Julia Lawall <julia@diku.dk>*
8540 * Add --strict-warnings option to Configure script to include devteam
8541 warnings in other configurations.
8545 * Add support for --libdir option and LIBDIR variable in makefiles. This
8546 makes it possible to install openssl libraries in locations which
8547 have names other than "lib", for example "/usr/lib64" which some
8550 *Steve Henson, based on patch from Jeremy Utley*
8552 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8553 X690 8.9.12 and can produce some misleading textual output of OIDs.
8555 *Steve Henson, reported by Dan Kaminsky*
8557 * Delete MD2 from algorithm tables. This follows the recommendation in
8558 several standards that it is not used in new applications due to
8559 several cryptographic weaknesses. For binary compatibility reasons
8560 the MD2 API is still compiled in by default.
8564 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8569 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8570 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8573 *Guenter <lists@gknw.net>*
8575 * Fix the server certificate chain building code to use X509_verify_cert(),
8576 it used to have an ad-hoc builder which was unable to cope with anything
8577 other than a simple chain.
8579 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8581 * Don't check self signed certificate signatures in X509_verify_cert()
8582 by default (a flag can override this): it just wastes time without
8583 adding any security. As a useful side effect self signed root CAs
8584 with non-FIPS digests are now usable in FIPS mode.
8588 * In dtls1_process_out_of_seq_message() the check if the current message
8589 is already buffered was missing. For every new message was memory
8590 allocated, allowing an attacker to perform an denial of service attack
8591 with sending out of seq handshake messages until there is no memory
8592 left. Additionally every future message was buffered, even if the
8593 sequence number made no sense and would be part of another handshake.
8594 So only messages with sequence numbers less than 10 in advance will be
8595 buffered. [CVE-2009-1378][]
8597 *Robin Seggelmann, discovered by Daniel Mentz*
8599 * Records are buffered if they arrive with a future epoch to be
8600 processed after finishing the corresponding handshake. There is
8601 currently no limitation to this buffer allowing an attacker to perform
8602 a DOS attack with sending records with future epochs until there is no
8603 memory left. This patch adds the pqueue_size() function to determine
8604 the size of a buffer and limits the record buffer to 100 entries.
8607 *Robin Seggelmann, discovered by Daniel Mentz*
8609 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8610 parent structure is freed. [CVE-2009-1379][]
8614 * Handle non-blocking I/O properly in SSL_shutdown() call.
8616 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8620 *Ilya O. <vrghost@gmail.com>*
8622 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009] ###
8624 * Disable renegotiation completely - this fixes a severe security
8625 problem [CVE-2009-3555][] at the cost of breaking all
8626 renegotiation. Renegotiation can be re-enabled by setting
8627 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8628 run-time. This is really not recommended unless you know what
8633 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009] ###
8635 * Don't set val to NULL when freeing up structures, it is freed up by
8636 underlying code. If sizeof(void *) > sizeof(long) this can result in
8637 zeroing past the valid field. [CVE-2009-0789][]
8639 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8641 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8642 checked correctly. This would allow some invalid signed attributes to
8643 appear to verify correctly. [CVE-2009-0591][]
8645 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8647 * Reject UniversalString and BMPString types with invalid lengths. This
8648 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8649 a legal length. [CVE-2009-0590][]
8653 * Set S/MIME signing as the default purpose rather than setting it
8654 unconditionally. This allows applications to override it at the store
8659 * Permit restricted recursion of ASN1 strings. This is needed in practice
8660 to handle some structures.
8664 * Improve efficiency of mem_gets: don't search whole buffer each time
8667 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8669 * New -hex option for openssl rand.
8673 * Print out UTF8String and NumericString when parsing ASN1.
8677 * Support NumericString type for name components.
8681 * Allow CC in the environment to override the automatically chosen
8682 compiler. Note that nothing is done to ensure flags work with the
8687 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009] ###
8689 * Properly check EVP_VerifyFinal() and similar return values
8692 *Ben Laurie, Bodo Moeller, Google Security Team*
8694 * Enable TLS extensions by default.
8698 * Allow the CHIL engine to be loaded, whether the application is
8699 multithreaded or not. (This does not release the developer from the
8700 obligation to set up the dynamic locking callbacks.)
8702 *Sander Temme <sander@temme.net>*
8704 * Use correct exit code if there is an error in dgst command.
8706 *Steve Henson; problem pointed out by Roland Dirlewanger*
8708 * Tweak Configure so that you need to say "experimental-jpake" to enable
8709 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8713 * Add experimental JPAKE support, including demo authentication in
8714 s_client and s_server.
8718 * Set the comparison function in v3_addr_canonize().
8720 *Rob Austein <sra@hactrn.net>*
8722 * Add support for XMPP STARTTLS in s_client.
8724 *Philip Paeps <philip@freebsd.org>*
8726 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8727 to ensure that even with this option, only ciphersuites in the
8728 server's preference list will be accepted. (Note that the option
8729 applies only when resuming a session, so the earlier behavior was
8730 just about the algorithm choice for symmetric cryptography.)
8734 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008] ###
8736 * Fix NULL pointer dereference if a DTLS server received
8737 ChangeCipherSpec as first record [CVE-2009-1386][].
8741 * Fix a state transition in s3_srvr.c and d1_srvr.c
8742 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
8746 * The fix in 0.9.8c that supposedly got rid of unsafe
8747 double-checked locking was incomplete for RSA blinding,
8748 addressing just one layer of what turns out to have been
8749 doubly unsafe triple-checked locking.
8751 So now fix this for real by retiring the MONT_HELPER macro
8752 in crypto/rsa/rsa_eay.c.
8755 *Bodo Moeller; problem pointed out by Marius Schilder*
8757 * Various precautionary measures:
8759 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8761 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8762 (NB: This would require knowledge of the secret session ticket key
8763 to exploit, in which case you'd be SOL either way.)
8765 - Change bn_nist.c so that it will properly handle input BIGNUMs
8766 outside the expected range.
8768 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8772 *Neel Mehta, Bodo Moeller*
8774 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8775 the load fails. Useful for distros.
8777 *Ben Laurie and the FreeBSD team*
8779 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8783 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8787 * Expand ENGINE to support engine supplied SSL client certificate functions.
8789 This work was sponsored by Logica.
8793 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8794 keystores. Support for SSL/TLS client authentication too.
8795 Not compiled unless enable-capieng specified to Configure.
8797 This work was sponsored by Logica.
8801 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8802 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8803 attribute creation routines such as certificate requests and PKCS#12
8808 ### Changes between 0.9.8g and 0.9.8h [28 May 2008] ###
8810 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8811 handshake which could lead to a client crash as found using the
8812 Codenomicon TLS test suite [CVE-2008-1672][]
8814 *Steve Henson, Mark Cox*
8816 * Fix double free in TLS server name extensions which could lead to
8817 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8821 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8823 Clear the error queue to ensure that error entries left from
8824 older function calls do not interfere with the correct operation.
8826 *Lutz Jaenicke, Erik de Castro Lopo*
8828 * Remove root CA certificates of commercial CAs:
8830 The OpenSSL project does not recommend any specific CA and does not
8831 have any policy with respect to including or excluding any CA.
8832 Therefore it does not make any sense to ship an arbitrary selection
8833 of root CA certificates with the OpenSSL software.
8837 * RSA OAEP patches to fix two separate invalid memory reads.
8838 The first one involves inputs when 'lzero' is greater than
8839 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8840 before the beginning of from). The second one involves inputs where
8841 the 'db' section contains nothing but zeroes (there is a one-byte
8842 invalid read after the end of 'db').
8844 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8846 * Partial backport from 0.9.9-dev:
8848 Introduce bn_mul_mont (dedicated Montgomery multiplication
8849 procedure) as a candidate for BIGNUM assembler implementation.
8850 While 0.9.9-dev uses assembler for various architectures, only
8851 x86_64 is available by default here in the 0.9.8 branch, and
8852 32-bit x86 is available through a compile-time setting.
8854 To try the 32-bit x86 assembler implementation, use Configure
8855 option "enable-montasm" (which exists only for this backport).
8857 As "enable-montasm" for 32-bit x86 disclaims code stability
8858 anyway, in this constellation we activate additional code
8859 backported from 0.9.9-dev for further performance improvements,
8860 namely BN_from_montgomery_word. (To enable this otherwise,
8861 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8864 *Andy Polyakov (backport partially by Bodo Moeller)*
8866 * Add TLS session ticket callback. This allows an application to set
8867 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8868 values. This is useful for key rollover for example where several key
8869 sets may exist with different names.
8873 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8874 This was broken until now in 0.9.8 releases, such that the only way
8875 a registered ENGINE could be used (assuming it initialises
8876 successfully on the host) was to explicitly set it as the default
8877 for the relevant algorithms. This is in contradiction with 0.9.7
8878 behaviour and the documentation. With this fix, when an ENGINE is
8879 registered into a given algorithm's table of implementations, the
8880 'uptodate' flag is reset so that auto-discovery will be used next
8881 time a new context for that algorithm attempts to select an
8884 *Ian Lister (tweaked by Geoff Thorpe)*
8886 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8887 implementation in the following ways:
8889 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8892 Lack of BER streaming support means one pass streaming processing is
8893 only supported if data is detached: setting the streaming flag is
8894 ignored for embedded content.
8896 CMS support is disabled by default and must be explicitly enabled
8897 with the enable-cms configuration option.
8901 * Update the GMP engine glue to do direct copies between BIGNUM and
8902 mpz_t when openssl and GMP use the same limb size. Otherwise the
8903 existing "conversion via a text string export" trick is still used.
8905 *Paul Sheer <paulsheer@gmail.com>*
8907 * Zlib compression BIO. This is a filter BIO which compressed and
8908 uncompresses any data passed through it.
8912 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
8913 RFC3394 compatible AES key wrapping.
8917 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
8918 sets string data without copying. X509_ALGOR_set0() and
8919 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
8920 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
8921 from an X509_ATTRIBUTE structure optionally checking it occurs only
8922 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
8927 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
8928 to get the expected BN_FLG_CONSTTIME behavior.
8930 *Bodo Moeller (Google)*
8934 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
8935 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
8936 - added some more tests to do_tests.pl
8937 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
8938 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
8939 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
8940 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
8941 - various changes to netware.pl to enable gcc-cross builds on Win32
8943 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
8944 - various changes to fix missing prototype warnings
8945 - fixed x86nasm.pl to create correct asm files for NASM COFF output
8946 - added AES, WHIRLPOOL and CPUID assembler code to build files
8947 - added missing AES assembler make rules to mk1mf.pl
8948 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
8950 *Guenter Knauf <eflash@gmx.net>*
8952 * Implement certificate status request TLS extension defined in RFC3546.
8953 A client can set the appropriate parameters and receive the encoded
8954 OCSP response via a callback. A server can query the supplied parameters
8955 and set the encoded OCSP response in the callback. Add simplified examples
8956 to s_client and s_server.
8960 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007] ###
8963 + Binary incompatibility of ssl_ctx_st structure
8964 + DTLS interoperation with non-compliant servers
8965 + Don't call get_session_cb() without proposed session
8966 + Fix ia64 assembler code
8968 *Andy Polyakov, Steve Henson*
8970 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007] ###
8972 * DTLS Handshake overhaul. There were longstanding issues with
8973 OpenSSL DTLS implementation, which were making it impossible for
8974 RFC 4347 compliant client to communicate with OpenSSL server.
8975 Unfortunately just fixing these incompatibilities would "cut off"
8976 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
8977 server keeps tolerating non RFC compliant syntax. The opposite is
8978 not true, 0.9.8f client can not communicate with earlier server.
8979 This update even addresses CVE-2007-4995.
8983 * Changes to avoid need for function casts in OpenSSL: some compilers
8984 (gcc 4.2 and later) reject their use.
8985 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
8988 * Add RFC4507 support to OpenSSL. This includes the corrections in
8989 RFC4507bis. The encrypted ticket format is an encrypted encoded
8990 SSL_SESSION structure, that way new session features are automatically
8993 If a client application caches session in an SSL_SESSION structure
8994 support is transparent because tickets are now stored in the encoded
8997 The SSL_CTX structure automatically generates keys for ticket
8998 protection in servers so again support should be possible
8999 with no application modification.
9001 If a client or server wishes to disable RFC4507 support then the option
9002 SSL_OP_NO_TICKET can be set.
9004 Add a TLS extension debugging callback to allow the contents of any client
9005 or server extensions to be examined.
9007 This work was sponsored by Google.
9011 * Add initial support for TLS extensions, specifically for the server_name
9012 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9013 have new members for a host name. The SSL data structure has an
9014 additional member SSL_CTX *initial_ctx so that new sessions can be
9015 stored in that context to allow for session resumption, even after the
9016 SSL has been switched to a new SSL_CTX in reaction to a client's
9017 server_name extension.
9019 New functions (subject to change):
9021 SSL_get_servername()
9022 SSL_get_servername_type()
9025 New CTRL codes and macros (subject to change):
9027 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9028 - SSL_CTX_set_tlsext_servername_callback()
9029 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9030 - SSL_CTX_set_tlsext_servername_arg()
9031 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9033 openssl s_client has a new '-servername ...' option.
9035 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9036 '-key2 ...', '-servername_fatal' (subject to change). This allows
9037 testing the HostName extension for a specific single host name ('-cert'
9038 and '-key' remain fallbacks for handshakes without HostName
9039 negotiation). If the unrecognized_name alert has to be sent, this by
9040 default is a warning; it becomes fatal with the '-servername_fatal'
9044 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9046 * Add AES and SSE2 assembly language support to VC++ build.
9050 * Mitigate attack on final subtraction in Montgomery reduction.
9054 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9055 (which previously caused an internal error).
9059 * Squeeze another 10% out of IGE mode when in != out.
9063 * AES IGE mode speedup.
9065 *Dean Gaudet (Google)*
9067 * Add the Korean symmetric 128-bit cipher SEED (see
9068 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
9069 add SEED ciphersuites from RFC 4162:
9071 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9072 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9073 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9074 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9076 To minimize changes between patchlevels in the OpenSSL 0.9.8
9077 series, SEED remains excluded from compilation unless OpenSSL
9078 is configured with 'enable-seed'.
9080 *KISA, Bodo Moeller*
9082 * Mitigate branch prediction attacks, which can be practical if a
9083 single processor is shared, allowing a spy process to extract
9084 information. For detailed background information, see
9085 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
9086 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9087 and Necessary Software Countermeasures"). The core of the change
9088 are new versions BN_div_no_branch() and
9089 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9090 respectively, which are slower, but avoid the security-relevant
9091 conditional branches. These are automatically called by BN_div()
9092 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9093 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9094 remove a conditional branch.
9096 BN_FLG_CONSTTIME is the new name for the previous
9097 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9098 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9099 in the exponent causes BN_mod_exp_mont() to use the alternative
9100 implementation in BN_mod_exp_mont_consttime().) The old name
9101 remains as a deprecated alias.
9103 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9104 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9105 constant-time implementations for more than just exponentiation.
9106 Here too the old name is kept as a deprecated alias.
9108 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9109 the BN_BLINDING structure gets an independent copy of the
9110 modulus. This means that the previous "BIGNUM *m" argument to
9111 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9112 essentially becomes "const BIGNUM *m", although we can't actually
9113 change this in the header file before 0.9.9. It allows
9114 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9115 enable BN_FLG_CONSTTIME.
9118 *Matthew D Wood (Intel Corp)*
9120 * In the SSL/TLS server implementation, be strict about session ID
9121 context matching (which matters if an application uses a single
9122 external cache for different purposes). Previously,
9123 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9124 set. This did ensure strict client verification, but meant that,
9125 with applications using a single external cache for quite
9126 different requirements, clients could circumvent ciphersuite
9127 restrictions for a given session ID context by starting a session
9128 in a different context.
9132 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9133 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9134 authentication-only ciphersuites.
9138 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9139 not complete and could lead to a possible single byte overflow
9140 [CVE-2007-5135][] [Ben Laurie]
9142 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007] ###
9144 * Since AES128 and AES256 (and similarly Camellia128 and
9145 Camellia256) share a single mask bit in the logic of
9146 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9147 kludge to work properly if AES128 is available and AES256 isn't
9148 (or if Camellia128 is available and Camellia256 isn't).
9152 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9153 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9154 When a point or a seed is encoded in a BIT STRING, we need to
9155 prevent the removal of trailing zero bits to get the proper DER
9156 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9157 of a NamedBitList, for which trailing 0 bits need to be removed.)
9161 * Have SSL/TLS server implementation tolerate "mismatched" record
9162 protocol version while receiving ClientHello even if the
9163 ClientHello is fragmented. (The server can't insist on the
9164 particular protocol version it has chosen before the ServerHello
9165 message has informed the client about his choice.)
9169 * Add RFC 3779 support.
9171 *Rob Austein for ARIN, Ben Laurie*
9173 * Load error codes if they are not already present instead of using a
9174 static variable. This allows them to be cleanly unloaded and reloaded.
9175 Improve header file function name parsing.
9179 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9180 or CAPABILITY handshake as required by RFCs.
9184 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006] ###
9186 * Introduce limits to prevent malicious keys being able to
9187 cause a denial of service. [CVE-2006-2940][]
9189 *Steve Henson, Bodo Moeller*
9191 * Fix ASN.1 parsing of certain invalid structures that can result
9192 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9194 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9195 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9197 * Fix SSL client code which could crash if connecting to a
9198 malicious SSLv2 server. [CVE-2006-4343][]
9200 *Tavis Ormandy and Will Drewry, Google Security Team*
9202 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9203 match only those. Before that, "AES256-SHA" would be interpreted
9204 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9205 the same strength classification in 0.9.7h) as we currently only
9206 have a single AES bit in the ciphersuite description bitmap.
9207 That change, however, also applied to ciphersuite strings such as
9208 "RC4-MD5" that intentionally matched multiple ciphersuites --
9209 namely, SSL 2.0 ciphersuites in addition to the more common ones
9210 from SSL 3.0/TLS 1.0.
9212 So we change the selection algorithm again: Naming an explicit
9213 ciphersuite selects this one ciphersuite, and any other similar
9214 ciphersuite (same bitmap) from *other* protocol versions.
9215 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9216 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9218 Since SSL 2.0 does not have any ciphersuites for which the
9219 128/256 bit distinction would be relevant, this works for now.
9220 The proper fix will be to use different bits for AES128 and
9221 AES256, which would have avoided the problems from the beginning;
9222 however, bits are scarce, so we can only do this in a new release
9223 (not just a patchlevel) when we can change the SSL_CIPHER
9224 definition to split the single 'unsigned long mask' bitmap into
9225 multiple values to extend the available space.
9230 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006] ###
9232 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9233 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9235 * Add AES IGE and biIGE modes.
9239 * Change the Unix randomness entropy gathering to use poll() when
9240 possible instead of select(), since the latter has some
9241 undesirable limitations.
9243 *Darryl Miles via Richard Levitte and Bodo Moeller*
9245 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9246 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9247 cannot be implicitly activated as part of, e.g., the "AES" alias.
9248 However, please upgrade to OpenSSL 0.9.9[-dev] for
9249 non-experimental use of the ECC ciphersuites to get TLS extension
9250 support, which is required for curve and point format negotiation
9251 to avoid potential handshake problems.
9255 * Disable rogue ciphersuites:
9257 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9258 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9259 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9261 The latter two were purportedly from
9262 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9265 Also deactivate the remaining ciphersuites from
9266 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9267 unofficial, and the ID has long expired.
9271 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9272 dual-core machines) and other potential thread-safety issues.
9276 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9277 versions), which is now available for royalty-free use
9278 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
9279 Also, add Camellia TLS ciphersuites from RFC 4132.
9281 To minimize changes between patchlevels in the OpenSSL 0.9.8
9282 series, Camellia remains excluded from compilation unless OpenSSL
9283 is configured with 'enable-camellia'.
9287 * Disable the padding bug check when compression is in use. The padding
9288 bug check assumes the first packet is of even length, this is not
9289 necessarily true if compression is enabled and can result in false
9290 positives causing handshake failure. The actual bug test is ancient
9291 code so it is hoped that implementations will either have fixed it by
9292 now or any which still have the bug do not support compression.
9296 ### Changes between 0.9.8a and 0.9.8b [04 May 2006] ###
9298 * When applying a cipher rule check to see if string match is an explicit
9299 cipher suite and only match that one cipher suite if it is.
9303 * Link in manifests for VC++ if needed.
9305 *Austin Ziegler <halostatue@gmail.com>*
9307 * Update support for ECC-based TLS ciphersuites according to
9308 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9309 TLS extensions, which are supported starting with the 0.9.9
9310 branch, not in the OpenSSL 0.9.8 branch).
9314 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9315 opaque EVP_CIPHER_CTX handling.
9319 * Fixes and enhancements to zlib compression code. We now only use
9320 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9321 to conform with the standards mentioned here:
9322 http://www.zlib.net/DLL_FAQ.txt
9323 Static zlib linking now works on Windows and the new --with-zlib-include
9324 --with-zlib-lib options to Configure can be used to supply the location
9325 of the headers and library. Gracefully handle case where zlib library
9330 * Several fixes and enhancements to the OID generation code. The old code
9331 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9332 handle numbers larger than ULONG_MAX, truncated printing and had a
9333 non standard OBJ_obj2txt() behaviour.
9337 * Add support for building of engines under engine/ as shared libraries
9338 under VC++ build system.
9342 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9343 Hopefully, we will not see any false combination of paths any more.
9347 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005] ###
9349 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9350 (part of SSL_OP_ALL). This option used to disable the
9351 countermeasure against man-in-the-middle protocol-version
9352 rollback in the SSL 2.0 server implementation, which is a bad
9353 idea. [CVE-2005-2969][]
9355 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9356 for Information Security, National Institute of Advanced Industrial
9357 Science and Technology [AIST*, Japan)]
9359 * Add two function to clear and return the verify parameter flags.
9363 * Keep cipherlists sorted in the source instead of sorting them at
9364 runtime, thus removing the need for a lock.
9368 * Avoid some small subgroup attacks in Diffie-Hellman.
9370 *Nick Mathewson and Ben Laurie*
9372 * Add functions for well-known primes.
9376 * Extended Windows CE support.
9378 *Satoshi Nakamura and Andy Polyakov*
9380 * Initialize SSL_METHOD structures at compile time instead of during
9381 runtime, thus removing the need for a lock.
9385 * Make PKCS7_decrypt() work even if no certificate is supplied by
9386 attempting to decrypt each encrypted key in turn. Add support to
9391 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005] ###
9393 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9396 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9400 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9401 key into the same file any more.
9405 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9409 * Add -utf8 command line and config file option to 'ca'.
9411 *Stefan <stf@udoma.org*
9413 * Removed the macro des_crypt(), as it seems to conflict with some
9414 libraries. Use DES_crypt().
9418 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9419 involves renaming the source and generated shared-libs for
9420 both. The engines will accept the corrected or legacy ids
9421 ('ncipher' and '4758_cca' respectively) when binding. NB,
9422 this only applies when building 'shared'.
9424 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9426 * Add attribute functions to EVP_PKEY structure. Modify
9427 PKCS12_create() to recognize a CSP name attribute and
9428 use it. Make -CSP option work again in pkcs12 utility.
9432 * Add new functionality to the bn blinding code:
9433 - automatic re-creation of the BN_BLINDING parameters after
9434 a fixed number of uses (currently 32)
9435 - add new function for parameter creation
9436 - introduce flags to control the update behaviour of the
9437 BN_BLINDING parameters
9438 - hide BN_BLINDING structure
9439 Add a second BN_BLINDING slot to the RSA structure to improve
9440 performance when a single RSA object is shared among several
9445 * Add support for DTLS.
9447 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9449 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9450 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9454 * Remove buggy and incomplete DH cert support from
9455 ssl/ssl_rsa.c and ssl/s3_both.c
9459 * Use SHA-1 instead of MD5 as the default digest algorithm for
9460 the apps/openssl applications.
9464 * Compile clean with "-Wall -Wmissing-prototypes
9465 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9466 DEBUG_SAFESTACK must also be set.
9470 * Change ./Configure so that certain algorithms can be disabled by default.
9471 The new counterpiece to "no-xxx" is "enable-xxx".
9473 The patented RC5 and MDC2 algorithms will now be disabled unless
9474 "enable-rc5" and "enable-mdc2", respectively, are specified.
9476 (IDEA remains enabled despite being patented. This is because IDEA
9477 is frequently required for interoperability, and there is no license
9478 fee for non-commercial use. As before, "no-idea" can be used to
9479 avoid this algorithm.)
9484 * Add processing of proxy certificates (see RFC 3820). This work was
9485 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9486 EGEE (Enabling Grids for E-science in Europe).
9490 * RC4 performance overhaul on modern architectures/implementations, such
9491 as Intel P4, IA-64 and AMD64.
9495 * New utility extract-section.pl. This can be used specify an alternative
9496 section number in a pod file instead of having to treat each file as
9497 a separate case in Makefile. This can be done by adding two lines to the
9500 =for comment openssl_section:XXX
9502 The blank line is mandatory.
9507 * New arguments -certform, -keyform and -pass for s_client and s_server
9508 to allow alternative format key and certificate files and passphrase
9513 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9514 update associated structures and add various utility functions.
9516 Add new policy related verify parameters, include policy checking in
9517 standard verify code. Enhance 'smime' application with extra parameters
9518 to support policy checking and print out.
9522 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9523 Nehemiah processors. These extensions support AES encryption in hardware
9524 as well as RNG (though RNG support is currently disabled).
9526 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9528 * Deprecate BN_[get|set]_params() functions (they were ignored internally).
9532 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9534 *Andy Polyakov and a number of other people*
9536 * Improved PowerPC platform support. Most notably BIGNUM assembler
9537 implementation contributed by IBM.
9539 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9541 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9542 exponent rather than 'unsigned long'. There is a corresponding change to
9543 the new 'rsa_keygen' element of the RSA_METHOD structure.
9545 *Jelte Jansen, Geoff Thorpe*
9547 * Functionality for creating the initial serial number file is now
9548 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9550 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9551 number file to 1, which is bound to cause problems. To avoid
9552 the problems while respecting compatibility between different 0.9.7
9553 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9554 CA.pl for serial number initialization. With the new release 0.9.8,
9555 we can fix the problem directly in the 'ca' utility.)
9559 * Reduced header interdependencies by declaring more opaque objects in
9560 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9561 give fewer recursive includes, which could break lazy source code - so
9562 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9563 developers should define this symbol when building and using openssl to
9564 ensure they track the recommended behaviour, interfaces, [etc], but
9565 backwards-compatible behaviour prevails when this isn't defined.
9569 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9573 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9574 This will generate a random key of the appropriate length based on the
9575 cipher context. The EVP_CIPHER can provide its own random key generation
9576 routine to support keys of a specific form. This is used in the des and
9577 3des routines to generate a key of the correct parity. Update S/MIME
9578 code to use new functions and hence generate correct parity DES keys.
9579 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9580 valid (weak or incorrect parity).
9584 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9585 as looking them up. This is useful when the verified structure may contain
9586 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9587 present unless the new PKCS7_NO_CRL flag is asserted.
9591 * Extend ASN1 oid configuration module. It now additionally accepts the
9594 shortName = some long name, 1.2.3.4
9598 * Reimplemented the BN_CTX implementation. There is now no more static
9599 limitation on the number of variables it can handle nor the depth of the
9600 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9601 information can now expand as required, and rather than having a single
9602 static array of bignums, BN_CTX now uses a linked-list of such arrays
9603 allowing it to expand on demand whilst maintaining the usefulness of
9604 BN_CTX's "bundling".
9608 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9609 to allow all RSA operations to function using a single BN_CTX.
9613 * Preliminary support for certificate policy evaluation and checking. This
9614 is initially intended to pass the tests outlined in "Conformance Testing
9615 of Relying Party Client Certificate Path Processing Logic" v1.07.
9619 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9620 remained unused and not that useful. A variety of other little bignum
9621 tweaks and fixes have also been made continuing on from the audit (see
9626 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9627 associated ASN1, EVP and SSL functions and old ASN1 macros.
9631 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9632 and this should never fail. So the return value from the use of
9633 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9634 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9638 * BN_CTX_get() should return zero-valued bignums, providing the same
9639 initialised value as BN_new().
9641 *Geoff Thorpe, suggested by Ulf Möller*
9643 * Support for inhibitAnyPolicy certificate extension.
9647 * An audit of the BIGNUM code is underway, for which debugging code is
9648 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9649 is considered valid when processing BIGNUMs, and causes execution to
9650 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9651 further steps are taken to deliberately pollute unused data in BIGNUM
9652 structures to try and expose faulty code further on. For now, openssl will
9653 (in its default mode of operation) continue to tolerate the inconsistent
9654 forms that it has tolerated in the past, but authors and packagers should
9655 consider trying openssl and their own applications when compiled with
9656 these debugging symbols defined. It will help highlight potential bugs in
9657 their own code, and will improve the test coverage for OpenSSL itself. At
9658 some point, these tighter rules will become openssl's default to improve
9659 maintainability, though the assert()s and other overheads will remain only
9660 in debugging configurations. See bn.h for more details.
9662 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9664 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9665 that can only be obtained through BN_CTX_new() (which implicitly
9666 initialises it). The presence of this function only made it possible
9667 to overwrite an existing structure (and cause memory leaks).
9671 * Because of the callback-based approach for implementing LHASH as a
9672 template type, lh_insert() adds opaque objects to hash-tables and
9673 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9674 to clean up those corresponding objects before destroying the hash table
9675 (and losing the object pointers). So some over-zealous constifications in
9676 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9677 objects as "const" and the lh_doall[_arg] callback wrappers are not
9678 prototyped to have "const" restrictions on the object pointers they are
9679 given (and so aren't required to cast them away any more).
9683 * The tmdiff.h API was so ugly and minimal that our own timing utility
9684 (speed) prefers to use its own implementation. The two implementations
9685 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9686 its object type properly exposed (MS_TM) instead of casting to/from "char
9687 *". This may still change yet if someone realises MS_TM and "ms_time_***"
9688 aren't necessarily the greatest nomenclatures - but this is what was used
9689 internally to the implementation so I've used that for now.
9693 * Ensure that deprecated functions do not get compiled when
9694 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9695 the self-tests were still using deprecated key-generation functions so
9696 these have been updated also.
9700 * Reorganise PKCS#7 code to separate the digest location functionality
9701 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9702 New function PKCS7_set_digest() to set the digest type for PKCS#7
9703 digestedData type. Add additional code to correctly generate the
9704 digestedData type and add support for this type in PKCS7 initialization
9709 * New function PKCS7_set0_type_other() this initializes a PKCS7
9710 structure of type "other".
9714 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9715 sure the loop does correctly stop and breaking ("division by zero")
9716 modulus operations are not performed. The (pre-generated) prime
9717 table crypto/bn/bn_prime.h was already correct, but it could not be
9718 re-generated on some platforms because of the "division by zero"
9719 situation in the script.
9721 *Ralf S. Engelschall*
9723 * Update support for ECC-based TLS ciphersuites according to
9724 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9725 SHA-1 now is only used for "small" curves (where the
9726 representation of a field element takes up to 24 bytes); for
9727 larger curves, the field element resulting from ECDH is directly
9728 used as premaster secret.
9730 *Douglas Stebila (Sun Microsystems Laboratories)*
9732 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9733 curve secp160r1 to the tests.
9735 *Douglas Stebila (Sun Microsystems Laboratories)*
9737 * Add the possibility to load symbols globally with DSO.
9739 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9741 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9742 control of the error stack.
9746 * Add support for STORE in ENGINE.
9750 * Add the STORE type. The intention is to provide a common interface
9751 to certificate and key stores, be they simple file-based stores, or
9752 HSM-type store, or LDAP stores, or...
9753 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9757 * Add a generic structure called OPENSSL_ITEM. This can be used to
9758 pass a list of arguments to any function as well as provide a way
9759 for a function to pass data back to the caller.
9763 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9764 works like BUF_strdup() but can be used to duplicate a portion of
9765 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9770 * Add the function sk_find_ex() which works like sk_find(), but will
9771 return an index to an element even if an exact match couldn't be
9772 found. The index is guaranteed to point at the element where the
9773 searched-for key would be inserted to preserve sorting order.
9777 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9778 takes an extra flags argument for optional functionality. Currently,
9779 the following flags are defined:
9781 OBJ_BSEARCH_VALUE_ON_NOMATCH
9782 This one gets OBJ_bsearch_ex() to return a pointer to the first
9783 element where the comparing function returns a negative or zero
9786 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9787 This one gets OBJ_bsearch_ex() to return a pointer to the first
9788 element where the comparing function returns zero. This is useful
9789 if there are more than one element where the comparing function
9794 * Make it possible to create self-signed certificates with 'openssl ca'
9795 in such a way that the self-signed certificate becomes part of the
9796 CA database and uses the same mechanisms for serial number generation
9797 as all other certificate signing. The new flag '-selfsign' enables
9798 this functionality. Adapt CA.sh and CA.pl.in.
9802 * Add functionality to check the public key of a certificate request
9803 against a given private. This is useful to check that a certificate
9804 request can be signed by that key (self-signing).
9808 * Make it possible to have multiple active certificates with the same
9809 subject in the CA index file. This is done only if the keyword
9810 'unique_subject' is set to 'no' in the main CA section (default
9811 if 'CA_default') of the configuration file. The value is saved
9812 with the database itself in a separate index attribute file,
9813 named like the index file with '.attr' appended to the name.
9817 * Generate multi-valued AVAs using '+' notation in config files for
9822 * Support for nameConstraints certificate extension.
9826 * Support for policyConstraints certificate extension.
9830 * Support for policyMappings certificate extension.
9834 * Make sure the default DSA_METHOD implementation only uses its
9835 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9836 and change its own handlers to be NULL so as to remove unnecessary
9837 indirection. This lets alternative implementations fallback to the
9838 default implementation more easily.
9842 * Support for directoryName in GeneralName related extensions
9847 * Make it possible to link applications using Makefile.shared.
9848 Make that possible even when linking against static libraries!
9852 * Support for single pass processing for S/MIME signing. This now
9853 means that S/MIME signing can be done from a pipe, in addition
9854 cleartext signing (multipart/signed type) is effectively streaming
9855 and the signed data does not need to be all held in memory.
9857 This is done with a new flag PKCS7_STREAM. When this flag is set
9858 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9859 is done after the data is output (and digests calculated) in
9860 SMIME_write_PKCS7().
9864 * Add full support for -rpath/-R, both in shared libraries and
9865 applications, at least on the platforms where it's known how
9870 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9871 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9872 will now compute a table of multiples of the generator that
9873 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9874 faster (notably in the case of a single point multiplication,
9875 scalar * generator).
9877 *Nils Larsch, Bodo Moeller*
9879 * IPv6 support for certificate extensions. The various extensions
9880 which use the IP:a.b.c.d can now take IPv6 addresses using the
9881 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9886 * Added an ENGINE that implements RSA by performing private key
9887 exponentiations with the GMP library. The conversions to and from
9888 GMP's mpz_t format aren't optimised nor are any montgomery forms
9889 cached, and on x86 it appears OpenSSL's own performance has caught up.
9890 However there are likely to be other architectures where GMP could
9891 provide a boost. This ENGINE is not built in by default, but it can be
9892 specified at Configure time and should be accompanied by the necessary
9893 linker additions, eg;
9894 ./config -DOPENSSL_USE_GMP -lgmp
9898 * "openssl engine" will not display ENGINE/DSO load failure errors when
9899 testing availability of engines with "-t" - the old behaviour is
9900 produced by increasing the feature's verbosity with "-tt".
9904 * ECDSA routines: under certain error conditions uninitialized BN objects
9905 could be freed. Solution: make sure initialization is performed early
9906 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
9911 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
9912 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9913 software implementations. For DSA and DH, parameter generation can
9914 also be overridden by providing the appropriate method callbacks.
9918 * Change the "progress" mechanism used in key-generation and
9919 primality testing to functions that take a new BN_GENCB pointer in
9920 place of callback/argument pairs. The new API functions have "_ex"
9921 postfixes and the older functions are reimplemented as wrappers for
9922 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
9923 declarations of the old functions to help (graceful) attempts to
9924 migrate to the new functions. Also, the new key-generation API
9925 functions operate on a caller-supplied key-structure and return
9926 success/failure rather than returning a key or NULL - this is to
9927 help make "keygen" another member function of RSA_METHOD etc.
9929 Example for using the new callback interface:
9931 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
9935 BN_GENCB_set(&my_cb, my_callback, my_arg);
9937 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
9938 /* For the meaning of a, b in calls to my_callback(), see the
9939 * documentation of the function that calls the callback.
9940 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
9941 * my_callback should return 1 if it wants BN_is_prime_ex()
9942 * to continue, or 0 to stop.
9947 * Change the ZLIB compression method to be stateful, and make it
9948 available to TLS with the number defined in
9949 draft-ietf-tls-compression-04.txt.
9953 * Add the ASN.1 structures and functions for CertificatePair, which
9954 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
9956 CertificatePair ::= SEQUENCE {
9957 forward [0] Certificate OPTIONAL,
9958 reverse [1] Certificate OPTIONAL,
9959 -- at least one of the pair shall be present -- }
9961 Also implement the PEM functions to read and write certificate
9962 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
9964 This needed to be defined, mostly for the sake of the LDAP
9965 attribute crossCertificatePair, but may prove useful elsewhere as
9970 * Make it possible to inhibit symlinking of shared libraries in
9971 Makefile.shared, for Cygwin's sake.
9975 * Extend the BIGNUM API by creating a function
9976 void BN_set_negative(BIGNUM *a, int neg);
9977 and a macro that behave like
9978 int BN_is_negative(const BIGNUM *a);
9980 to avoid the need to access 'a->neg' directly in applications.
9984 * Implement fast modular reduction for pseudo-Mersenne primes
9985 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
9986 EC_GROUP_new_curve_GFp() will now automatically use this
9989 *Nils Larsch <nla@trustcenter.de>*
9991 * Add new lock type (CRYPTO_LOCK_BN).
9995 * Change the ENGINE framework to automatically load engines
9996 dynamically from specific directories unless they could be
9997 found to already be built in or loaded. Move all the
9998 current engines except for the cryptodev one to a new
10000 The engines in engines/ are built as shared libraries if
10001 the "shared" options was given to ./Configure or ./config.
10002 Otherwise, they are inserted in libcrypto.a.
10003 /usr/local/ssl/engines is the default directory for dynamic
10004 engines, but that can be overridden at configure time through
10005 the usual use of --prefix and/or --openssldir, and at run
10006 time with the environment variable OPENSSL_ENGINES.
10008 *Geoff Thorpe and Richard Levitte*
10010 * Add Makefile.shared, a helper makefile to build shared
10011 libraries. Adapt Makefile.org.
10015 * Add version info to Win32 DLLs.
10017 *Peter 'Luna' Runestig" <peter@runestig.com>*
10019 * Add new 'medium level' PKCS#12 API. Certificates and keys
10020 can be added using this API to created arbitrary PKCS#12
10021 files while avoiding the low level API.
10023 New options to PKCS12_create(), key or cert can be NULL and
10024 will then be omitted from the output file. The encryption
10025 algorithm NIDs can be set to -1 for no encryption, the mac
10026 iteration count can be set to 0 to omit the mac.
10028 Enhance pkcs12 utility by making the -nokeys and -nocerts
10029 options work when creating a PKCS#12 file. New option -nomac
10030 to omit the mac, NONE can be set for an encryption algorithm.
10031 New code is modified to use the enhanced PKCS12_create()
10032 instead of the low level API.
10036 * Extend ASN1 encoder to support indefinite length constructed
10037 encoding. This can output sequences tags and octet strings in
10038 this form. Modify pk7_asn1.c to support indefinite length
10039 encoding. This is experimental and needs additional code to
10040 be useful, such as an ASN1 bio and some enhanced streaming
10043 Extend template encode functionality so that tagging is passed
10044 down to the template encoder.
10048 * Let 'openssl req' fail if an argument to '-newkey' is not
10049 recognized instead of using RSA as a default.
10053 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10054 As these are not official, they are not included in "ALL";
10055 the "ECCdraft" ciphersuite group alias can be used to select them.
10057 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10059 * Add ECDH engine support.
10061 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10063 * Add ECDH in new directory crypto/ecdh/.
10065 *Douglas Stebila (Sun Microsystems Laboratories)*
10067 * Let BN_rand_range() abort with an error after 100 iterations
10068 without success (which indicates a broken PRNG).
10072 * Change BN_mod_sqrt() so that it verifies that the input value
10073 is really the square of the return value. (Previously,
10074 BN_mod_sqrt would show GIGO behaviour.)
10078 * Add named elliptic curves over binary fields from X9.62, SECG,
10079 and WAP/WTLS; add OIDs that were still missing.
10081 *Sheueling Chang Shantz and Douglas Stebila
10082 (Sun Microsystems Laboratories)*
10084 * Extend the EC library for elliptic curves over binary fields
10085 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10088 EC_GF2m_simple_method
10092 EC_GROUP_new_curve_GF2m
10093 EC_GROUP_set_curve_GF2m
10094 EC_GROUP_get_curve_GF2m
10095 EC_POINT_set_affine_coordinates_GF2m
10096 EC_POINT_get_affine_coordinates_GF2m
10097 EC_POINT_set_compressed_coordinates_GF2m
10099 Point compression for binary fields is disabled by default for
10100 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10103 As binary polynomials are represented as BIGNUMs, various members
10104 of the EC_GROUP and EC_POINT data structures can be shared
10105 between the implementations for prime fields and binary fields;
10106 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
10107 are essentially identical to their ..._GFp counterparts.
10108 (For simplicity, the '..._GFp' prefix has been dropped from
10109 various internal method names.)
10111 An internal 'field_div' method (similar to 'field_mul' and
10112 'field_sqr') has been added; this is used only for binary fields.
10114 *Sheueling Chang Shantz and Douglas Stebila
10115 (Sun Microsystems Laboratories)*
10117 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10118 through methods ('mul', 'precompute_mult').
10120 The generic implementations (now internally called 'ec_wNAF_mul'
10121 and 'ec_wNAF_precomputed_mult') remain the default if these
10122 methods are undefined.
10124 *Sheueling Chang Shantz and Douglas Stebila
10125 (Sun Microsystems Laboratories)*
10127 * New function EC_GROUP_get_degree, which is defined through
10128 EC_METHOD. For curves over prime fields, this returns the bit
10129 length of the modulus.
10131 *Sheueling Chang Shantz and Douglas Stebila
10132 (Sun Microsystems Laboratories)*
10134 * New functions EC_GROUP_dup, EC_POINT_dup.
10135 (These simply call ..._new and ..._copy).
10137 *Sheueling Chang Shantz and Douglas Stebila
10138 (Sun Microsystems Laboratories)*
10140 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10141 Polynomials are represented as BIGNUMs (where the sign bit is not
10142 used) in the following functions [macros]:
10145 BN_GF2m_sub [= BN_GF2m_add]
10146 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10147 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10148 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10150 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10151 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10152 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10153 BN_GF2m_cmp [= BN_ucmp]
10155 (Note that only the 'mod' functions are actually for fields GF(2^m).
10156 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10158 For some functions, an the irreducible polynomial defining a
10159 field can be given as an 'unsigned int[]' with strictly
10160 decreasing elements giving the indices of those bits that are set;
10161 i.e., p[] represents the polynomial
10162 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10164 p[0] > p[1] > ... > p[k] = 0.
10165 This applies to the following functions:
10168 BN_GF2m_mod_mul_arr
10169 BN_GF2m_mod_sqr_arr
10170 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10171 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10172 BN_GF2m_mod_exp_arr
10173 BN_GF2m_mod_sqrt_arr
10174 BN_GF2m_mod_solve_quad_arr
10178 Conversion can be performed by the following functions:
10183 bntest.c has additional tests for binary polynomial arithmetic.
10185 Two implementations for BN_GF2m_mod_div() are available.
10186 The default algorithm simply uses BN_GF2m_mod_inv() and
10187 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10188 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10189 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10191 *Sheueling Chang Shantz and Douglas Stebila
10192 (Sun Microsystems Laboratories)*
10194 * Add new error code 'ERR_R_DISABLED' that can be used when some
10195 functionality is disabled at compile-time.
10197 *Douglas Stebila <douglas.stebila@sun.com>*
10199 * Change default behaviour of 'openssl asn1parse' so that more
10200 information is visible when viewing, e.g., a certificate:
10202 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10203 mode the content of non-printable OCTET STRINGs is output in a
10204 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10205 avoid the appearance of a printable string.
10207 *Nils Larsch <nla@trustcenter.de>*
10209 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10211 EC_GROUP_set_asn1_flag()
10212 EC_GROUP_get_asn1_flag()
10213 EC_GROUP_set_point_conversion_form()
10214 EC_GROUP_get_point_conversion_form()
10215 These control ASN1 encoding details:
10216 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10217 has been set to OPENSSL_EC_NAMED_CURVE.
10218 - Points are encoded in uncompressed form by default; options for
10219 asn1_for are as for point2oct, namely
10220 POINT_CONVERSION_COMPRESSED
10221 POINT_CONVERSION_UNCOMPRESSED
10222 POINT_CONVERSION_HYBRID
10224 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10226 EC_GROUP_set_seed()
10227 EC_GROUP_get0_seed()
10228 EC_GROUP_get_seed_len()
10229 This is used only for ASN1 purposes (so far).
10231 *Nils Larsch <nla@trustcenter.de>*
10233 * Add 'field_type' member to EC_METHOD, which holds the NID
10234 of the appropriate field type OID. The new function
10235 EC_METHOD_get_field_type() returns this value.
10237 *Nils Larsch <nla@trustcenter.de>*
10240 EC_POINT_point2bn()
10241 EC_POINT_bn2point()
10242 EC_POINT_point2hex()
10243 EC_POINT_hex2point()
10244 providing useful interfaces to EC_POINT_point2oct() and
10245 EC_POINT_oct2point().
10247 *Nils Larsch <nla@trustcenter.de>*
10249 * Change internals of the EC library so that the functions
10250 EC_GROUP_set_generator()
10251 EC_GROUP_get_generator()
10252 EC_GROUP_get_order()
10253 EC_GROUP_get_cofactor()
10254 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10255 to methods, which would lead to unnecessary code duplication when
10256 adding different types of curves.
10258 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10260 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10261 arithmetic, and such that modified wNAFs are generated
10262 (which avoid length expansion in many cases).
10266 * Add a function EC_GROUP_check_discriminant() (defined via
10267 EC_METHOD) that verifies that the curve discriminant is non-zero.
10269 Add a function EC_GROUP_check() that makes some sanity tests
10270 on a EC_GROUP, its generator and order. This includes
10271 EC_GROUP_check_discriminant().
10273 *Nils Larsch <nla@trustcenter.de>*
10275 * Add ECDSA in new directory crypto/ecdsa/.
10277 Add applications 'openssl ecparam' and 'openssl ecdsa'
10278 (these are based on 'openssl dsaparam' and 'openssl dsa').
10280 ECDSA support is also included in various other files across the
10281 library. Most notably,
10282 - 'openssl req' now has a '-newkey ecdsa:file' option;
10283 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10284 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10285 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10286 them suitable for ECDSA where domain parameters must be
10287 extracted before the specific public key;
10288 - ECDSA engine support has been added.
10290 *Nils Larsch <nla@trustcenter.de>*
10292 * Include some named elliptic curves, and add OIDs from X9.62,
10293 SECG, and WAP/WTLS. Each curve can be obtained from the new
10295 EC_GROUP_new_by_curve_name(),
10296 and the list of available named curves can be obtained with
10297 EC_get_builtin_curves().
10298 Also add a 'curve_name' member to EC_GROUP objects, which can be
10300 EC_GROUP_set_curve_name()
10301 EC_GROUP_get_curve_name()
10303 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10305 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10306 was actually never needed) and in BN_mul(). The removal in BN_mul()
10307 required a small change in bn_mul_part_recursive() and the addition
10308 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10309 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10310 bn_sub_words() and bn_add_words() except they take arrays with
10315 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007] ###
10317 * Cleanse PEM buffers before freeing them since they may contain
10320 *Benjamin Bennett <ben@psc.edu>*
10322 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10323 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10324 authentication-only ciphersuites.
10328 * Since AES128 and AES256 share a single mask bit in the logic of
10329 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10330 kludge to work properly if AES128 is available and AES256 isn't.
10334 * Expand security boundary to match 1.1.1 module.
10338 * Remove redundant features: hash file source, editing of test vectors
10339 modify fipsld to use external fips_premain.c signature.
10343 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10344 run algorithm test programs.
10348 * Make algorithm test programs more tolerant of whitespace.
10352 * Have SSL/TLS server implementation tolerate "mismatched" record
10353 protocol version while receiving ClientHello even if the
10354 ClientHello is fragmented. (The server can't insist on the
10355 particular protocol version it has chosen before the ServerHello
10356 message has informed the client about his choice.)
10360 * Load error codes if they are not already present instead of using a
10361 static variable. This allows them to be cleanly unloaded and reloaded.
10365 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006] ###
10367 * Introduce limits to prevent malicious keys being able to
10368 cause a denial of service. [CVE-2006-2940][]
10370 *Steve Henson, Bodo Moeller*
10372 * Fix ASN.1 parsing of certain invalid structures that can result
10373 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10375 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10376 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10378 * Fix SSL client code which could crash if connecting to a
10379 malicious SSLv2 server. [CVE-2006-4343][]
10381 *Tavis Ormandy and Will Drewry, Google Security Team*
10383 * Change ciphersuite string processing so that an explicit
10384 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10385 will no longer include "AES128-SHA"), and any other similar
10386 ciphersuite (same bitmap) from *other* protocol versions (so that
10387 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10388 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10389 changes from 0.9.8b and 0.9.8d.
10393 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006] ###
10395 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10396 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10398 * Change the Unix randomness entropy gathering to use poll() when
10399 possible instead of select(), since the latter has some
10400 undesirable limitations.
10402 *Darryl Miles via Richard Levitte and Bodo Moeller*
10404 * Disable rogue ciphersuites:
10406 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10407 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10408 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10410 The latter two were purportedly from
10411 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10414 Also deactivate the remaining ciphersuites from
10415 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10416 unofficial, and the ID has long expired.
10420 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10421 dual-core machines) and other potential thread-safety issues.
10425 ### Changes between 0.9.7i and 0.9.7j [04 May 2006] ###
10427 * Adapt fipsld and the build system to link against the validated FIPS
10428 module in FIPS mode.
10432 * Fixes for VC++ 2005 build under Windows.
10436 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10437 from a Windows bash shell such as MSYS. It is autodetected from the
10438 "config" script when run from a VC++ environment. Modify standard VC++
10439 build to use fipscanister.o from the GNU make build.
10443 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005] ###
10445 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10446 The value now differs depending on if you build for FIPS or not.
10447 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10448 safely run with a non-FIPSed libcrypto, as it may crash because of
10449 the difference induced by this change.
10453 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005] ###
10455 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10456 (part of SSL_OP_ALL). This option used to disable the
10457 countermeasure against man-in-the-middle protocol-version
10458 rollback in the SSL 2.0 server implementation, which is a bad
10459 idea. [CVE-2005-2969][]
10461 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10462 for Information Security, National Institute of Advanced Industrial
10463 Science and Technology [AIST*, Japan)]
10465 * Minimal support for X9.31 signatures and PSS padding modes. This is
10466 mainly for FIPS compliance and not fully integrated at this stage.
10470 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10471 the exponentiation using a fixed-length exponent. (Otherwise,
10472 the information leaked through timing could expose the secret key
10473 after many signatures; cf. Bleichenbacher's attack on DSA with
10478 * Make a new fixed-window mod_exp implementation the default for
10479 RSA, DSA, and DH private-key operations so that the sequence of
10480 squares and multiplies and the memory access pattern are
10481 independent of the particular secret key. This will mitigate
10482 cache-timing and potential related attacks.
10484 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10485 and this is automatically used by BN_mod_exp_mont() if the new flag
10486 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10487 will use this BN flag for private exponents unless the flag
10488 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10489 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10492 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10494 * Change the client implementation for SSLv23_method() and
10495 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10496 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10497 (Previously, the SSL 2.0 backwards compatible Client Hello
10498 message format would be used even with SSL_OP_NO_SSLv2.)
10502 * Add support for smime-type MIME parameter in S/MIME messages which some
10507 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10508 a threadsafe manner. Modify rsa code to use new function and add calls
10509 to dsa and dh code (which had race conditions before).
10513 * Include the fixed error library code in the C error file definitions
10514 instead of fixing them up at runtime. This keeps the error code
10515 structures constant.
10519 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005] ###
10521 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10524 * Fixes for newer kerberos headers. NB: the casts are needed because
10525 the 'length' field is signed on one version and unsigned on another
10526 with no (?) obvious way to tell the difference, without these VC++
10527 complains. Also the "definition" of FAR (blank) is no longer included
10528 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10529 some needed definitions.
10533 * Undo Cygwin change.
10537 * Added support for proxy certificates according to RFC 3820.
10538 Because they may be a security thread to unaware applications,
10539 they must be explicitly allowed in run-time. See
10540 docs/HOWTO/proxy_certificates.txt for further information.
10544 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005] ###
10546 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10547 server and client random values. Previously
10548 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10549 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10551 This change has negligible security impact because:
10553 1. Server and client random values still have 24 bytes of pseudo random
10556 2. Server and client random values are sent in the clear in the initial
10559 3. The master secret is derived using the premaster secret (48 bytes in
10560 size for static RSA ciphersuites) as well as client server and random
10563 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10566 *Stephen Henson, reported by UK NISCC*
10568 * Use Windows randomness collection on Cygwin.
10572 * Fix hang in EGD/PRNGD query when communication socket is closed
10573 prematurely by EGD/PRNGD.
10575 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10577 * Prompt for pass phrases when appropriate for PKCS12 input format.
10581 * Back-port of selected performance improvements from development
10582 branch, as well as improved support for PowerPC platforms.
10586 * Add lots of checks for memory allocation failure, error codes to indicate
10587 failure and freeing up memory if a failure occurs.
10589 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10591 * Add new -passin argument to dgst.
10595 * Perform some character comparisons of different types in X509_NAME_cmp:
10596 this is needed for some certificates that re-encode DNs into UTF8Strings
10597 (in violation of RFC3280) and can't or won't issue name rollover
10602 * Make an explicit check during certificate validation to see that
10603 the CA setting in each certificate on the chain is correct. As a
10604 side effect always do the following basic checks on extensions,
10605 not just when there's an associated purpose to the check:
10607 - if there is an unhandled critical extension (unless the user
10608 has chosen to ignore this fault)
10609 - if the path length has been exceeded (if one is set at all)
10610 - that certain extensions fit the associated purpose (if one has
10615 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004] ###
10617 * Avoid a race condition when CRLs are checked in a multi threaded
10618 environment. This would happen due to the reordering of the revoked
10619 entries during signature checking and serial number lookup. Now the
10620 encoding is cached and the serial number sort performed under a lock.
10621 Add new STACK function sk_is_sorted().
10625 * Add Delta CRL to the extension code.
10629 * Various fixes to s3_pkt.c so alerts are sent properly.
10631 *David Holmes <d.holmes@f5.com>*
10633 * Reduce the chances of duplicate issuer name and serial numbers (in
10634 violation of RFC3280) using the OpenSSL certificate creation utilities.
10635 This is done by creating a random 64 bit value for the initial serial
10636 number when a serial number file is created or when a self signed
10637 certificate is created using 'openssl req -x509'. The initial serial
10638 number file is created using 'openssl x509 -next_serial' in CA.pl
10639 rather than being initialized to 1.
10643 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004] ###
10645 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10646 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10648 *Joe Orton, Steve Henson*
10650 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10653 *Joe Orton, Steve Henson*
10655 * Make it possible to have multiple active certificates with the same
10656 subject in the CA index file. This is done only if the keyword
10657 'unique_subject' is set to 'no' in the main CA section (default
10658 if 'CA_default') of the configuration file. The value is saved
10659 with the database itself in a separate index attribute file,
10660 named like the index file with '.attr' appended to the name.
10664 * X509 verify fixes. Disable broken certificate workarounds when
10665 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10666 keyUsage extension present. Don't accept CRLs with unhandled critical
10667 extensions: since verify currently doesn't process CRL extensions this
10668 rejects a CRL with *any* critical extensions. Add new verify error codes
10673 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10674 A clarification of RFC2560 will require the use of OCTET STRINGs and
10675 some implementations cannot handle the current raw format. Since OpenSSL
10676 copies and compares OCSP nonces as opaque blobs without any attempt at
10677 parsing them this should not create any compatibility issues.
10681 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10682 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10683 this HMAC (and other) operations are several times slower than OpenSSL
10688 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10690 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10692 * Use the correct content when signing type "other".
10696 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003] ###
10698 * Fix various bugs revealed by running the NISCC test suite:
10700 Stop out of bounds reads in the ASN1 code when presented with
10701 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10703 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10705 If verify callback ignores invalid public key errors don't try to check
10706 certificate signature with the NULL public key.
10711 * New -ignore_err option in ocsp application to stop the server
10712 exiting on the first error in a request.
10716 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10717 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10722 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10723 extra data after the compression methods not only for TLS 1.0
10724 but also for SSL 3.0 (as required by the specification).
10726 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10728 * Change X509_certificate_type() to mark the key as exported/exportable
10729 when it's 512 *bits* long, not 512 bytes.
10733 * Change AES_cbc_encrypt() so it outputs exact multiple of
10734 blocks during encryption.
10738 * Various fixes to base64 BIO and non blocking I/O. On write
10739 flushes were not handled properly if the BIO retried. On read
10740 data was not being buffered properly and had various logic bugs.
10741 This also affects blocking I/O when the data being decoded is a
10746 * Various S/MIME bugfixes and compatibility changes:
10747 output correct application/pkcs7 MIME type if
10748 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10749 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10750 of files as .eml work). Correctly handle very long lines in MIME
10755 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003] ###
10757 * Countermeasure against the Klima-Pokorny-Rosa extension of
10758 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10759 a protocol version number mismatch like a decryption error
10760 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10764 * Turn on RSA blinding by default in the default implementation
10765 to avoid a timing attack. Applications that don't want it can call
10766 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10767 They would be ill-advised to do so in most cases.
10769 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10771 * Change RSA blinding code so that it works when the PRNG is not
10772 seeded (in this case, the secret RSA exponent is abused as
10773 an unpredictable seed -- if it is not unpredictable, there
10774 is no point in blinding anyway). Make RSA blinding thread-safe
10775 by remembering the creator's thread ID in rsa->blinding and
10776 having all other threads use local one-time blinding factors
10777 (this requires more computation than sharing rsa->blinding, but
10778 avoids excessive locking; and if an RSA object is not shared
10779 between threads, blinding will still be very fast).
10783 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10784 ENGINE as defaults for all supported algorithms irrespective of
10785 the 'flags' parameter. 'flags' is now honoured, so applications
10786 should make sure they are passing it correctly.
10790 * Target "mingw" now allows native Windows code to be generated in
10791 the Cygwin environment as well as with the MinGW compiler.
10795 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003] ###
10797 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10798 via timing by performing a MAC computation even if incorrect
10799 block cipher padding has been found. This is a countermeasure
10800 against active attacks where the attacker has to distinguish
10801 between bad padding and a MAC verification error. [CVE-2003-0078][]
10803 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10804 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10805 Martin Vuagnoux (EPFL, Ilion)*
10807 * Make the no-err option work as intended. The intention with no-err
10808 is not to have the whole error stack handling routines removed from
10809 libcrypto, it's only intended to remove all the function name and
10810 reason texts, thereby removing some of the footprint that may not
10811 be interesting if those errors aren't displayed anyway.
10813 NOTE: it's still possible for any application or module to have its
10814 own set of error texts inserted. The routines are there, just not
10815 used by default when no-err is given.
10819 * Add support for FreeBSD on IA64.
10821 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10823 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10824 Kerberos function mit_des_cbc_cksum(). Before this change,
10825 the value returned by DES_cbc_cksum() was like the one from
10826 mit_des_cbc_cksum(), except the bytes were swapped.
10828 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10830 * Allow an application to disable the automatic SSL chain building.
10831 Before this a rather primitive chain build was always performed in
10832 ssl3_output_cert_chain(): an application had no way to send the
10833 correct chain if the automatic operation produced an incorrect result.
10835 Now the chain builder is disabled if either:
10837 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10839 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10841 The reasoning behind this is that an application would not want the
10842 auto chain building to take place if extra chain certificates are
10843 present and it might also want a means of sending no additional
10844 certificates (for example the chain has two certificates and the
10849 * Add the possibility to build without the ENGINE framework.
10851 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10853 * Under Win32 gmtime() can return NULL: check return value in
10854 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10858 * DSA routines: under certain error conditions uninitialized BN objects
10859 could be freed. Solution: make sure initialization is performed early
10860 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10861 Nils Larsch <nla@trustcenter.de> via PR#459)
10865 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10866 checked on reconnect on the client side, therefore session resumption
10867 could still fail with a "ssl session id is different" error. This
10868 behaviour is masked when SSL_OP_ALL is used due to
10869 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10870 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10871 followup to PR #377.
10875 * IA-32 assembler support enhancements: unified ELF targets, support
10876 for SCO/Caldera platforms, fix for Cygwin shared build.
10880 * Add support for FreeBSD on sparc64. As a consequence, support for
10881 FreeBSD on non-x86 processors is separate from x86 processors on
10882 the config script, much like the NetBSD support.
10884 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10886 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002] ###
10888 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10891 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10892 code (06) was taken as the first octet of the session ID and the last
10893 octet was ignored consequently. As a result SSLv2 client side session
10894 caching could not have worked due to the session ID mismatch between
10896 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10901 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10902 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10907 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10908 seems that in spite of existing for more than a year, many application
10909 author have done nothing to provide the necessary callbacks, which
10910 means that this particular engine will not work properly anywhere.
10911 This is a very unfortunate situation which forces us, in the name
10912 of usability, to give the hw_ncipher.c a static lock, which is part
10914 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10915 appear in 0.9.8 or later. We EXPECT application authors to have
10916 dealt properly with this when 0.9.8 is released (unless we actually
10917 make such changes in the libcrypto locking code that changes will
10918 have to be made anyway).
10922 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
10923 octets have been read, EOF or an error occurs. Without this change
10924 some truncated ASN1 structures will not produce an error.
10928 * Disable Heimdal support, since it hasn't been fully implemented.
10929 Still give the possibility to force the use of Heimdal, but with
10930 warnings and a request that patches get sent to openssl-dev.
10934 * Add the VC-CE target, introduce the WINCE sysname, and add
10935 INSTALL.WCE and appropriate conditionals to make it build.
10937 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10939 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
10940 cygssl-x.y.z.dll, where x, y and z are the major, minor and
10941 edit numbers of the version.
10943 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10945 * Introduce safe string copy and catenation functions
10946 (BUF_strlcpy() and BUF_strlcat()).
10948 *Ben Laurie (CHATS) and Richard Levitte*
10950 * Avoid using fixed-size buffers for one-line DNs.
10952 *Ben Laurie (CHATS)*
10954 * Add BUF_MEM_grow_clean() to avoid information leakage when
10955 resizing buffers containing secrets, and use where appropriate.
10957 *Ben Laurie (CHATS)*
10959 * Avoid using fixed size buffers for configuration file location.
10961 *Ben Laurie (CHATS)*
10963 * Avoid filename truncation for various CA files.
10965 *Ben Laurie (CHATS)*
10967 * Use sizeof in preference to magic numbers.
10969 *Ben Laurie (CHATS)*
10971 * Avoid filename truncation in cert requests.
10973 *Ben Laurie (CHATS)*
10975 * Add assertions to check for (supposedly impossible) buffer
10978 *Ben Laurie (CHATS)*
10980 * Don't cache truncated DNS entries in the local cache (this could
10981 potentially lead to a spoofing attack).
10983 *Ben Laurie (CHATS)*
10985 * Fix various buffers to be large enough for hex/decimal
10986 representations in a platform independent manner.
10988 *Ben Laurie (CHATS)*
10990 * Add CRYPTO_realloc_clean() to avoid information leakage when
10991 resizing buffers containing secrets, and use where appropriate.
10993 *Ben Laurie (CHATS)*
10995 * Add BIO_indent() to avoid much slightly worrying code to do
10998 *Ben Laurie (CHATS)*
11000 * Convert sprintf()/BIO_puts() to BIO_printf().
11002 *Ben Laurie (CHATS)*
11004 * buffer_gets() could terminate with the buffer only half
11007 *Ben Laurie (CHATS)*
11009 * Add assertions to prevent user-supplied crypto functions from
11010 overflowing internal buffers by having large block sizes, etc.
11012 *Ben Laurie (CHATS)*
11014 * New OPENSSL_assert() macro (similar to assert(), but enabled
11017 *Ben Laurie (CHATS)*
11019 * Eliminate unused copy of key in RC4.
11021 *Ben Laurie (CHATS)*
11023 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11025 *Ben Laurie (CHATS)*
11027 * Fix off-by-one error in EGD path.
11029 *Ben Laurie (CHATS)*
11031 * If RANDFILE path is too long, ignore instead of truncating.
11033 *Ben Laurie (CHATS)*
11035 * Eliminate unused and incorrectly sized X.509 structure
11038 *Ben Laurie (CHATS)*
11040 * Eliminate unused and dangerous function knumber().
11042 *Ben Laurie (CHATS)*
11044 * Eliminate unused and dangerous structure, KSSL_ERR.
11046 *Ben Laurie (CHATS)*
11048 * Protect against overlong session ID context length in an encoded
11049 session object. Since these are local, this does not appear to be
11052 *Ben Laurie (CHATS)*
11054 * Change from security patch (see 0.9.6e below) that did not affect
11055 the 0.9.6 release series:
11057 Remote buffer overflow in SSL3 protocol - an attacker could
11058 supply an oversized master key in Kerberos-enabled versions.
11061 *Ben Laurie (CHATS)*
11063 * Change the SSL kerb5 codes to match RFC 2712.
11067 * Make -nameopt work fully for req and add -reqopt switch.
11069 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11071 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11073 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11075 * Make sure tests can be performed even if the corresponding algorithms
11076 have been removed entirely. This was also the last step to make
11077 OpenSSL compilable with DJGPP under all reasonable conditions.
11079 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11081 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11082 to allow version independent disabling of normally unselected ciphers,
11083 which may be activated as a side-effect of selecting a single cipher.
11085 (E.g., cipher list string "RSA" enables ciphersuites that are left
11086 out of "ALL" because they do not provide symmetric encryption.
11087 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11089 *Lutz Jaenicke, Bodo Moeller*
11091 * Add appropriate support for separate platform-dependent build
11092 directories. The recommended way to make a platform-dependent
11093 build directory is the following (tested on Linux), maybe with
11096 # Place yourself outside of the OpenSSL source tree. In
11097 # this example, the environment variable OPENSSL_SOURCE
11098 # is assumed to contain the absolute OpenSSL source directory.
11099 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11100 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11101 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11102 mkdir -p `dirname $F`
11103 ln -s $OPENSSL_SOURCE/$F $F
11106 To be absolutely sure not to disturb the source tree, a "make clean"
11107 is a good thing. If it isn't successful, don't worry about it,
11108 it probably means the source directory is very clean.
11112 * Make sure any ENGINE control commands make local copies of string
11113 pointers passed to them whenever necessary. Otherwise it is possible
11114 the caller may have overwritten (or deallocated) the original string
11115 data when a later ENGINE operation tries to use the stored values.
11117 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11119 * Improve diagnostics in file reading and command-line digests.
11121 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11123 * Add AES modes CFB and OFB to the object database. Correct an
11124 error in AES-CFB decryption.
11128 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11129 allows existing EVP_CIPHER_CTX structures to be reused after
11130 calling EVP_*Final(). This behaviour is used by encryption
11131 BIOs and some applications. This has the side effect that
11132 applications must explicitly clean up cipher contexts with
11133 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11137 * Check the values of dna and dnb in bn_mul_recursive before calling
11138 bn_mul_comba (a non zero value means the a or b arrays do not contain
11139 n2 elements) and fallback to bn_mul_normal if either is not zero.
11143 * Fix escaping of non-ASCII characters when using the -subj option
11144 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11148 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11149 form for "surname", serialNumber has no short form.
11150 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11151 therefore remove "mail" short name for "internet 7".
11152 The OID for unique identifiers in X509 certificates is
11153 x500UniqueIdentifier, not uniqueIdentifier.
11154 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11158 * Add an "init" command to the ENGINE config module and auto initialize
11159 ENGINEs. Without any "init" command the ENGINE will be initialized
11160 after all ctrl commands have been executed on it. If init=1 the
11161 ENGINE is initialized at that point (ctrls before that point are run
11162 on the uninitialized ENGINE and after on the initialized one). If
11163 init=0 then the ENGINE will not be initialized at all.
11167 * Fix the 'app_verify_callback' interface so that the user-defined
11168 argument is actually passed to the callback: In the
11169 SSL_CTX_set_cert_verify_callback() prototype, the callback
11170 declaration has been changed from
11173 int (*cb)(X509_STORE_CTX *,void *);
11174 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11175 i=s->ctx->app_verify_callback(&ctx)
11176 has been changed into
11177 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11179 To update applications using SSL_CTX_set_cert_verify_callback(),
11180 a dummy argument can be added to their callback functions.
11182 *D. K. Smetters <smetters@parc.xerox.com>*
11184 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11186 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11188 * Add and OPENSSL_LOAD_CONF define which will cause
11189 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11190 This allows older applications to transparently support certain
11191 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11192 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11193 load the config file and OPENSSL_add_all_algorithms_conf() which will
11194 always load it have also been added.
11198 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11199 Adjust NIDs and EVP layer.
11201 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11203 * Config modules support in openssl utility.
11205 Most commands now load modules from the config file,
11206 though in a few (such as version) this isn't done
11207 because it couldn't be used for anything.
11209 In the case of ca and req the config file used is
11210 the same as the utility itself: that is the -config
11211 command line option can be used to specify an
11216 * Move default behaviour from OPENSSL_config(). If appname is NULL
11217 use "openssl_conf" if filename is NULL use default openssl config file.
11221 * Add an argument to OPENSSL_config() to allow the use of an alternative
11222 config section name. Add a new flag to tolerate a missing config file
11223 and move code to CONF_modules_load_file().
11227 * Support for crypto accelerator cards from Accelerated Encryption
11228 Processing, www.aep.ie. (Use engine 'aep')
11229 The support was copied from 0.9.6c [engine] and adapted/corrected
11230 to work with the new engine framework.
11232 *AEP Inc. and Richard Levitte*
11234 * Support for SureWare crypto accelerator cards from Baltimore
11235 Technologies. (Use engine 'sureware')
11236 The support was copied from 0.9.6c [engine] and adapted
11237 to work with the new engine framework.
11241 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11242 make the newer ENGINE framework commands for the CHIL engine work.
11244 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11246 * Make it possible to produce shared libraries on ReliantUNIX.
11248 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11250 * Add the configuration target debug-linux-ppro.
11251 Make 'openssl rsa' use the general key loading routines
11252 implemented in apps.c, and make those routines able to
11253 handle the key format FORMAT_NETSCAPE and the variant
11256 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11258 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11260 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11262 * Add -keyform to rsautl, and document -engine.
11264 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11266 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11267 BIO_R_NO_SUCH_FILE error code rather than the generic
11268 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11272 * Add new functions
11273 ERR_peek_last_error
11274 ERR_peek_last_error_line
11275 ERR_peek_last_error_line_data.
11276 These are similar to
11278 ERR_peek_error_line
11279 ERR_peek_error_line_data,
11280 but report on the latest error recorded rather than the first one
11281 still in the error queue.
11283 *Ben Laurie, Bodo Moeller*
11285 * default_algorithms option in ENGINE config module. This allows things
11287 default_algorithms = ALL
11288 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11292 * Preliminary ENGINE config module.
11296 * New experimental application configuration code.
11300 * Change the AES code to follow the same name structure as all other
11301 symmetric ciphers, and behave the same way. Move everything to
11302 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11304 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11306 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11308 *Ben Laurie and Theo de Raadt*
11310 * Add option to output public keys in req command.
11312 *Massimiliano Pala madwolf@openca.org*
11314 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11315 (up to about 10% better than before for P-192 and P-224).
11319 * New functions/macros
11321 SSL_CTX_set_msg_callback(ctx, cb)
11322 SSL_CTX_set_msg_callback_arg(ctx, arg)
11323 SSL_set_msg_callback(ssl, cb)
11324 SSL_set_msg_callback_arg(ssl, arg)
11326 to request calling a callback function
11328 void cb(int write_p, int version, int content_type,
11329 const void *buf, size_t len, SSL *ssl, void *arg)
11331 whenever a protocol message has been completely received
11332 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11333 protocol version according to which the SSL library interprets
11334 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11335 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11336 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11337 specification (change_cipher_spec(20), alert(21), handshake(22)).
11338 'buf' and 'len' point to the actual message, 'ssl' to the
11339 SSL object, and 'arg' is the application-defined value set by
11340 SSL[_CTX]_set_msg_callback_arg().
11342 'openssl s_client' and 'openssl s_server' have new '-msg' options
11343 to enable a callback that displays all protocol messages.
11347 * Change the shared library support so shared libraries are built as
11348 soon as the corresponding static library is finished, and thereby get
11349 openssl and the test programs linked against the shared library.
11350 This still only happens when the keyword "shard" has been given to
11351 the configuration scripts.
11353 NOTE: shared library support is still an experimental thing, and
11354 backward binary compatibility is still not guaranteed.
11356 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11358 * Add support for Subject Information Access extension.
11360 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11362 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11363 additional bytes when new memory had to be allocated, not just
11364 when reusing an existing buffer.
11368 * New command line and configuration option 'utf8' for the req command.
11369 This allows field values to be specified as UTF8 strings.
11373 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11374 runs for the former and machine-readable output for the latter.
11378 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11379 of the e-mail address in the DN (i.e., it will go into a certificate
11380 extension only). The new configuration file option 'email_in_dn = no'
11381 has the same effect.
11383 *Massimiliano Pala madwolf@openca.org*
11385 * Change all functions with names starting with des_ to be starting
11386 with DES_ instead. Add wrappers that are compatible with libdes,
11387 but are named _ossl_old_des_*. Finally, add macros that map the
11388 des_* symbols to the corresponding _ossl_old_des_* if libdes
11389 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11390 desired, the des_* symbols will be mapped to DES_*, with one
11393 Since we provide two compatibility mappings, the user needs to
11394 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11395 compatibility is desired. The default (i.e., when that macro
11396 isn't defined) is OpenSSL 0.9.6c compatibility.
11398 There are also macros that enable and disable the support of old
11399 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11400 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11401 are defined, the default will apply: to support the old des routines.
11403 In either case, one must include openssl/des.h to get the correct
11404 definitions. Do not try to just include openssl/des_old.h, that
11407 NOTE: This is a major break of an old API into a new one. Software
11408 authors are encouraged to switch to the DES_ style functions. Some
11409 time in the future, des_old.h and the libdes compatibility functions
11410 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11411 default), and then completely removed.
11415 * Test for certificates which contain unsupported critical extensions.
11416 If such a certificate is found during a verify operation it is
11417 rejected by default: this behaviour can be overridden by either
11418 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11419 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11420 X509_supported_extension() has also been added which returns 1 if a
11421 particular extension is supported.
11425 * Modify the behaviour of EVP cipher functions in similar way to digests
11426 to retain compatibility with existing code.
11430 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11431 compatibility with existing code. In particular the 'ctx' parameter does
11432 not have to be to be initialized before the call to EVP_DigestInit() and
11433 it is tidied up after a call to EVP_DigestFinal(). New function
11434 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11435 EVP_MD_CTX_copy() changed to not require the destination to be
11436 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11437 requires the destination to be valid.
11439 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11440 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11444 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11445 so that complete 'Handshake' protocol structures are kept in memory
11446 instead of overwriting 'msg_type' and 'length' with 'body' data.
11450 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11452 *Massimo Santin via Richard Levitte*
11454 * Major restructuring to the underlying ENGINE code. This includes
11455 reduction of linker bloat, separation of pure "ENGINE" manipulation
11456 (initialisation, etc) from functionality dealing with implementations
11457 of specific crypto interfaces. This change also introduces integrated
11458 support for symmetric ciphers and digest implementations - so ENGINEs
11459 can now accelerate these by providing EVP_CIPHER and EVP_MD
11460 implementations of their own. This is detailed in crypto/engine/README
11461 as it couldn't be adequately described here. However, there are a few
11462 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11463 were changed in the original introduction of ENGINE code have now
11464 reverted back - the hooking from this code to ENGINE is now a good
11465 deal more passive and at run-time, operations deal directly with
11466 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11467 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11468 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
11469 they were not being used by the framework as there is no concept of a
11470 BIGNUM_METHOD and they could not be generalised to the new
11471 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11472 ENGINE_cpy() has been removed as it cannot be consistently defined in
11477 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11481 * Change mkdef.pl to sort symbols that get the same entry number,
11482 and make sure the automatically generated functions ERR_load_*
11483 become part of libeay.num as well.
11487 * New function SSL_renegotiate_pending(). This returns true once
11488 renegotiation has been requested (either SSL_renegotiate() call
11489 or HelloRequest/ClientHello received from the peer) and becomes
11490 false once a handshake has been completed.
11491 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11492 sends a HelloRequest, but does not ensure that a handshake takes
11493 place. SSL_renegotiate_pending() is useful for checking if the
11494 client has followed the request.)
11498 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11499 By default, clients may request session resumption even during
11500 renegotiation (if session ID contexts permit); with this option,
11501 session resumption is possible only in the first handshake.
11503 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11504 more bits available for options that should not be part of
11505 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11509 * Add some demos for certificate and certificate request creation.
11513 * Make maximum certificate chain size accepted from the peer application
11514 settable (SSL*_get/set_max_cert_list()), as proposed by
11515 "Douglas E. Engert" <deengert@anl.gov>.
11519 * Add support for shared libraries for Unixware-7
11520 (Boyd Lynn Gerber <gerberb@zenez.com>).
11524 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11525 be done prior to destruction. Use this to unload error strings from
11526 ENGINEs that load their own error strings. NB: This adds two new API
11527 functions to "get" and "set" this destroy handler in an ENGINE.
11531 * Alter all existing ENGINE implementations (except "openssl" and
11532 "openbsd") to dynamically instantiate their own error strings. This
11533 makes them more flexible to be built both as statically-linked ENGINEs
11534 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11535 Also, add stub code to each that makes building them as self-contained
11536 shared-libraries easier (see README.ENGINE).
11540 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11541 implementations into applications that are completely implemented in
11542 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11543 commands that can be used to configure what shared-library to load and
11544 to control aspects of the way it is handled. Also, made an update to
11545 the README.ENGINE file that brings its information up-to-date and
11546 provides some information and instructions on the "dynamic" ENGINE
11547 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11551 * Make it possible to unload ranges of ERR strings with a new
11552 "ERR_unload_strings" function.
11556 * Add a copy() function to EVP_MD.
11560 * Make EVP_MD routines take a context pointer instead of just the
11561 md_data void pointer.
11565 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11566 that the digest can only process a single chunk of data
11567 (typically because it is provided by a piece of
11568 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11569 is only going to provide a single chunk of data, and hence the
11570 framework needn't accumulate the data for oneshot drivers.
11574 * As with "ERR", make it possible to replace the underlying "ex_data"
11575 functions. This change also alters the storage and management of global
11576 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11577 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11578 index counters. The API functions that use this state have been changed
11579 to take a "class_index" rather than pointers to the class's local STACK
11580 and counter, and there is now an API function to dynamically create new
11581 classes. This centralisation allows us to (a) plug a lot of the
11582 thread-safety problems that existed, and (b) makes it possible to clean
11583 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11584 such data would previously have always leaked in application code and
11585 workarounds were in place to make the memory debugging turn a blind eye
11586 to it. Application code that doesn't use this new function will still
11587 leak as before, but their memory debugging output will announce it now
11588 rather than letting it slide.
11590 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11591 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11592 has a return value to indicate success or failure.
11596 * Make it possible to replace the underlying "ERR" functions such that the
11597 global state (2 LHASH tables and 2 locks) is only used by the "default"
11598 implementation. This change also adds two functions to "get" and "set"
11599 the implementation prior to it being automatically set the first time
11600 any other ERR function takes place. Ie. an application can call "get",
11601 pass the return value to a module it has just loaded, and that module
11602 can call its own "set" function using that value. This means the
11603 module's "ERR" operations will use (and modify) the error state in the
11604 application and not in its own statically linked copy of OpenSSL code.
11608 * Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
11609 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11610 the operation, and provides a more encapsulated way for external code
11611 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11612 to use these functions rather than manually incrementing the counts.
11614 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11618 * Add EVP test program.
11622 * Add symmetric cipher support to ENGINE. Expect the API to change!
11626 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11627 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11628 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11629 These allow a CRL to be built without having to access X509_CRL fields
11630 directly. Modify 'ca' application to use new functions.
11634 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11635 bug workarounds. Rollback attack detection is a security feature.
11636 The problem will only arise on OpenSSL servers when TLSv1 is not
11637 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11638 Software authors not wanting to support TLSv1 will have special reasons
11639 for their choice and can explicitly enable this option.
11641 *Bodo Moeller, Lutz Jaenicke*
11643 * Rationalise EVP so it can be extended: don't include a union of
11644 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11645 (similar to those existing for EVP_CIPHER_CTX).
11650 EVP_MD_CTX_init(&md); /* new function call */
11651 EVP_DigestInit(&md, EVP_sha1());
11652 EVP_DigestUpdate(&md, in, len);
11653 EVP_DigestFinal(&md, out, NULL);
11654 EVP_MD_CTX_cleanup(&md); /* new function call */
11659 * Make DES key schedule conform to the usual scheme, as well as
11660 correcting its structure. This means that calls to DES functions
11661 now have to pass a pointer to a des_key_schedule instead of a
11662 plain des_key_schedule (which was actually always a pointer
11665 des_key_schedule ks;
11667 des_set_key_checked(..., &ks);
11668 des_ncbc_encrypt(..., &ks, ...);
11670 (Note that a later change renames 'des_...' into 'DES_...'.)
11674 * Initial reduction of linker bloat: the use of some functions, such as
11675 PEM causes large amounts of unused functions to be linked in due to
11676 poor organisation. For example pem_all.c contains every PEM function
11677 which has a knock on effect of linking in large amounts of (unused)
11678 ASN1 code. Grouping together similar functions and splitting unrelated
11679 functions prevents this.
11683 * Cleanup of EVP macros.
11687 * Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
11688 correct _ecb suffix.
11692 * Add initial OCSP responder support to ocsp application. The
11693 revocation information is handled using the text based index
11694 use by the ca application. The responder can either handle
11695 requests generated internally, supplied in files (for example
11696 via a CGI script) or using an internal minimal server.
11700 * Add configuration choices to get zlib compression for TLS.
11704 * Changes to Kerberos SSL for RFC 2712 compliance:
11705 1. Implemented real KerberosWrapper, instead of just using
11706 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11707 2. Implemented optional authenticator field of KerberosWrapper.
11709 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11710 and authenticator structs; see crypto/krb5/.
11712 Generalized Kerberos calls to support multiple Kerberos libraries.
11713 *Vern Staats <staatsvr@asc.hpc.mil>,
11714 Jeffrey Altman <jaltman@columbia.edu>
11715 via Richard Levitte*
11717 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11718 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11719 values for each of the key sizes rather than having just
11720 parameters (and 'speed' generating keys each time).
11724 * Speed up EVP routines.
11727 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11728 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11729 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11730 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11732 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11733 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11734 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11737 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11739 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11743 * Added the OS2-EMX target.
11745 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11747 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11748 to support NCONF routines in extension code. New function CONF_set_nconf()
11749 to allow functions which take an NCONF to also handle the old LHASH
11750 structure: this means that the old CONF compatible routines can be
11751 retained (in particular wrt extensions) without having to duplicate the
11752 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11756 * Enhance the general user interface with mechanisms for inner control
11757 and with possibilities to have yes/no kind of prompts.
11761 * Change all calls to low level digest routines in the library and
11762 applications to use EVP. Add missing calls to HMAC_cleanup() and
11763 don't assume HMAC_CTX can be copied using memcpy().
11765 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11767 * Add the possibility to control engines through control names but with
11768 arbitrary arguments instead of just a string.
11769 Change the key loaders to take a UI_METHOD instead of a callback
11770 function pointer. NOTE: this breaks binary compatibility with earlier
11771 versions of OpenSSL [engine].
11772 Adapt the nCipher code for these new conditions and add a card insertion
11777 * Enhance the general user interface with mechanisms to better support
11778 dialog box interfaces, application-defined prompts, the possibility
11779 to use defaults (for example default passwords from somewhere else)
11780 and interrupts/cancellations.
11784 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11785 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11789 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11790 tidy up some unnecessarily weird code in 'sk_new()').
11792 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11794 * Change the key loading routines for ENGINEs to use the same kind
11795 callback (pem_password_cb) as all other routines that need this
11800 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11801 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11802 than this minimum value is recommended.
11806 * New random seeder for OpenVMS, using the system process statistics
11807 that are easily reachable.
11811 * Windows apparently can't transparently handle global
11812 variables defined in DLLs. Initialisations such as:
11814 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11816 won't compile. This is used by the any applications that need to
11817 declare their own ASN1 modules. This was fixed by adding the option
11818 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11819 needed for static libraries under Win32.
11823 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11824 setting of purpose and trust fields. New X509_STORE trust and
11825 purpose functions and tidy up setting in other SSL functions.
11829 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11830 structure. These are inherited by X509_STORE_CTX when it is
11831 initialised. This allows various defaults to be set in the
11832 X509_STORE structure (such as flags for CRL checking and custom
11833 purpose or trust settings) for functions which only use X509_STORE_CTX
11834 internally such as S/MIME.
11836 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11837 trust settings if they are not set in X509_STORE. This allows X509_STORE
11838 purposes and trust (in S/MIME for example) to override any set by default.
11840 Add command line options for CRL checking to smime, s_client and s_server
11845 * Initial CRL based revocation checking. If the CRL checking flag(s)
11846 are set then the CRL is looked up in the X509_STORE structure and
11847 its validity and signature checked, then if the certificate is found
11848 in the CRL the verify fails with a revoked error.
11850 Various new CRL related callbacks added to X509_STORE_CTX structure.
11852 Command line options added to 'verify' application to support this.
11854 This needs some additional work, such as being able to handle multiple
11855 CRLs with different times, extension based lookup (rather than just
11856 by subject name) and ultimately more complete V2 CRL extension
11861 * Add a general user interface API (crypto/ui/). This is designed
11862 to replace things like des_read_password and friends (backward
11863 compatibility functions using this new API are provided).
11864 The purpose is to remove prompting functions from the DES code
11865 section as well as provide for prompting through dialog boxes in
11866 a window system and the like.
11870 * Add "ex_data" support to ENGINE so implementations can add state at a
11871 per-structure level rather than having to store it globally.
11875 * Make it possible for ENGINE structures to be copied when retrieved by
11876 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11877 This causes the "original" ENGINE structure to act like a template,
11878 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11879 operational state can be localised to each ENGINE structure, despite the
11880 fact they all share the same "methods". New ENGINE structures returned in
11881 this case have no functional references and the return value is the single
11882 structural reference. This matches the single structural reference returned
11883 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11888 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11889 needs to match any other type at all we need to manually clear the
11894 * Changes to the "openssl engine" utility to include;
11895 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11896 about an ENGINE's available control commands.
11897 - executing control commands from command line arguments using the
11898 '-pre' and '-post' switches. '-post' is only used if '-t' is
11899 specified and the ENGINE is successfully initialised. The syntax for
11900 the individual commands are colon-separated, for example;
11901 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11905 * New dynamic control command support for ENGINEs. ENGINEs can now
11906 declare their own commands (numbers), names (strings), descriptions,
11907 and input types for run-time discovery by calling applications. A
11908 subset of these commands are implicitly classed as "executable"
11909 depending on their input type, and only these can be invoked through
11910 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11911 can be based on user input, config files, etc). The distinction is
11912 that "executable" commands cannot return anything other than a boolean
11913 result and can only support numeric or string input, whereas some
11914 discoverable commands may only be for direct use through
11915 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11916 pointers, or other custom uses. The "executable" commands are to
11917 support parameterisations of ENGINE behaviour that can be
11918 unambiguously defined by ENGINEs and used consistently across any
11919 OpenSSL-based application. Commands have been added to all the
11920 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
11921 control over shared-library paths without source code alterations.
11925 * Changed all ENGINE implementations to dynamically allocate their
11926 ENGINEs rather than declaring them statically. Apart from this being
11927 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
11928 this also allows the implementations to compile without using the
11929 internal engine_int.h header.
11933 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
11934 'const' value. Any code that should be able to modify a RAND_METHOD
11935 should already have non-const pointers to it (ie. they should only
11936 modify their own ones).
11940 * Made a variety of little tweaks to the ENGINE code.
11941 - "atalla" and "ubsec" string definitions were moved from header files
11942 to C code. "nuron" string definitions were placed in variables
11943 rather than hard-coded - allowing parameterisation of these values
11944 later on via ctrl() commands.
11945 - Removed unused "#if 0"'d code.
11946 - Fixed engine list iteration code so it uses ENGINE_free() to release
11947 structural references.
11948 - Constified the RAND_METHOD element of ENGINE structures.
11949 - Constified various get/set functions as appropriate and added
11950 missing functions (including a catch-all ENGINE_cpy that duplicates
11951 all ENGINE values onto a new ENGINE except reference counts/state).
11952 - Removed NULL parameter checks in get/set functions. Setting a method
11953 or function to NULL is a way of cancelling out a previously set
11954 value. Passing a NULL ENGINE parameter is just plain stupid anyway
11955 and doesn't justify the extra error symbols and code.
11956 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
11957 flags from engine_int.h to engine.h.
11958 - Changed prototypes for ENGINE handler functions (init(), finish(),
11959 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
11963 * Implement binary inversion algorithm for BN_mod_inverse in addition
11964 to the algorithm using long division. The binary algorithm can be
11965 used only if the modulus is odd. On 32-bit systems, it is faster
11966 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
11967 roughly 5-15% for 256-bit moduli), so we use it only for moduli
11968 up to 450 bits. In 64-bit environments, the binary algorithm
11969 appears to be advantageous for much longer moduli; here we use it
11970 for moduli up to 2048 bits.
11974 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
11975 could not support the combine flag in choice fields.
11979 * Add a 'copy_extensions' option to the 'ca' utility. This copies
11980 extensions from a certificate request to the certificate.
11984 * Allow multiple 'certopt' and 'nameopt' options to be separated
11985 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
11986 file: this allows the display of the certificate about to be
11987 signed to be customised, to allow certain fields to be included
11988 or excluded and extension details. The old system didn't display
11989 multicharacter strings properly, omitted fields not in the policy
11990 and couldn't display additional details such as extensions.
11994 * Function EC_POINTs_mul for multiple scalar multiplication
11995 of an arbitrary number of elliptic curve points
11996 \sum scalars[i]*points[i],
11997 optionally including the generator defined for the EC_GROUP:
11998 scalar*generator + \sum scalars[i]*points[i].
12000 EC_POINT_mul is a simple wrapper function for the typical case
12001 that the point list has just one item (besides the optional
12006 * First EC_METHODs for curves over GF(p):
12008 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12009 operations and provides various method functions that can also
12010 operate with faster implementations of modular arithmetic.
12012 EC_GFp_mont_method() reuses most functions that are part of
12013 EC_GFp_simple_method, but uses Montgomery arithmetic.
12015 *Bodo Moeller; point addition and point doubling
12016 implementation directly derived from source code provided by
12017 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12019 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12020 crypto/ec/ec_lib.c):
12022 Curves are EC_GROUP objects (with an optional group generator)
12023 based on EC_METHODs that are built into the library.
12025 Points are EC_POINT objects based on EC_GROUP objects.
12027 Most of the framework would be able to handle curves over arbitrary
12028 finite fields, but as there are no obvious types for fields other
12029 than GF(p), some functions are limited to that for now.
12033 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12034 that the file contains a complete HTTP response.
12038 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12039 change the def and num file printf format specifier from "%-40sXXX"
12040 to "%-39s XXX". The latter will always guarantee a space after the
12041 field while the former will cause them to run together if the field
12042 is 40 of more characters long.
12046 * Constify the cipher and digest 'method' functions and structures
12047 and modify related functions to take constant EVP_MD and EVP_CIPHER
12052 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12053 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12057 * Modify EVP_Digest*() routines so they now return values. Although the
12058 internal software routines can never fail additional hardware versions
12063 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12065 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12066 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12069 ERR_R_NESTED_ASN1_ERROR
12071 ERR_R_MISSING_ASN1_EOS
12072 were 4 .. 9, conflicting with
12073 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12075 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12076 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12078 Add new error code 'ERR_R_INTERNAL_ERROR'.
12082 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12087 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12088 sets the subject name for a new request or supersedes the
12089 subject name in a given request. Formats that can be parsed are
12090 'CN=Some Name, OU=myOU, C=IT'
12092 'CN=Some Name/OU=myOU/C=IT'.
12094 Add options '-batch' and '-verbose' to 'openssl req'.
12096 *Massimiliano Pala <madwolf@hackmasters.net>*
12098 * Introduce the possibility to access global variables through
12099 functions on platform were that's the best way to handle exporting
12100 global variables in shared libraries. To enable this functionality,
12101 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12102 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12103 is normally done by Configure or something similar).
12105 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12106 in the source file (foo.c) like this:
12108 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12109 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12111 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12112 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12114 OPENSSL_DECLARE_GLOBAL(int,foo);
12115 #define foo OPENSSL_GLOBAL_REF(foo)
12116 OPENSSL_DECLARE_GLOBAL(double,bar);
12117 #define bar OPENSSL_GLOBAL_REF(bar)
12119 The #defines are very important, and therefore so is including the
12120 header file everywhere where the defined globals are used.
12122 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12123 of ASN.1 items, but that structure is a bit different.
12125 The largest change is in util/mkdef.pl which has been enhanced with
12126 better and easier to understand logic to choose which symbols should
12127 go into the Windows .def files as well as a number of fixes and code
12128 cleanup (among others, algorithm keywords are now sorted
12129 lexicographically to avoid constant rewrites).
12133 * In BN_div() keep a copy of the sign of 'num' before writing the
12134 result to 'rm' because if rm==num the value will be overwritten
12135 and produce the wrong result if 'num' is negative: this caused
12136 problems with BN_mod() and BN_nnmod().
12140 * Function OCSP_request_verify(). This checks the signature on an
12141 OCSP request and verifies the signer certificate. The signer
12142 certificate is just checked for a generic purpose and OCSP request
12147 * Add OCSP_check_validity() function to check the validity of OCSP
12148 responses. OCSP responses are prepared in real time and may only
12149 be a few seconds old. Simply checking that the current time lies
12150 between thisUpdate and nextUpdate max reject otherwise valid responses
12151 caused by either OCSP responder or client clock inaccuracy. Instead
12152 we allow thisUpdate and nextUpdate to fall within a certain period of
12153 the current time. The age of the response can also optionally be
12154 checked. Two new options -validity_period and -status_age added to
12159 * If signature or public key algorithm is unrecognized print out its
12160 OID rather that just UNKNOWN.
12164 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12165 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12166 ID to be generated from the issuer certificate alone which can then be
12167 passed to OCSP_id_issuer_cmp().
12171 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12172 ASN1 modules to export functions returning ASN1_ITEM pointers
12173 instead of the ASN1_ITEM structures themselves. This adds several
12174 new macros which allow the underlying ASN1 function/structure to
12175 be accessed transparently. As a result code should not use ASN1_ITEM
12176 references directly (such as &X509_it) but instead use the relevant
12177 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12178 use of the new ASN1 code on platforms where exporting structures
12179 is problematical (for example in shared libraries) but exporting
12180 functions returning pointers to structures is not.
12184 * Add support for overriding the generation of SSL/TLS session IDs.
12185 These callbacks can be registered either in an SSL_CTX or per SSL.
12186 The purpose of this is to allow applications to control, if they wish,
12187 the arbitrary values chosen for use as session IDs, particularly as it
12188 can be useful for session caching in multiple-server environments. A
12189 command-line switch for testing this (and any client code that wishes
12190 to use such a feature) has been added to "s_server".
12192 *Geoff Thorpe, Lutz Jaenicke*
12194 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12195 of the form '#if defined(...) || defined(...) || ...' and
12196 '#if !defined(...) && !defined(...) && ...'. This also avoids
12197 the growing number of special cases it was previously handling.
12201 * Make all configuration macros available for application by making
12202 sure they are available in opensslconf.h, by giving them names starting
12203 with "OPENSSL_" to avoid conflicts with other packages and by making
12204 sure e_os2.h will cover all platform-specific cases together with
12206 Additionally, it is now possible to define configuration/platform-
12207 specific names (called "system identities"). In the C code, these
12208 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
12209 macro with the name beginning with "OPENSSL_SYS_", which is determined
12210 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
12215 * New option -set_serial to 'req' and 'x509' this allows the serial
12216 number to use to be specified on the command line. Previously self
12217 signed certificates were hard coded with serial number 0 and the
12218 CA options of 'x509' had to use a serial number in a file which was
12223 * New options to 'ca' utility to support V2 CRL entry extensions.
12224 Currently CRL reason, invalidity date and hold instruction are
12225 supported. Add new CRL extensions to V3 code and some new objects.
12229 * New function EVP_CIPHER_CTX_set_padding() this is used to
12230 disable standard block padding (aka PKCS#5 padding) in the EVP
12231 API, which was previously mandatory. This means that the data is
12232 not padded in any way and so the total length much be a multiple
12233 of the block size, otherwise an error occurs.
12237 * Initial (incomplete) OCSP SSL support.
12241 * New function OCSP_parse_url(). This splits up a URL into its host,
12242 port and path components: primarily to parse OCSP URLs. New -url
12243 option to ocsp utility.
12247 * New nonce behavior. The return value of OCSP_check_nonce() now
12248 reflects the various checks performed. Applications can decide
12249 whether to tolerate certain situations such as an absent nonce
12250 in a response when one was present in a request: the ocsp application
12251 just prints out a warning. New function OCSP_add1_basic_nonce()
12252 this is to allow responders to include a nonce in a response even if
12253 the request is nonce-less.
12257 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12258 skipped when using openssl x509 multiple times on a single input file,
12259 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
12263 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12264 set string type: to handle setting ASN1_TIME structures. Fix ca
12265 utility to correctly initialize revocation date of CRLs.
12269 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12270 the clients preferred ciphersuites and rather use its own preferences.
12271 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12272 Internet Explorer by ensuring unchanged hash method during stepup.
12273 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12277 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12278 to aes and add a new 'exist' option to print out symbols that don't
12283 * Additional options to ocsp utility to allow flags to be set and
12284 additional certificates supplied.
12288 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12289 OCSP client a number of certificate to only verify the response
12294 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12295 handle the new API. Currently only ECB, CBC modes supported. Add new
12298 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12299 Encryption Standard (AES) Ciphersuites for Transport Layer
12300 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12301 not enabled by default and were not part of the "ALL" ciphersuite
12302 alias because they were not yet official; they could be
12303 explicitly requested by specifying the "AESdraft" ciphersuite
12304 group alias. In the final release of OpenSSL 0.9.7, the group
12305 alias is called "AES" and is part of "ALL".)
12307 *Ben Laurie, Steve Henson, Bodo Moeller*
12309 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12310 request to response.
12314 * Functions for OCSP responders. OCSP_request_onereq_count(),
12315 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12316 extract information from a certificate request. OCSP_response_create()
12317 creates a response and optionally adds a basic response structure.
12318 OCSP_basic_add1_status() adds a complete single response to a basic
12319 response and returns the OCSP_SINGLERESP structure just added (to allow
12320 extensions to be included for example). OCSP_basic_add1_cert() adds a
12321 certificate to a basic response and OCSP_basic_sign() signs a basic
12322 response with various flags. New helper functions ASN1_TIME_check()
12323 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12324 (converts ASN1_TIME to GeneralizedTime).
12328 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12329 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12330 structure from a certificate. X509_pubkey_digest() digests the public_key
12331 contents: this is used in various key identifiers.
12335 * Make sk_sort() tolerate a NULL argument.
12337 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12339 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12340 passed by the function are trusted implicitly. If any of them signed the
12341 response then it is assumed to be valid and is not verified.
12345 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12346 to data. This was previously part of the PKCS7 ASN1 code. This
12347 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12348 *Steve Henson, reported by Kenneth R. Robinette
12349 <support@securenetterm.com>*
12351 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12352 routines: without these tracing memory leaks is very painful.
12353 Fix leaks in PKCS12 and PKCS7 routines.
12357 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12358 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12359 effectively meant GeneralizedTime would never be used. Now it
12360 is initialised to -1 but X509_time_adj() now has to check the value
12361 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12362 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12363 *Steve Henson, reported by Kenneth R. Robinette
12364 <support@securenetterm.com>*
12366 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12367 result in a zero length in the ASN1_INTEGER structure which was
12368 not consistent with the structure when d2i_ASN1_INTEGER() was used
12369 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12370 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12371 where it did not print out a minus for negative ASN1_INTEGER.
12375 * Add summary printout to ocsp utility. The various functions which
12376 convert status values to strings have been renamed to:
12377 OCSP_response_status_str(), OCSP_cert_status_str() and
12378 OCSP_crl_reason_str() and are no longer static. New options
12379 to verify nonce values and to disable verification. OCSP response
12380 printout format cleaned up.
12384 * Add additional OCSP certificate checks. These are those specified
12385 in RFC2560. This consists of two separate checks: the CA of the
12386 certificate being checked must either be the OCSP signer certificate
12387 or the issuer of the OCSP signer certificate. In the latter case the
12388 OCSP signer certificate must contain the OCSP signing extended key
12389 usage. This check is performed by attempting to match the OCSP
12390 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12391 in the OCSP_CERTID structures of the response.
12395 * Initial OCSP certificate verification added to OCSP_basic_verify()
12396 and related routines. This uses the standard OpenSSL certificate
12397 verify routines to perform initial checks (just CA validity) and
12398 to obtain the certificate chain. Then additional checks will be
12399 performed on the chain. Currently the root CA is checked to see
12400 if it is explicitly trusted for OCSP signing. This is used to set
12401 a root CA as a global signing root: that is any certificate that
12402 chains to that CA is an acceptable OCSP signing certificate.
12406 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12407 extensions from a separate configuration file.
12408 As when reading extensions from the main configuration file,
12409 the '-extensions ...' option may be used for specifying the
12412 *Massimiliano Pala <madwolf@comune.modena.it>*
12414 * New OCSP utility. Allows OCSP requests to be generated or
12415 read. The request can be sent to a responder and the output
12416 parsed, outputted or printed in text form. Not complete yet:
12417 still needs to check the OCSP response validity.
12421 * New subcommands for 'openssl ca':
12422 'openssl ca -status <serial>' prints the status of the cert with
12423 the given serial number (according to the index file).
12424 'openssl ca -updatedb' updates the expiry status of certificates
12427 *Massimiliano Pala <madwolf@comune.modena.it>*
12429 * New '-newreq-nodes' command option to CA.pl. This is like
12430 '-newreq', but calls 'openssl req' with the '-nodes' option
12431 so that the resulting key is not encrypted.
12433 *Damien Miller <djm@mindrot.org>*
12435 * New configuration for the GNU Hurd.
12437 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12439 * Initial code to implement OCSP basic response verify. This
12440 is currently incomplete. Currently just finds the signer's
12441 certificate and verifies the signature on the response.
12445 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12446 value of OPENSSLDIR. This is available via the new '-d' option
12447 to 'openssl version', and is also included in 'openssl version -a'.
12451 * Allowing defining memory allocation callbacks that will be given
12452 file name and line number information in additional arguments
12453 (a const char* and an int). The basic functionality remains, as
12454 well as the original possibility to just replace malloc(),
12455 realloc() and free() by functions that do not know about these
12456 additional arguments. To register and find out the current
12457 settings for extended allocation functions, the following
12458 functions are provided:
12460 CRYPTO_set_mem_ex_functions
12461 CRYPTO_set_locked_mem_ex_functions
12462 CRYPTO_get_mem_ex_functions
12463 CRYPTO_get_locked_mem_ex_functions
12465 These work the same way as CRYPTO_set_mem_functions and friends.
12466 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
12467 extended allocation function is enabled.
12468 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
12469 a conventional allocation function is enabled.
12471 *Richard Levitte, Bodo Moeller*
12473 * Finish off removing the remaining LHASH function pointer casts.
12474 There should no longer be any prototype-casting required when using
12475 the LHASH abstraction, and any casts that remain are "bugs". See
12476 the callback types and macros at the head of lhash.h for details
12477 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12481 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12482 If /dev/[u]random devices are not available or do not return enough
12483 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12485 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12486 /etc/entropy will be queried once each in this sequence, querying stops
12487 when enough entropy was collected without querying more sockets.
12491 * Change the Unix RAND_poll() variant to be able to poll several
12492 random devices, as specified by DEVRANDOM, until a sufficient amount
12493 of data has been collected. We spend at most 10 ms on each file
12494 (select timeout) and read in non-blocking mode. DEVRANDOM now
12495 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12496 (previously it was just the string "/dev/urandom"), so on typical
12497 platforms the 10 ms delay will never occur.
12498 Also separate out the Unix variant to its own file, rand_unix.c.
12499 For VMS, there's a currently-empty rand_vms.c.
12503 * Move OCSP client related routines to ocsp_cl.c. These
12504 provide utility functions which an application needing
12505 to issue a request to an OCSP responder and analyse the
12506 response will typically need: as opposed to those which an
12507 OCSP responder itself would need which will be added later.
12509 OCSP_request_sign() signs an OCSP request with an API similar
12510 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12511 response. OCSP_response_get1_basic() extracts basic response
12512 from response. OCSP_resp_find_status(): finds and extracts status
12513 information from an OCSP_CERTID structure (which will be created
12514 when the request structure is built). These are built from lower
12515 level functions which work on OCSP_SINGLERESP structures but
12516 won't normally be used unless the application wishes to examine
12517 extensions in the OCSP response for example.
12519 Replace nonce routines with a pair of functions.
12520 OCSP_request_add1_nonce() adds a nonce value and optionally
12521 generates a random value. OCSP_check_nonce() checks the
12522 validity of the nonce in an OCSP response.
12526 * Change function OCSP_request_add() to OCSP_request_add0_id().
12527 This doesn't copy the supplied OCSP_CERTID and avoids the
12528 need to free up the newly created id. Change return type
12529 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12530 This can then be used to add extensions to the request.
12531 Deleted OCSP_request_new(), since most of its functionality
12532 is now in OCSP_REQUEST_new() (and the case insensitive name
12533 clash) apart from the ability to set the request name which
12534 will be added elsewhere.
12538 * Update OCSP API. Remove obsolete extensions argument from
12539 various functions. Extensions are now handled using the new
12540 OCSP extension code. New simple OCSP HTTP function which
12541 can be used to send requests and parse the response.
12545 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12546 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12547 uses the special reorder version of SET OF to sort the attributes
12548 and reorder them to match the encoded order. This resolves a long
12549 standing problem: a verify on a PKCS7 structure just after signing
12550 it used to fail because the attribute order did not match the
12551 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12552 it uses the received order. This is necessary to tolerate some broken
12553 software that does not order SET OF. This is handled by encoding
12554 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12555 to produce the required SET OF.
12559 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12560 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12561 files to get correct declarations of the ASN.1 item variables.
12565 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12566 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12567 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12568 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12569 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12570 ASN1_ITEM and no wrapper functions.
12574 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12575 replace the old function pointer based I/O routines. Change most of
12576 the *_d2i_bio() and *_d2i_fp() functions to use these.
12580 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12581 lines, recognize more "algorithms" that can be deselected, and make
12582 it complain about algorithm deselection that isn't recognised.
12586 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12587 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12588 to use new functions. Add NO_ASN1_OLD which can be set to remove
12589 some old style ASN1 functions: this can be used to determine if old
12590 code will still work when these eventually go away.
12594 * New extension functions for OCSP structures, these follow the
12595 same conventions as certificates and CRLs.
12599 * New function X509V3_add1_i2d(). This automatically encodes and
12600 adds an extension. Its behaviour can be customised with various
12601 flags to append, replace or delete. Various wrappers added for
12602 certificates and CRLs.
12606 * Fix to avoid calling the underlying ASN1 print routine when
12607 an extension cannot be parsed. Correct a typo in the
12608 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12612 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12613 entries for variables.
12617 * Add functionality to apps/openssl.c for detecting locking
12618 problems: As the program is single-threaded, all we have
12619 to do is register a locking callback using an array for
12620 storing which locks are currently held by the program.
12624 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12625 SSL_get_ex_data_X509_STORE_idx(), which is used in
12626 ssl_verify_cert_chain() and thus can be called at any time
12627 during TLS/SSL handshakes so that thread-safety is essential.
12628 Unfortunately, the ex_data design is not at all suited
12629 for multi-threaded use, so it probably should be abolished.
12633 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12635 *Broadcom, tweaked and integrated by Geoff Thorpe*
12637 * Move common extension printing code to new function
12638 X509V3_print_extensions(). Reorganise OCSP print routines and
12639 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12643 * New function X509_signature_print() to remove duplication in some
12648 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12649 set (this was treated exactly the same as SET OF previously). This
12650 is used to reorder the STACK representing the structure to match the
12651 encoding. This will be used to get round a problem where a PKCS7
12652 structure which was signed could not be verified because the STACK
12653 order did not reflect the encoded order.
12657 * Reimplement the OCSP ASN1 module using the new code.
12661 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12662 for its ASN1 operations. The old style function pointers still exist
12663 for now but they will eventually go away.
12667 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12668 completely replaces the old ASN1 functionality with a table driven
12669 encoder and decoder which interprets an ASN1_ITEM structure describing
12670 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12671 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12672 has also been converted to the new form.
12676 * Change BN_mod_exp_recp so that negative moduli are tolerated
12677 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12678 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12679 for negative moduli.
12683 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12684 of not touching the result's sign bit.
12688 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12693 * Changed the LHASH code to use prototypes for callbacks, and created
12694 macros to declare and implement thin (optionally static) functions
12695 that provide type-safety and avoid function pointer casting for the
12696 type-specific callbacks.
12700 * Added Kerberos Cipher Suites to be used with TLS, as written in
12702 *Veers Staats <staatsvr@asc.hpc.mil>,
12703 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12705 * Reformat the FAQ so the different questions and answers can be divided
12706 in sections depending on the subject.
12710 * Have the zlib compression code load ZLIB.DLL dynamically under
12715 * New function BN_mod_sqrt for computing square roots modulo a prime
12716 (using the probabilistic Tonelli-Shanks algorithm unless
12717 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12718 be handled deterministically).
12720 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12722 * Make BN_mod_inverse faster by explicitly handling small quotients
12723 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12724 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12728 * New function BN_kronecker.
12732 * Fix BN_gcd so that it works on negative inputs; the result is
12733 positive unless both parameters are zero.
12734 Previously something reasonably close to an infinite loop was
12735 possible because numbers could be growing instead of shrinking
12736 in the implementation of Euclid's algorithm.
12740 * Fix BN_is_word() and BN_is_one() macros to take into account the
12741 sign of the number in question.
12743 Fix BN_is_word(a,w) to work correctly for w == 0.
12745 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12746 because its test if the absolute value of 'a' equals 'w'.
12747 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12748 it exists mostly for use in the implementations of BN_is_zero(),
12749 BN_is_one(), and BN_is_word().
12753 * New function BN_swap.
12757 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12758 the exponentiation functions are more likely to produce reasonable
12759 results on negative inputs.
12763 * Change BN_mod_mul so that the result is always non-negative.
12764 Previously, it could be negative if one of the factors was negative;
12765 I don't think anyone really wanted that behaviour.
12769 * Move BN_mod_... functions into new file crypto/bn/bn_mod.c
12770 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12771 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12772 and add new functions:
12781 BN_mod_lshift1_quick
12783 BN_mod_lshift_quick
12785 These functions always generate non-negative results.
12787 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12788 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12790 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12791 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12792 be reduced modulo m.
12794 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12797 The following entry accidentally appeared in the CHANGES file
12798 distributed with OpenSSL 0.9.7. The modifications described in
12799 it do *not* apply to OpenSSL 0.9.7.
12801 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12802 was actually never needed) and in BN_mul(). The removal in BN_mul()
12803 required a small change in bn_mul_part_recursive() and the addition
12804 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12805 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12806 bn_sub_words() and bn_add_words() except they take arrays with
12812 * In 'openssl passwd', verify passwords read from the terminal
12813 unless the '-salt' option is used (which usually means that
12814 verification would just waste user's time since the resulting
12815 hash is going to be compared with some given password hash)
12816 or the new '-noverify' option is used.
12818 This is an incompatible change, but it does not affect
12819 non-interactive use of 'openssl passwd' (passwords on the command
12820 line, '-stdin' option, '-in ...' option) and thus should not
12821 cause any problems.
12825 * Remove all references to RSAref, since there's no more need for it.
12829 * Make DSO load along a path given through an environment variable
12830 (SHLIB_PATH) with shl_load().
12834 * Constify the ENGINE code as a result of BIGNUM constification.
12835 Also constify the RSA code and most things related to it. In a
12836 few places, most notable in the depth of the ASN.1 code, ugly
12837 casts back to non-const were required (to be solved at a later
12842 * Make it so the openssl application has all engines loaded by default.
12846 * Constify the BIGNUM routines a little more.
12850 * Add the following functions:
12852 ENGINE_load_cswift()
12854 ENGINE_load_atalla()
12855 ENGINE_load_nuron()
12856 ENGINE_load_builtin_engines()
12858 That way, an application can itself choose if external engines that
12859 are built-in in OpenSSL shall ever be used or not. The benefit is
12860 that applications won't have to be linked with libdl or other dso
12861 libraries unless it's really needed.
12863 Changed 'openssl engine' to load all engines on demand.
12864 Changed the engine header files to avoid the duplication of some
12865 declarations (they differed!).
12869 * 'openssl engine' can now list capabilities.
12873 * Better error reporting in 'openssl engine'.
12877 * Never call load_dh_param(NULL) in s_server.
12881 * Add engine application. It can currently list engines by name and
12882 identity, and test if they are actually available.
12886 * Improve RPM specification file by forcing symbolic linking and making
12887 sure the installed documentation is also owned by root.root.
12889 *Damien Miller <djm@mindrot.org>*
12891 * Give the OpenSSL applications more possibilities to make use of
12892 keys (public as well as private) handled by engines.
12896 * Add OCSP code that comes from CertCo.
12900 * Add VMS support for the Rijndael code.
12904 * Added untested support for Nuron crypto accelerator.
12908 * Add support for external cryptographic devices. This code was
12909 previously distributed separately as the "engine" branch.
12911 *Geoff Thorpe, Richard Levitte*
12913 * Rework the filename-translation in the DSO code. It is now possible to
12914 have far greater control over how a "name" is turned into a filename
12915 depending on the operating environment and any oddities about the
12916 different shared library filenames on each system.
12920 * Support threads on FreeBSD-elf in Configure.
12924 * Fix for SHA1 assembly problem with MASM: it produces
12925 warnings about corrupt line number information when assembling
12926 with debugging information. This is caused by the overlapping
12929 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
12932 NCONF_get_number() has no error checking at all. As a replacement,
12933 NCONF_get_number_e() is defined (_e for "error checking") and is
12934 promoted strongly. The old NCONF_get_number is kept around for
12935 binary backward compatibility.
12936 Make it possible for methods to load from something other than a BIO,
12937 by providing a function pointer that is given a name instead of a BIO.
12938 For example, this could be used to load configuration data from an
12943 * Fix for non blocking accept BIOs. Added new I/O special reason
12944 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
12945 with non blocking I/O was not possible because no retry code was
12946 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
12951 * Added the beginnings of Rijndael support.
12955 * Fix for bug in DirectoryString mask setting. Add support for
12956 X509_NAME_print_ex() in 'req' and X509_print_ex() function
12957 to allow certificate printing to more controllable, additional
12958 'certopt' option to 'x509' to allow new printing options to be
12963 * Clean old EAY MD5 hack from e_os.h.
12967 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004] ###
12969 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12970 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
12972 *Joe Orton, Steve Henson*
12974 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003] ###
12976 * Fix additional bug revealed by the NISCC test suite:
12978 Stop bug triggering large recursion when presented with
12979 certain ASN.1 tags [CVE-2003-0851][]
12983 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003] ###
12985 * Fix various bugs revealed by running the NISCC test suite:
12987 Stop out of bounds reads in the ASN1 code when presented with
12988 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12990 If verify callback ignores invalid public key errors don't try to check
12991 certificate signature with the NULL public key.
12996 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12997 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13002 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13003 extra data after the compression methods not only for TLS 1.0
13004 but also for SSL 3.0 (as required by the specification).
13006 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13008 * Change X509_certificate_type() to mark the key as exported/exportable
13009 when it's 512 *bits* long, not 512 bytes.
13013 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003] ###
13015 * Countermeasure against the Klima-Pokorny-Rosa extension of
13016 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13017 a protocol version number mismatch like a decryption error
13018 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13022 * Turn on RSA blinding by default in the default implementation
13023 to avoid a timing attack. Applications that don't want it can call
13024 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13025 They would be ill-advised to do so in most cases.
13027 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13029 * Change RSA blinding code so that it works when the PRNG is not
13030 seeded (in this case, the secret RSA exponent is abused as
13031 an unpredictable seed -- if it is not unpredictable, there
13032 is no point in blinding anyway). Make RSA blinding thread-safe
13033 by remembering the creator's thread ID in rsa->blinding and
13034 having all other threads use local one-time blinding factors
13035 (this requires more computation than sharing rsa->blinding, but
13036 avoids excessive locking; and if an RSA object is not shared
13037 between threads, blinding will still be very fast).
13041 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003] ###
13043 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13044 via timing by performing a MAC computation even if incorrect
13045 block cipher padding has been found. This is a countermeasure
13046 against active attacks where the attacker has to distinguish
13047 between bad padding and a MAC verification error. [CVE-2003-0078][]
13049 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13050 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13051 Martin Vuagnoux (EPFL, Ilion)*
13053 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002] ###
13055 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13056 memory from its contents. This is done with a counter that will
13057 place alternating values in each byte. This can be used to solve
13058 two issues: 1) the removal of calls to memset() by highly optimizing
13059 compilers, and 2) cleansing with other values than 0, since those can
13060 be read through on certain media, for example a swap space on disk.
13064 * Bugfix: client side session caching did not work with external caching,
13065 because the session->cipher setting was not restored when reloading
13066 from the external cache. This problem was masked, when
13067 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13068 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13072 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13073 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13075 *Zeev Lieber <zeev-l@yahoo.com>*
13077 * Undo an undocumented change introduced in 0.9.6e which caused
13078 repeated calls to OpenSSL_add_all_ciphers() and
13079 OpenSSL_add_all_digests() to be ignored, even after calling
13084 * Change the default configuration reader to deal with last line not
13085 being properly terminated.
13089 * Change X509_NAME_cmp() so it applies the special rules on handling
13090 DN values that are of type PrintableString, as well as RDNs of type
13091 emailAddress where the value has the type ia5String.
13093 *stefank@valicert.com via Richard Levitte*
13095 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13096 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13097 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13098 the bitwise-OR of the two for use by the majority of applications
13099 wanting this behaviour, and update the docs. The documented
13100 behaviour and actual behaviour were inconsistent and had been
13101 changing anyway, so this is more a bug-fix than a behavioural
13104 *Geoff Thorpe, diagnosed by Nadav Har'El*
13106 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13107 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13111 * Fix initialization code race conditions in
13112 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13113 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13114 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13115 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13116 ssl2_get_cipher_by_char(),
13117 ssl3_get_cipher_by_char().
13119 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13121 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13122 the cached sessions are flushed, as the remove_cb() might use ex_data
13123 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13124 (see [openssl.org #212]).
13126 *Geoff Thorpe, Lutz Jaenicke*
13128 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13129 length, instead of the encoding length to d2i_ASN1_OBJECT.
13133 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002] ###
13135 * [In 0.9.6g-engine release:]
13136 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
13138 *Lynn Gazis <lgazis@rainbow.com>*
13140 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002] ###
13142 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13143 and get fix the header length calculation.
13144 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13145 Alon Kantor <alonk@checkpoint.com> (and others),
13148 * Use proper error handling instead of 'assertions' in buffer
13149 overflow checks added in 0.9.6e. This prevents DoS (the
13150 assertions could call abort()).
13152 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13154 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002] ###
13156 * Add various sanity checks to asn1_get_length() to reject
13157 the ASN1 length bytes if they exceed sizeof(long), will appear
13158 negative or the content length exceeds the length of the
13161 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13163 * Fix cipher selection routines: ciphers without encryption had no flags
13164 for the cipher strength set and where therefore not handled correctly
13165 by the selection routines (PR #130).
13169 * Fix EVP_dsa_sha macro.
13174 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13175 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13176 that was added in OpenSSL 0.9.6d.
13178 As the countermeasure turned out to be incompatible with some
13179 broken SSL implementations, the new option is part of SSL_OP_ALL.
13180 SSL_OP_ALL is usually employed when compatibility with weird SSL
13181 implementations is desired (e.g. '-bugs' option to 's_client' and
13182 's_server'), so the new option is automatically set in many
13187 * Changes in security patch:
13189 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13190 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13191 Air Force Materiel Command, USAF, under agreement number
13194 * Add various sanity checks to asn1_get_length() to reject
13195 the ASN1 length bytes if they exceed sizeof(long), will appear
13196 negative or the content length exceeds the length of the
13197 supplied buffer. [CVE-2002-0659][]
13199 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13201 * Assertions for various potential buffer overflows, not known to
13202 happen in practice.
13204 *Ben Laurie (CHATS)*
13206 * Various temporary buffers to hold ASCII versions of integers were
13207 too small for 64 bit platforms. [CVE-2002-0655][]
13208 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
13210 * Remote buffer overflow in SSL3 protocol - an attacker could
13211 supply an oversized session ID to a client. [CVE-2002-0656][]
13213 *Ben Laurie (CHATS)*
13215 * Remote buffer overflow in SSL2 protocol - an attacker could
13216 supply an oversized client master key. [CVE-2002-0656][]
13218 *Ben Laurie (CHATS)*
13220 ### Changes between 0.9.6c and 0.9.6d [9 May 2002] ###
13222 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13223 encoded as NULL) with id-dsa-with-sha1.
13225 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13227 * Check various X509_...() return values in apps/req.c.
13229 *Nils Larsch <nla@trustcenter.de>*
13231 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13232 an end-of-file condition would erroneously be flagged, when the CRLF
13233 was just at the end of a processed block. The bug was discovered when
13234 processing data through a buffering memory BIO handing the data to a
13235 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13236 <ptsekov@syntrex.com> and Nedelcho Stanev.
13240 * Implement a countermeasure against a vulnerability recently found
13241 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13242 before application data chunks to avoid the use of known IVs
13243 with data potentially chosen by the attacker.
13247 * Fix length checks in ssl3_get_client_hello().
13251 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13252 to prevent ssl3_read_internal() from incorrectly assuming that
13253 ssl3_read_bytes() found application data while handshake
13254 processing was enabled when in fact s->s3->in_read_app_data was
13255 merely automatically cleared during the initial handshake.
13257 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13259 * Fix object definitions for Private and Enterprise: they were not
13260 recognized in their shortname (=lowercase) representation. Extend
13261 obj_dat.pl to issue an error when using undefined keywords instead
13262 of silently ignoring the problem (Svenning Sorensen
13263 <sss@sss.dnsalias.net>).
13267 * Fix DH_generate_parameters() so that it works for 'non-standard'
13268 generators, i.e. generators other than 2 and 5. (Previously, the
13269 code did not properly initialise the 'add' and 'rem' values to
13270 BN_generate_prime().)
13272 In the new general case, we do not insist that 'generator' is
13273 actually a primitive root: This requirement is rather pointless;
13274 a generator of the order-q subgroup is just as good, if not
13279 * Map new X509 verification errors to alerts. Discovered and submitted by
13280 Tom Wu <tom@arcot.com>.
13284 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13285 returning non-zero before the data has been completely received
13286 when using non-blocking I/O.
13288 *Bodo Moeller; problem pointed out by John Hughes*
13290 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13292 *Ben Laurie, Lutz Jaenicke*
13294 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13295 Yoram Zahavi <YoramZ@gilian.com>).
13299 * Add information about CygWin 1.3 and on, and preserve proper
13300 configuration for the versions before that.
13302 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13304 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13305 check whether we deal with a copy of a session and do not delete from
13306 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13307 <izhar@checkpoint.com>.
13311 * Do not store session data into the internal session cache, if it
13312 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13313 flag is set). Proposed by Aslam <aslam@funk.com>.
13317 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13322 * [In 0.9.6d-engine release:]
13323 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13325 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13327 * Add the configuration target linux-s390x.
13329 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13331 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13332 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13333 variable as an indication that a ClientHello message has been
13334 received. As the flag value will be lost between multiple
13335 invocations of ssl3_accept when using non-blocking I/O, the
13336 function may not be aware that a handshake has actually taken
13337 place, thus preventing a new session from being added to the
13340 To avoid this problem, we now set s->new_session to 2 instead of
13341 using a local variable.
13343 *Lutz Jaenicke, Bodo Moeller*
13345 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13346 if the SSL_R_LENGTH_MISMATCH error is detected.
13348 *Geoff Thorpe, Bodo Moeller*
13350 * New 'shared_ldflag' column in Configure platform table.
13354 * Fix EVP_CIPHER_mode macro.
13356 *"Dan S. Camper" <dan@bti.net>*
13358 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13359 type, we must throw them away by setting rr->length to 0.
13361 *D P Chang <dpc@qualys.com>*
13363 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001] ###
13365 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13366 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13367 worked incorrectly for those cases where range = 10..._2 and
13368 3*range is two bits longer than range.)
13372 * Only add signing time to PKCS7 structures if it is not already
13377 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13378 OBJ_ld_ce should be OBJ_id_ce.
13379 Also some ip-pda OIDs in crypto/objects/objects.txt were
13380 incorrect (cf. RFC 3039).
13382 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13384 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13385 returns early because it has nothing to do.
13387 *Andy Schneider <andy.schneider@bjss.co.uk>*
13389 * [In 0.9.6c-engine release:]
13390 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13392 *Andy Schneider <andy.schneider@bjss.co.uk>*
13394 * [In 0.9.6c-engine release:]
13395 Add support for Cryptographic Appliance's keyserver technology.
13396 (Use engine 'keyclient')
13398 *Cryptographic Appliances and Geoff Thorpe*
13400 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13401 is called via tools/c89.sh because arguments have to be
13402 rearranged (all '-L' options must appear before the first object
13405 *Richard Shapiro <rshapiro@abinitio.com>*
13407 * [In 0.9.6c-engine release:]
13408 Add support for Broadcom crypto accelerator cards, backported
13411 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13413 * [In 0.9.6c-engine release:]
13414 Add support for SureWare crypto accelerator cards from
13415 Baltimore Technologies. (Use engine 'sureware')
13417 *Baltimore Technologies and Mark Cox*
13419 * [In 0.9.6c-engine release:]
13420 Add support for crypto accelerator cards from Accelerated
13421 Encryption Processing, www.aep.ie. (Use engine 'aep')
13423 *AEP Inc. and Mark Cox*
13425 * Add a configuration entry for gcc on UnixWare.
13427 *Gary Benson <gbenson@redhat.com>*
13429 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13430 messages are stored in a single piece (fixed-length part and
13431 variable-length part combined) and fix various bugs found on the way.
13435 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13436 instead. BIO_gethostbyname() does not know what timeouts are
13437 appropriate, so entries would stay in cache even when they have
13439 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
13441 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13442 faced with a pathologically small ClientHello fragment that does
13443 not contain client_version: Instead of aborting with an error,
13444 simply choose the highest available protocol version (i.e.,
13445 TLS 1.0 unless it is disabled). In practice, ClientHello
13446 messages are never sent like this, but this change gives us
13447 strictly correct behaviour at least for TLS.
13451 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13452 never resets s->method to s->ctx->method when called from within
13453 one of the SSL handshake functions.
13455 *Bodo Moeller; problem pointed out by Niko Baric*
13457 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13458 (sent using the client's version number) if client_version is
13459 smaller than the protocol version in use. Also change
13460 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13461 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13462 the client will at least see that alert.
13466 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13471 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13472 client receives HelloRequest while in a handshake.
13474 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13476 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13477 should end in 'break', not 'goto end' which circumvents various
13478 cleanups done in state SSL_ST_OK. But session related stuff
13479 must be disabled for SSL_ST_OK in the case that we just sent a
13482 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13483 before just sending a HelloRequest.
13485 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13487 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13488 reveal whether illegal block cipher padding was found or a MAC
13489 verification error occurred. (Neither SSLerr() codes nor alerts
13490 are directly visible to potential attackers, but the information
13491 may leak via logfiles.)
13493 Similar changes are not required for the SSL 2.0 implementation
13494 because the number of padding bytes is sent in clear for SSL 2.0,
13495 and the extra bytes are just ignored. However ssl/s2_pkt.c
13496 failed to verify that the purported number of padding bytes is in
13501 * Add OpenUNIX-8 support including shared libraries
13502 (Boyd Lynn Gerber <gerberb@zenez.com>).
13506 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13507 'wristwatch attack' using huge encoding parameters (cf.
13508 James H. Manger's CRYPTO 2001 paper). Note that the
13509 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13510 encoding parameters and hence was not vulnerable.
13514 * BN_sqr() bug fix.
13516 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13518 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13519 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13520 followed by modular reduction.
13522 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13524 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13525 equivalent based on BN_pseudo_rand() instead of BN_rand().
13529 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13530 This function was broken, as the check for a new client hello message
13531 to handle SGC did not allow these large messages.
13532 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13536 * Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
13540 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13541 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13545 * Rework the configuration and shared library support for Tru64 Unix.
13546 The configuration part makes use of modern compiler features and
13547 still retains old compiler behavior for those that run older versions
13548 of the OS. The shared library support part includes a variant that
13549 uses the RPATH feature, and is available through the special
13550 configuration target "alpha-cc-rpath", which will never be selected
13553 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13555 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13556 with the same message size as in ssl3_get_certificate_request().
13557 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13558 messages might inadvertently be reject as too long.
13560 *Petr Lampa <lampa@fee.vutbr.cz>*
13562 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13566 * Modified SSL library such that the verify_callback that has been set
13567 specifically for an SSL object with SSL_set_verify() is actually being
13568 used. Before the change, a verify_callback set with this function was
13569 ignored and the verify_callback() set in the SSL_CTX at the time of
13570 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13571 to allow the necessary settings.
13575 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13576 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13577 done automatically (in contradiction to the requirements of the C
13578 standard). This made problems when used from OpenSSH.
13582 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13583 dh->length and always used
13585 BN_rand_range(priv_key, dh->p).
13587 BN_rand_range() is not necessary for Diffie-Hellman, and this
13588 specific range makes Diffie-Hellman unnecessarily inefficient if
13589 dh->length (recommended exponent length) is much smaller than the
13590 length of dh->p. We could use BN_rand_range() if the order of
13591 the subgroup was stored in the DH structure, but we only have
13596 BN_rand(priv_key, l, ...)
13598 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13605 RSA_eay_public_encrypt
13606 RSA_eay_private_decrypt
13607 RSA_eay_private_encrypt (signing)
13608 RSA_eay_public_decrypt (signature verification)
13610 (default implementations for RSA_public_encrypt,
13611 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13612 always reject numbers >= n.
13616 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13617 to synchronize access to 'locking_thread'. This is necessary on
13618 systems where access to 'locking_thread' (an 'unsigned long'
13619 variable) is not atomic.
13623 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13624 *before* setting the 'crypto_lock_rand' flag. The previous code had
13625 a race condition if 0 is a valid thread ID.
13627 *Travis Vitek <vitek@roguewave.com>*
13629 * Add support for shared libraries under Irix.
13631 *Albert Chin-A-Young <china@thewrittenword.com>*
13633 * Add configuration option to build on Linux on both big-endian and
13634 little-endian MIPS.
13636 *Ralf Baechle <ralf@uni-koblenz.de>*
13638 * Add the possibility to create shared libraries on HP-UX.
13642 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001] ###
13644 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13645 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13646 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13647 PRNG state recovery was possible based on the output of
13648 one PRNG request appropriately sized to gain knowledge on
13649 'md' followed by enough consecutive 1-byte PRNG requests
13650 to traverse all of 'state'.
13652 1. When updating 'md_local' (the current thread's copy of 'md')
13653 during PRNG output generation, hash all of the previous
13654 'md_local' value, not just the half used for PRNG output.
13656 2. Make the number of bytes from 'state' included into the hash
13657 independent from the number of PRNG bytes requested.
13659 The first measure alone would be sufficient to avoid
13660 Markku-Juhani's attack. (Actually it had never occurred
13661 to me that the half of 'md_local' used for chaining was the
13662 half from which PRNG output bytes were taken -- I had always
13663 assumed that the secret half would be used.) The second
13664 measure makes sure that additional data from 'state' is never
13665 mixed into 'md_local' in small portions; this heuristically
13666 further strengthens the PRNG.
13670 * Fix crypto/bn/asm/mips3.s.
13674 * When only the key is given to "enc", the IV is undefined. Print out
13675 an error message in this case.
13679 * Handle special case when X509_NAME is empty in X509 printing routines.
13683 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13684 positive and less than q.
13688 * Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
13689 used: it isn't thread safe and the add_lock_callback should handle
13692 *Paul Rose <Paul.Rose@bridge.com>*
13694 * Verify that incoming data obeys the block size in
13695 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13701 *Ulf Möller, Bodo Möller*
13703 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13704 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13705 when fixing the server behaviour for backwards-compatible 'client
13706 hello' messages. (Note that the attack is impractical against
13707 SSL 3.0 and TLS 1.0 anyway because length and version checking
13708 means that the probability of guessing a valid ciphertext is
13709 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13712 Before 0.9.5, the countermeasure (hide the error by generating a
13713 random 'decryption result') did not work properly because
13714 ERR_clear_error() was missing, meaning that SSL_get_error() would
13715 detect the supposedly ignored error.
13717 Both problems are now fixed.
13721 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13722 (previously it was 1024).
13726 * Fix for compatibility mode trust settings: ignore trust settings
13727 unless some valid trust or reject settings are present.
13731 * Fix for blowfish EVP: its a variable length cipher.
13735 * Fix various bugs related to DSA S/MIME verification. Handle missing
13736 parameters in DSA public key structures and return an error in the
13737 DSA routines if parameters are absent.
13741 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13742 in the current directory if neither $RANDFILE nor $HOME was set.
13743 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13744 caused some confusion to Windows users who haven't defined $HOME.
13745 Thus RAND_file_name() is changed again: e_os.h can define a
13746 DEFAULT_HOME, which will be used if $HOME is not set.
13747 For Windows, we use "C:"; on other platforms, we still require
13748 environment variables.
13750 * Move 'if (!initialized) RAND_poll()' into regions protected by
13751 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13752 having multiple threads call RAND_poll() concurrently.
13756 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13757 combination of a flag and a thread ID variable.
13758 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13759 flag), *other* threads can enter ssleay_add_bytes without obeying
13760 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13761 that they do not hold after the first thread unsets add_do_not_lock).
13765 * Change bctest again: '-x' expressions are not available in all
13766 versions of 'test'.
13770 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001] ###
13772 * Fix a couple of memory leaks in PKCS7_dataDecode()
13774 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13776 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13777 the default extension for executables, if any. Also, make the perl
13778 scripts that use symlink() to test if it really exists and use "cp"
13779 if it doesn't. All this made OpenSSL compilable and installable in
13784 * Fix for asn1_GetSequence() for indefinite length constructed data.
13785 If SEQUENCE is length is indefinite just set c->slen to the total
13786 amount of data available.
13788 *Steve Henson, reported by shige@FreeBSD.org*
13790 *This change does not apply to 0.9.7.*
13792 * Change bctest to avoid here-documents inside command substitution
13793 (workaround for FreeBSD /bin/sh bug).
13794 For compatibility with Ultrix, avoid shell functions (introduced
13795 in the bctest version that searches along $PATH).
13799 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13800 with des_encrypt() defined on some operating systems, like Solaris
13805 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13806 On the Importance of Eliminating Errors in Cryptographic
13807 Computations, J. Cryptology 14 (2001) 2, 101-119,
13808 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
13812 * MIPS assembler BIGNUM division bug fix.
13816 * Disabled incorrect Alpha assembler code.
13820 * Fix PKCS#7 decode routines so they correctly update the length
13821 after reading an EOC for the EXPLICIT tag.
13825 *This change does not apply to 0.9.7.*
13827 * Fix bug in PKCS#12 key generation routines. This was triggered
13828 if a 3DES key was generated with a 0 initial byte. Include
13829 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13830 (but broken) behaviour.
13834 * Enhance bctest to search for a working bc along $PATH and print
13837 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13839 * Fix memory leaks in err.c: free err_data string if necessary;
13840 don't write to the wrong index in ERR_set_error_data.
13844 * Implement ssl23_peek (analogous to ssl23_read), which previously
13849 * Replace rdtsc with _emit statements for VC++ version 5.
13851 *Jeremy Cooper <jeremy@baymoo.org>*
13853 * Make it possible to reuse SSLv2 sessions.
13857 * In copy_email() check for >= 0 as a return value for
13858 X509_NAME_get_index_by_NID() since 0 is a valid index.
13860 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13862 * Avoid coredump with unsupported or invalid public keys by checking if
13863 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13864 PKCS7_verify() fails with non detached data.
13868 * Don't use getenv in library functions when run as setuid/setgid.
13869 New function OPENSSL_issetugid().
13873 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13874 due to incorrect handling of multi-threading:
13876 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13878 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13880 3. Count how many times MemCheck_off() has been called so that
13881 nested use can be treated correctly. This also avoids
13882 inband-signalling in the previous code (which relied on the
13883 assumption that thread ID 0 is impossible).
13887 * Add "-rand" option also to s_client and s_server.
13891 * Fix CPU detection on Irix 6.x.
13892 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13893 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13895 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13900 *This change does not apply to 0.9.7.*
13902 * Use the cached encoding of an X509_NAME structure rather than
13903 copying it. This is apparently the reason for the libsafe "errors"
13904 but the code is actually correct.
13908 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13909 Bleichenbacher's DSA attack.
13910 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13911 to be set and top=0 forces the highest bit to be set; top=-1 is new
13912 and leaves the highest bit random.
13914 *Ulf Moeller, Bodo Moeller*
13916 * In the NCONF_...-based implementations for CONF_... queries
13917 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13918 a temporary CONF structure with the data component set to NULL
13919 (which gives segmentation faults in lh_retrieve).
13920 Instead, use NULL for the CONF pointer in CONF_get_string and
13921 CONF_get_number (which may use environment variables) and directly
13922 return NULL from CONF_get_section.
13926 * Fix potential buffer overrun for EBCDIC.
13930 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
13931 keyUsage if basicConstraints absent for a CA.
13935 * Make SMIME_write_PKCS7() write mail header values with a format that
13936 is more generally accepted (no spaces before the semicolon), since
13937 some programs can't parse those values properly otherwise. Also make
13938 sure BIO's that break lines after each write do not create invalid
13943 * Make the CRL encoding routines work with empty SEQUENCE OF. The
13944 macros previously used would not encode an empty SEQUENCE OF
13945 and break the signature.
13949 *This change does not apply to 0.9.7.*
13951 * Zero the premaster secret after deriving the master secret in
13956 * Add some EVP_add_digest_alias registrations (as found in
13957 OpenSSL_add_all_digests()) to SSL_library_init()
13958 aka OpenSSL_add_ssl_algorithms(). This provides improved
13959 compatibility with peers using X.509 certificates
13960 with unconventional AlgorithmIdentifier OIDs.
13964 * Fix for Irix with NO_ASM.
13966 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13968 * ./config script fixes.
13970 *Ulf Moeller, Richard Levitte*
13972 * Fix 'openssl passwd -1'.
13976 * Change PKCS12_key_gen_asc() so it can cope with non null
13977 terminated strings whose length is passed in the passlen
13978 parameter, for example from PEM callbacks. This was done
13979 by adding an extra length parameter to asc2uni().
13981 *Steve Henson, reported by <oddissey@samsung.co.kr>*
13983 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
13984 call failed, free the DSA structure.
13988 * Fix to uni2asc() to cope with zero length Unicode strings.
13989 These are present in some PKCS#12 files.
13993 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
13994 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
13995 when writing a 32767 byte record.
13997 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
13999 * In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
14000 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
14002 (RSA objects have a reference count access to which is protected
14003 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14004 so they are meant to be shared between threads.)
14005 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14006 "Reddie, Steven" <Steven.Reddie@ca.com>*
14008 * Fix a deadlock in CRYPTO_mem_leaks().
14012 * Use better test patterns in bntest.
14016 * rand_win.c fix for Borland C.
14020 * BN_rshift bugfix for n == 0.
14024 * Add a 'bctest' script that checks for some known 'bc' bugs
14025 so that 'make test' does not abort just because 'bc' is broken.
14029 * Store verify_result within SSL_SESSION also for client side to
14030 avoid potential security hole. (Re-used sessions on the client side
14031 always resulted in verify_result==X509_V_OK, not using the original
14032 result of the server certificate verification.)
14036 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14037 SSL3_RT_APPLICATION_DATA, return 0.
14038 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14043 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14044 releases, have been re-implemented by renaming the previous
14045 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14046 and ssl3_read_internal, respectively, and adding 'peek' parameters
14047 to them. The new ssl[23]_{read,peek} functions are calls to
14048 ssl[23]_read_internal with the 'peek' flag set appropriately.
14049 A 'peek' parameter has also been added to ssl3_read_bytes, which
14050 does the actual work for ssl3_read_internal.
14054 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14055 the method-specific "init()" handler. Also clean up ex_data after
14056 calling the method-specific "finish()" handler. Previously, this was
14057 happening the other way round.
14061 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14062 The previous value, 12, was not always sufficient for BN_mod_exp().
14066 * Make sure that shared libraries get the internal name engine with
14067 the full version number and not just 0. This should mark the
14068 shared libraries as not backward compatible. Of course, this should
14069 be changed again when we can guarantee backward binary compatibility.
14073 * Fix typo in get_cert_by_subject() in by_dir.c
14075 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14077 * Rework the system to generate shared libraries:
14079 - Make note of the expected extension for the shared libraries and
14080 if there is a need for symbolic links from for example libcrypto.so.0
14081 to libcrypto.so.0.9.7. There is extended info in Configure for
14084 - Make as few rebuilds of the shared libraries as possible.
14086 - Still avoid linking the OpenSSL programs with the shared libraries.
14088 - When installing, install the shared libraries separately from the
14093 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14095 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14096 and not in SSL_clear because the latter is also used by the
14097 accept/connect functions; previously, the settings made by
14098 SSL_set_read_ahead would be lost during the handshake.
14100 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14102 * Correct util/mkdef.pl to be selective about disabled algorithms.
14103 Previously, it would create entries for disabled algorithms no
14108 * Added several new manual pages for SSL_* function.
14112 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000] ###
14114 * In ssl23_get_client_hello, generate an error message when faced
14115 with an initial SSL 3.0/TLS record that is too small to contain the
14116 first two bytes of the ClientHello message, i.e. client_version.
14117 (Note that this is a pathologic case that probably has never happened
14118 in real life.) The previous approach was to use the version number
14119 from the record header as a substitute; but our protocol choice
14120 should not depend on that one because it is not authenticated
14121 by the Finished messages.
14125 * More robust randomness gathering functions for Windows.
14127 *Jeffrey Altman <jaltman@columbia.edu>*
14129 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14130 not set then we don't setup the error code for issuer check errors
14131 to avoid possibly overwriting other errors which the callback does
14132 handle. If an application does set the flag then we assume it knows
14133 what it is doing and can handle the new informational codes
14138 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14139 a general "ANY" type, as such it should be able to decode anything
14140 including tagged types. However it didn't check the class so it would
14141 wrongly interpret tagged types in the same way as their universal
14142 counterpart and unknown types were just rejected. Changed so that the
14143 tagged and unknown types are handled in the same way as a SEQUENCE:
14144 that is the encoding is stored intact. There is also a new type
14145 "V_ASN1_OTHER" which is used when the class is not universal, in this
14146 case we have no idea what the actual type is so we just lump them all
14151 * On VMS, stdout may very well lead to a file that is written to
14152 in a record-oriented fashion. That means that every write() will
14153 write a separate record, which will be read separately by the
14154 programs trying to read from it. This can be very confusing.
14156 The solution is to put a BIO filter in the way that will buffer
14157 text until a linefeed is reached, and then write everything a
14158 line at a time, so every record written will be an actual line,
14159 not chunks of lines and not (usually doesn't happen, but I've
14160 seen it once) several lines in one record. BIO_f_linebuffer() is
14163 Currently, it's a VMS-only method, because that's where it has
14164 been tested well enough.
14168 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14169 it can return incorrect results.
14170 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14171 but it was in 0.9.6-beta[12].)
14175 * Disable the check for content being present when verifying detached
14176 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14177 include zero length content when signing messages.
14181 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14182 BIO_ctrl (for BIO pairs).
14186 * Add DSO method for VMS.
14190 * Bug fix: Montgomery multiplication could produce results with the
14195 * Add RPM specification openssl.spec and modify it to build three
14196 packages. The default package contains applications, application
14197 documentation and run-time libraries. The devel package contains
14198 include files, static libraries and function documentation. The
14199 doc package contains the contents of the doc directory. The original
14200 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14204 * Add a large number of documentation files for many SSL routines.
14206 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14208 * Add a configuration entry for Sony News 4.
14210 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14212 * Don't set the two most significant bits to one when generating a
14213 random number < q in the DSA library.
14217 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14218 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14219 the underlying transport is blocking) if a handshake took place.
14220 (The default behaviour is needed by applications such as s_client
14221 and s_server that use select() to determine when to use SSL_read;
14222 but for applications that know in advance when to expect data, it
14223 just makes things more complicated.)
14227 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14232 * Add a few more EBCDIC conditionals that make `req' and `x509'
14233 work better on such systems.
14235 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14237 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14238 Update PKCS12_parse() so it copies the friendlyName and the
14239 keyid to the certificates aux info.
14243 * Fix bug in PKCS7_verify() which caused an infinite loop
14244 if there was more than one signature.
14246 *Sven Uszpelkat <su@celocom.de>*
14248 * Major change in util/mkdef.pl to include extra information
14249 about each symbol, as well as presenting variables as well
14250 as functions. This change means that there's n more need
14251 to rebuild the .num files when some algorithms are excluded.
14255 * Allow the verify time to be set by an application,
14256 rather than always using the current time.
14260 * Phase 2 verify code reorganisation. The certificate
14261 verify code now looks up an issuer certificate by a
14262 number of criteria: subject name, authority key id
14263 and key usage. It also verifies self signed certificates
14264 by the same criteria. The main comparison function is
14265 X509_check_issued() which performs these checks.
14267 Lot of changes were necessary in order to support this
14268 without completely rewriting the lookup code.
14270 Authority and subject key identifier are now cached.
14272 The LHASH 'certs' is X509_STORE has now been replaced
14273 by a STACK_OF(X509_OBJECT). This is mainly because an
14274 LHASH can't store or retrieve multiple objects with
14275 the same hash value.
14277 As a result various functions (which were all internal
14278 use only) have changed to handle the new X509_STORE
14279 structure. This will break anything that messed round
14280 with X509_STORE internally.
14282 The functions X509_STORE_add_cert() now checks for an
14283 exact match, rather than just subject name.
14285 The X509_STORE API doesn't directly support the retrieval
14286 of multiple certificates matching a given criteria, however
14287 this can be worked round by performing a lookup first
14288 (which will fill the cache with candidate certificates)
14289 and then examining the cache for matches. This is probably
14290 the best we can do without throwing out X509_LOOKUP
14291 entirely (maybe later...).
14293 The X509_VERIFY_CTX structure has been enhanced considerably.
14295 All certificate lookup operations now go via a get_issuer()
14296 callback. Although this currently uses an X509_STORE it
14297 can be replaced by custom lookups. This is a simple way
14298 to bypass the X509_STORE hackery necessary to make this
14299 work and makes it possible to use more efficient techniques
14300 in future. A very simple version which uses a simple
14301 STACK for its trusted certificate store is also provided
14302 using X509_STORE_CTX_trusted_stack().
14304 The verify_cb() and verify() callbacks now have equivalents
14305 in the X509_STORE_CTX structure.
14307 X509_STORE_CTX also has a 'flags' field which can be used
14308 to customise the verify behaviour.
14312 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14313 excludes S/MIME capabilities.
14317 * When a certificate request is read in keep a copy of the
14318 original encoding of the signed data and use it when outputting
14319 again. Signatures then use the original encoding rather than
14320 a decoded, encoded version which may cause problems if the
14321 request is improperly encoded.
14325 * For consistency with other BIO_puts implementations, call
14326 buffer_write(b, ...) directly in buffer_puts instead of calling
14329 In BIO_puts, increment b->num_write as in BIO_write.
14331 *Peter.Sylvester@EdelWeb.fr*
14333 * Fix BN_mul_word for the case where the word is 0. (We have to use
14334 BN_zero, we may not return a BIGNUM with an array consisting of
14335 words set to zero.)
14339 * Avoid calling abort() from within the library when problems are
14340 detected, except if preprocessor symbols have been defined
14341 (such as REF_CHECK, BN_DEBUG etc.).
14345 * New openssl application 'rsautl'. This utility can be
14346 used for low level RSA operations. DER public key
14347 BIO/fp routines also added.
14351 * New Configure entry and patches for compiling on QNX 4.
14353 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14355 * A demo state-machine implementation was sponsored by
14356 Nuron (http://www.nuron.com/) and is now available in
14357 demos/state_machine.
14361 * New options added to the 'dgst' utility for signature
14362 generation and verification.
14366 * Unrecognized PKCS#7 content types are now handled via a
14367 catch all ASN1_TYPE structure. This allows unsupported
14368 types to be stored as a "blob" and an application can
14369 encode and decode it manually.
14373 * Fix various signed/unsigned issues to make a_strex.c
14374 compile under VC++.
14376 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14378 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14379 length if passed a buffer. ASN1_INTEGER_to_BN failed
14380 if passed a NULL BN and its argument was negative.
14382 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14384 * Modification to PKCS#7 encoding routines to output definite
14385 length encoding. Since currently the whole structures are in
14386 memory there's not real point in using indefinite length
14387 constructed encoding. However if OpenSSL is compiled with
14388 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14392 * Added BIO_vprintf() and BIO_vsnprintf().
14396 * Added more prefixes to parse for in the strings written
14397 through a logging bio, to cover all the levels that are available
14398 through syslog. The prefixes are now:
14400 PANIC, EMERG, EMR => LOG_EMERG
14401 ALERT, ALR => LOG_ALERT
14402 CRIT, CRI => LOG_CRIT
14403 ERROR, ERR => LOG_ERR
14404 WARNING, WARN, WAR => LOG_WARNING
14405 NOTICE, NOTE, NOT => LOG_NOTICE
14406 INFO, INF => LOG_INFO
14407 DEBUG, DBG => LOG_DEBUG
14409 and as before, if none of those prefixes are present at the
14410 beginning of the string, LOG_ERR is chosen.
14412 On Win32, the LOG_* levels are mapped according to this:
14414 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14415 LOG_WARNING => EVENTLOG_WARNING_TYPE
14416 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14421 * Made it possible to reconfigure with just the configuration
14422 argument "reconf" or "reconfigure". The command line arguments
14423 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14424 and are retrieved from there when reconfiguring.
14430 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14432 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14436 * The obj_dat.pl script was messing up the sorting of object
14437 names. The reason was that it compared the quoted version
14438 of strings as a result "OCSP" > "OCSP Signing" because
14439 " > SPACE. Changed script to store unquoted versions of
14440 names and add quotes on output. It was also omitting some
14441 names from the lookup table if they were given a default
14442 value (that is if SN is missing it is given the same
14443 value as LN and vice versa), these are now added on the
14444 grounds that if an object has a name we should be able to
14445 look it up. Finally added warning output when duplicate
14446 short or long names are found.
14450 * Changes needed for Tandem NSK.
14452 *Scott Uroff <scott@xypro.com>*
14454 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14455 RSA_padding_check_SSLv23(), special padding was never detected
14456 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14457 version rollback attacks was not effective.
14459 In s23_clnt.c, don't use special rollback-attack detection padding
14460 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14461 client; similarly, in s23_srvr.c, don't do the rollback check if
14462 SSL 2.0 is the only protocol enabled in the server.
14466 * Make it possible to get hexdumps of unprintable data with 'openssl
14467 asn1parse'. By implication, the functions ASN1_parse_dump() and
14468 BIO_dump_indent() are added.
14472 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14473 these print out strings and name structures based on various
14474 flags including RFC2253 support and proper handling of
14475 multibyte characters. Added options to the 'x509' utility
14476 to allow the various flags to be set.
14480 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14481 Also change the functions X509_cmp_current_time() and
14482 X509_gmtime_adj() work with an ASN1_TIME structure,
14483 this will enable certificates using GeneralizedTime in validity
14484 dates to be checked.
14488 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14489 negative public key encodings) on by default,
14490 NO_NEG_PUBKEY_BUG can be set to disable it.
14494 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14495 content octets. An i2c_ASN1_OBJECT is unnecessary because
14496 the encoding can be trivially obtained from the structure.
14500 * crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
14501 not read locks (CRYPTO_r_[un]lock).
14505 * A first attempt at creating official support for shared
14506 libraries through configuration. I've kept it so the
14507 default is static libraries only, and the OpenSSL programs
14508 are always statically linked for now, but there are
14509 preparations for dynamic linking in place.
14510 This has been tested on Linux and Tru64.
14514 * Randomness polling function for Win9x, as described in:
14515 Peter Gutmann, Software Generation of Practically Strong
14520 * Fix so PRNG is seeded in req if using an already existing
14525 * New options to smime application. -inform and -outform
14526 allow alternative formats for the S/MIME message including
14527 PEM and DER. The -content option allows the content to be
14528 specified separately. This should allow things like Netscape
14529 form signing output easier to verify.
14533 * Fix the ASN1 encoding of tags using the 'long form'.
14537 * New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
14538 STRING types. These convert content octets to and from the
14539 underlying type. The actual tag and length octets are
14540 already assumed to have been read in and checked. These
14541 are needed because all other string types have virtually
14542 identical handling apart from the tag. By having versions
14543 of the ASN1 functions that just operate on content octets
14544 IMPLICIT tagging can be handled properly. It also allows
14545 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14546 and ASN1_INTEGER are identical apart from the tag.
14550 * Change the handling of OID objects as follows:
14552 - New object identifiers are inserted in objects.txt, following
14553 the syntax given in objects.README.
14554 - objects.pl is used to process obj_mac.num and create a new
14556 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14559 This is currently kind of a hack, and the perl code in objects.pl
14560 isn't very elegant, but it works as I intended. The simplest way
14561 to check that it worked correctly is to look in obj_dat.h and
14562 check the array nid_objs and make sure the objects haven't moved
14563 around (this is important!). Additions are OK, as well as
14564 consistent name changes.
14568 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14572 * Addition of the command line parameter '-rand file' to 'openssl req'.
14573 The given file adds to whatever has already been seeded into the
14574 random pool through the RANDFILE configuration file option or
14575 environment variable, or the default random state file.
14579 * mkstack.pl now sorts each macro group into lexical order.
14580 Previously the output order depended on the order the files
14581 appeared in the directory, resulting in needless rewriting
14586 * Patches to make OpenSSL compile under Win32 again. Mostly
14587 work arounds for the VC++ problem that it treats func() as
14588 func(void). Also stripped out the parts of mkdef.pl that
14589 added extra typesafe functions: these no longer exist.
14593 * Reorganisation of the stack code. The macros are now all
14594 collected in safestack.h . Each macro is defined in terms of
14595 a "stack macro" of the form SKM_<name>(type, a, b). The
14596 DEBUG_SAFESTACK is now handled in terms of function casts,
14597 this has the advantage of retaining type safety without the
14598 use of additional functions. If DEBUG_SAFESTACK is not defined
14599 then the non typesafe macros are used instead. Also modified the
14600 mkstack.pl script to handle the new form. Needs testing to see
14601 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14602 the default if no major problems. Similar behaviour for ASN1_SET_OF
14603 and PKCS12_STACK_OF.
14607 * When some versions of IIS use the 'NET' form of private key the
14608 key derivation algorithm is different. Normally MD5(password) is
14609 used as a 128 bit RC4 key. In the modified case
14610 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14611 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14612 as the old Netscape_RSA functions except they have an additional
14613 'sgckey' parameter which uses the modified algorithm. Also added
14614 an -sgckey command line option to the rsa utility. Thanks to
14615 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14616 algorithm to openssl-dev.
14620 * The evp_local.h macros were using 'c.##kname' which resulted in
14621 invalid expansion on some systems (SCO 5.0.5 for example).
14622 Corrected to 'c.kname'.
14624 *Phillip Porch <root@theporch.com>*
14626 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14627 a STACK of email addresses from a certificate or request, these look
14628 in the subject name and the subject alternative name extensions and
14629 omit any duplicate addresses.
14633 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14634 This makes DSA verification about 2 % faster.
14638 * Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
14639 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14640 plus overhead for 1024 bit moduli).
14641 This makes exponentiations about 0.5 % faster for 1024 bit
14642 exponents (as measured by "openssl speed rsa2048").
14646 * Rename memory handling macros to avoid conflicts with other
14648 Malloc => OPENSSL_malloc
14649 Malloc_locked => OPENSSL_malloc_locked
14650 Realloc => OPENSSL_realloc
14651 Free => OPENSSL_free
14655 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14656 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14660 * CygWin32 support.
14662 *John Jarvie <jjarvie@newsguy.com>*
14664 * The type-safe stack code has been rejigged. It is now only compiled
14665 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14666 by default all type-specific stack functions are "#define"d back to
14667 standard stack functions. This results in more streamlined output
14668 but retains the type-safety checking possibilities of the original
14673 * The STACK code has been cleaned up, and certain type declarations
14674 that didn't make a lot of sense have been brought in line. This has
14675 also involved a cleanup of sorts in safestack.h to more correctly
14676 map type-safe stack functions onto their plain stack counterparts.
14677 This work has also resulted in a variety of "const"ifications of
14678 lots of the code, especially "_cmp" operations which should normally
14679 be prototyped with "const" parameters anyway.
14683 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14684 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14685 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14686 where all of 'md' is used each time the PRNG is used, but 'state'
14687 is used only indexed by a cyclic counter. As entropy may not be
14688 well distributed from the beginning, 'md' is important as a
14689 chaining variable. However, the output function chains only half
14690 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14691 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14692 in all of 'state' being rewritten, with the new values depending
14693 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14697 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14698 the handshake is continued after ssl_verify_cert_chain();
14699 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14700 can lead to 'unexplainable' connection aborts later.
14702 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14704 * Major EVP API cipher revision.
14705 Add hooks for extra EVP features. This allows various cipher
14706 parameters to be set in the EVP interface. Support added for variable
14707 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14708 setting of RC2 and RC5 parameters.
14710 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14713 Remove lots of duplicated code from the EVP library. For example *every*
14714 cipher init() function handles the 'iv' in the same way according to the
14715 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14716 for CFB and OFB modes they zero ctx->num.
14718 New functionality allows removal of S/MIME code RC2 hack.
14720 Most of the routines have the same form and so can be declared in terms
14723 By shifting this to the top level EVP_CipherInit() it can be removed from
14724 all individual ciphers. If the cipher wants to handle IVs or keys
14725 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14728 Change lots of functions like EVP_EncryptUpdate() to now return a
14729 value: although software versions of the algorithms cannot fail
14730 any installed hardware versions can.
14734 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14735 this option is set, tolerate broken clients that send the negotiated
14736 protocol version number instead of the requested protocol version
14741 * Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
14742 i.e. non-zero for export ciphersuites, zero otherwise.
14743 Previous versions had this flag inverted, inconsistent with
14744 rsa_tmp_cb (..._TMP_RSA_CB).
14746 *Bodo Moeller; problem reported by Amit Chopra*
14748 * Add missing DSA library text string. Work around for some IIS
14749 key files with invalid SEQUENCE encoding.
14753 * Add a document (doc/standards.txt) that list all kinds of standards
14754 and so on that are implemented in OpenSSL.
14758 * Enhance c_rehash script. Old version would mishandle certificates
14759 with the same subject name hash and wouldn't handle CRLs at all.
14760 Added -fingerprint option to crl utility, to support new c_rehash
14765 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14769 * Fix for SSL server purpose checking. Server checking was
14770 rejecting certificates which had extended key usage present
14771 but no ssl client purpose.
14773 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14775 * Make PKCS#12 code work with no password. The PKCS#12 spec
14776 is a little unclear about how a blank password is handled.
14777 Since the password in encoded as a BMPString with terminating
14778 double NULL a zero length password would end up as just the
14779 double NULL. However no password at all is different and is
14780 handled differently in the PKCS#12 key generation code. NS
14781 treats a blank password as zero length. MSIE treats it as no
14782 password on export: but it will try both on import. We now do
14783 the same: PKCS12_parse() tries zero length and no password if
14784 the password is set to "" or NULL (NULL is now a valid password:
14785 it wasn't before) as does the pkcs12 application.
14789 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14790 perror when PEM_read_bio_X509_REQ fails, the error message must
14791 be obtained from the error queue.
14795 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14796 it in ERR_remove_state if appropriate, and change ERR_get_state
14797 accordingly to avoid race conditions (this is necessary because
14798 thread_hash is no longer constant once set).
14802 * Bugfix for linux-elf makefile.one.
14806 * RSA_get_default_method() will now cause a default
14807 RSA_METHOD to be chosen if one doesn't exist already.
14808 Previously this was only set during a call to RSA_new()
14809 or RSA_new_method(NULL) meaning it was possible for
14810 RSA_get_default_method() to return NULL.
14814 * Added native name translation to the existing DSO code
14815 that will convert (if the flag to do so is set) filenames
14816 that are sufficiently small and have no path information
14817 into a canonical native form. Eg. "blah" converted to
14818 "libblah.so" or "blah.dll" etc.
14822 * New function ERR_error_string_n(e, buf, len) which is like
14823 ERR_error_string(e, buf), but writes at most 'len' bytes
14824 including the 0 terminator. For ERR_error_string_n, 'buf'
14827 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14829 * CONF library reworked to become more general. A new CONF
14830 configuration file reader "class" is implemented as well as a
14831 new functions (NCONF_*, for "New CONF") to handle it. The now
14832 old CONF_* functions are still there, but are reimplemented to
14833 work in terms of the new functions. Also, a set of functions
14834 to handle the internal storage of the configuration data is
14835 provided to make it easier to write new configuration file
14836 reader "classes" (I can definitely see something reading a
14837 configuration file in XML format, for example), called _CONF_*,
14838 or "the configuration storage API"...
14840 The new configuration file reading functions are:
14842 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14843 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14845 NCONF_default, NCONF_WIN32
14847 NCONF_dump_fp, NCONF_dump_bio
14849 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14850 NCONF_new creates a new CONF object. This works in the same way
14851 as other interfaces in OpenSSL, like the BIO interface.
14852 NCONF_dump_* dump the internal storage of the configuration file,
14853 which is useful for debugging. All other functions take the same
14854 arguments as the old CONF_* functions with the exception of the
14855 first that must be a `CONF *' instead of a `LHASH *'.
14857 To make it easier to use the new classes with the old CONF_* functions,
14858 the function CONF_set_default_method is provided.
14862 * Add '-tls1' option to 'openssl ciphers', which was already
14863 mentioned in the documentation but had not been implemented.
14864 (This option is not yet really useful because even the additional
14865 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14869 * Initial DSO code added into libcrypto for letting OpenSSL (and
14870 OpenSSL-based applications) load shared libraries and bind to
14871 them in a portable way.
14873 *Geoff Thorpe, with contributions from Richard Levitte*
14875 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000] ###
14877 * Make sure _lrotl and _lrotr are only used with MSVC.
14879 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14880 (the default implementation of RAND_status).
14882 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14883 to '-clrext' (= clear extensions), as intended and documented.
14884 *Bodo Moeller; inconsistency pointed out by Michael Attili
14885 <attili@amaxo.com>*
14887 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14888 was larger than the MD block size.
14890 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14892 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14893 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14894 using the passed key: if the passed key was a private key the result
14895 of X509_print(), for example, would be to print out all the private key
14900 * des_quad_cksum() byte order bug fix.
14901 *Ulf Möller, using the problem description in krb4-0.9.7, where
14902 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14904 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14907 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14909 * For easily testing in shell scripts whether some command
14910 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14911 returns with exit code 0 iff no command of the given name is available.
14912 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14913 the output goes to stdout and nothing is printed to stderr.
14914 Additional arguments are always ignored.
14916 Since for each cipher there is a command of the same name,
14917 the 'no-cipher' compilation switches can be tested this way.
14919 ('openssl no-XXX' is not able to detect pseudo-commands such
14920 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
14924 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
14928 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
14929 is set; it will be thrown away anyway because each handshake creates
14931 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
14932 to parameters -- in previous versions (since OpenSSL 0.9.3) the
14933 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
14934 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
14938 * New s_client option -ign_eof: EOF at stdin is ignored, and
14939 'Q' and 'R' lose their special meanings (quit/renegotiate).
14940 This is part of what -quiet does; unlike -quiet, -ign_eof
14941 does not suppress any output.
14945 * Add compatibility options to the purpose and trust code. The
14946 purpose X509_PURPOSE_ANY is "any purpose" which automatically
14947 accepts a certificate or CA, this was the previous behaviour,
14948 with all the associated security issues.
14950 X509_TRUST_COMPAT is the old trust behaviour: only and
14951 automatically trust self signed roots in certificate store. A
14952 new trust setting X509_TRUST_DEFAULT is used to specify that
14953 a purpose has no associated trust setting and it should instead
14954 use the value in the default purpose.
14958 * Fix the PKCS#8 DSA private key code so it decodes keys again
14959 and fix a memory leak.
14963 * In util/mkerr.pl (which implements 'make errors'), preserve
14964 reason strings from the previous version of the .c file, as
14965 the default to have only downcase letters (and digits) in
14966 automatically generated reasons codes is not always appropriate.
14970 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
14971 using strerror. Previously, ERR_reason_error_string() returned
14972 library names as reason strings for SYSerr; but SYSerr is a special
14973 case where small numbers are errno values, not library numbers.
14977 * Add '-dsaparam' option to 'openssl dhparam' application. This
14978 converts DSA parameters into DH parameters. (When creating parameters,
14979 DSA_generate_parameters is used.)
14983 * Include 'length' (recommended exponent length) in C code generated
14984 by 'openssl dhparam -C'.
14988 * The second argument to set_label in perlasm was already being used
14989 so couldn't be used as a "file scope" flag. Moved to third argument
14994 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
14995 instead of RAND_bytes for encryption IVs and salts.
14999 * Include RAND_status() into RAND_METHOD instead of implementing
15000 it only for md_rand.c Otherwise replacing the PRNG by calling
15001 RAND_set_rand_method would be impossible.
15005 * Don't let DSA_generate_key() enter an infinite loop if the random
15006 number generation fails.
15010 * New 'rand' application for creating pseudo-random output.
15014 * Added configuration support for Linux/IA64
15016 *Rolf Haberrecker <rolf@suse.de>*
15018 * Assembler module support for Mingw32.
15022 * Shared library support for HPUX (in shlib/).
15024 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15026 * Shared library support for Solaris gcc.
15028 *Lutz Behnke <behnke@trustcenter.de>*
15030 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000] ###
15032 * PKCS7_encrypt() was adding text MIME headers twice because they
15033 were added manually and by SMIME_crlf_copy().
15037 * In bntest.c don't call BN_rand with zero bits argument.
15039 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15041 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15042 case was implemented. This caused BN_div_recp() to fail occasionally.
15046 * Add an optional second argument to the set_label() in the perl
15047 assembly language builder. If this argument exists and is set
15048 to 1 it signals that the assembler should use a symbol whose
15049 scope is the entire file, not just the current function. This
15050 is needed with MASM which uses the format label:: for this scope.
15052 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15054 * Change the ASN1 types so they are typedefs by default. Before
15055 almost all types were #define'd to ASN1_STRING which was causing
15056 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15061 * Change names of new functions to the new get1/get0 naming
15062 convention: After 'get1', the caller owns a reference count
15063 and has to call ..._free; 'get0' returns a pointer to some
15064 data structure without incrementing reference counters.
15065 (Some of the existing 'get' functions increment a reference
15066 counter, some don't.)
15067 Similarly, 'set1' and 'add1' functions increase reference
15068 counters or duplicate objects.
15072 * Allow for the possibility of temp RSA key generation failure:
15073 the code used to assume it always worked and crashed on failure.
15077 * Fix potential buffer overrun problem in BIO_printf().
15078 *Ulf Möller, using public domain code by Patrick Powell; problem
15079 pointed out by David Sacerdote <das33@cornell.edu>*
15081 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15082 RAND_egd() and RAND_status(). In the command line application,
15083 the EGD socket can be specified like a seed file using RANDFILE
15088 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15089 Some CAs (e.g. Verisign) distribute certificates in this form.
15093 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15094 list to exclude them. This means that no special compilation option
15095 is needed to use anonymous DH: it just needs to be included in the
15100 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15101 EVP_MD_type. The old functionality is available in a new macro called
15102 EVP_MD_md(). Change code that uses it and update docs.
15106 * ..._ctrl functions now have corresponding ..._callback_ctrl functions
15107 where the 'void *' argument is replaced by a function pointer argument.
15108 Previously 'void *' was abused to point to functions, which works on
15109 many platforms, but is not correct. As these functions are usually
15110 called by macros defined in OpenSSL header files, most source code
15111 should work without changes.
15115 * <openssl/opensslconf.h> (which is created by Configure) now contains
15116 sections with information on -D... compiler switches used for
15117 compiling the library so that applications can see them. To enable
15118 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
15119 must be defined. E.g.,
15120 #define OPENSSL_ALGORITHM_DEFINES
15121 #include <openssl/opensslconf.h>
15122 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
15124 *Richard Levitte, Ulf and Bodo Möller*
15126 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15131 * Change the 'other' type in certificate aux info to a STACK_OF
15132 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15133 the required ASN1 format: arbitrary types determined by an OID.
15137 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15138 argument to 'req'. This is not because the function is newer or
15139 better than others it just uses the work 'NEW' in the certificate
15140 request header lines. Some software needs this.
15144 * Reorganise password command line arguments: now passwords can be
15145 obtained from various sources. Delete the PEM_cb function and make
15146 it the default behaviour: i.e. if the callback is NULL and the
15147 usrdata argument is not NULL interpret it as a null terminated pass
15148 phrase. If usrdata and the callback are NULL then the pass phrase
15149 is prompted for as usual.
15153 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15154 the support is automatically enabled. The resulting binaries will
15155 autodetect the card and use it if present.
15157 *Ben Laurie and Compaq Inc.*
15159 * Work around for Netscape hang bug. This sends certificate request
15160 and server done in one record. Since this is perfectly legal in the
15161 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15162 the bugs/SSLv3 entry for more info.
15166 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15170 * Add -rand argument to smime and pkcs12 applications and read/write
15175 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15179 * Add command line password options to the remaining applications.
15183 * Bug fix for BN_div_recp() for numerators with an even number of
15188 * More tests in bntest.c, and changed test_bn output.
15192 * ./config recognizes MacOS X now.
15196 * Bug fix for BN_div() when the first words of num and divisor are
15197 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
15201 * Add support for various broken PKCS#8 formats, and command line
15202 options to produce them.
15206 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15207 get temporary BIGNUMs from a BN_CTX.
15211 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15216 * Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
15217 include a #define from the old name to the new. The original intent
15218 was that statically linked binaries could for example just call
15219 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15220 link with digests. This never worked because SSLeay_add_all_digests()
15221 and SSLeay_add_all_ciphers() were in the same source file so calling
15222 one would link with the other. They are now in separate source files.
15226 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15230 * Use a less unusual form of the Miller-Rabin primality test (it used
15231 a binary algorithm for exponentiation integrated into the Miller-Rabin
15232 loop, our standard modexp algorithms are faster).
15236 * Support for the EBCDIC character set completed.
15238 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15240 * Source code cleanups: use const where appropriate, eliminate casts,
15241 use void * instead of char * in lhash.
15245 * Bugfix: ssl3_send_server_key_exchange was not restartable
15246 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15247 this the server could overwrite ephemeral keys that the client
15252 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15253 using 50 iterations of the Rabin-Miller test.
15255 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15256 iterations of the Rabin-Miller test as required by the appendix
15257 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15258 As BN_is_prime_fasttest includes trial division, DSA parameter
15259 generation becomes much faster.
15261 This implies a change for the callback functions in DSA_is_prime
15262 and DSA_generate_parameters: The callback function is called once
15263 for each positive witness in the Rabin-Miller test, not just
15264 occasionally in the inner loop; and the parameters to the
15265 callback function now provide an iteration count for the outer
15266 loop rather than for the current invocation of the inner loop.
15267 DSA_generate_parameters additionally can call the callback
15268 function with an 'iteration count' of -1, meaning that a
15269 candidate has passed the trial division test (when q is generated
15270 from an application-provided seed, trial division is skipped).
15274 * New function BN_is_prime_fasttest that optionally does trial
15275 division before starting the Rabin-Miller test and has
15276 an additional BN_CTX * argument (whereas BN_is_prime always
15277 has to allocate at least one BN_CTX).
15278 'callback(1, -1, cb_arg)' is called when a number has passed the
15279 trial division stage.
15283 * Fix for bug in CRL encoding. The validity dates weren't being handled
15288 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15292 * New function BN_pseudo_rand().
15296 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15297 bignum version of BN_from_montgomery() with the working code from
15298 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15303 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15304 made it impossible to use the same SSL_SESSION data structure in
15305 SSL2 clients in multiple threads.
15309 * The return value of RAND_load_file() no longer counts bytes obtained
15310 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15311 to seed the PRNG (previously an explicit byte count was required).
15313 *Ulf Möller, Bodo Möller*
15315 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15316 used (char *) instead of (void *) and had casts all over the place.
15320 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15324 * Retain source code compatibility for BN_prime_checks macro:
15325 BN_is_prime(..., BN_prime_checks, ...) now uses
15326 BN_prime_checks_for_size to determine the appropriate number of
15327 Rabin-Miller iterations.
15331 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15332 DH_CHECK_P_NOT_SAFE_PRIME.
15333 (Check if this is true? OpenPGP calls them "strong".)
15337 * Merge the functionality of "dh" and "gendh" programs into a new program
15338 "dhparam". The old programs are retained for now but will handle DH keys
15339 (instead of parameters) in future.
15343 * Make the ciphers, s_server and s_client programs check the return values
15344 when a new cipher list is set.
15348 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15349 ciphers. Before when the 56bit ciphers were enabled the sorting was
15352 The syntax for the cipher sorting has been extended to support sorting by
15353 cipher-strength (using the strength_bits hard coded in the tables).
15354 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15356 Fix a bug in the cipher-command parser: when supplying a cipher command
15357 string with an "undefined" symbol (neither command nor alphanumeric
15358 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15359 an error is flagged.
15361 Due to the strength-sorting extension, the code of the
15362 ssl_create_cipher_list() function was completely rearranged. I hope that
15363 the readability was also increased :-)
15365 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15367 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15368 for the first serial number and places 2 in the serial number file. This
15369 avoids problems when the root CA is created with serial number zero and
15370 the first user certificate has the same issuer name and serial number
15375 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15376 the new code. Add documentation for this stuff.
15380 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15381 X509_*() to X509at_*() on the grounds that they don't handle X509
15382 structures and behave in an analogous way to the X509v3 functions:
15383 they shouldn't be called directly but wrapper functions should be used
15386 So we also now have some wrapper functions that call the X509at functions
15387 when passed certificate requests. (TO DO: similar things can be done with
15388 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15389 things. Some of these need some d2i or i2d and print functionality
15390 because they handle more complex structures.)
15394 * Add missing #ifndefs that caused missing symbols when building libssl
15395 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15396 NO_RSA in ssl/s2*.c.
15398 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15400 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15401 has a return value which indicates the quality of the random data
15402 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15403 error queue. New function RAND_pseudo_bytes() generates output that is
15404 guaranteed to be unique but not unpredictable. RAND_add is like
15405 RAND_seed, but takes an extra argument for an entropy estimate
15406 (RAND_seed always assumes full entropy).
15410 * Do more iterations of Rabin-Miller probable prime test (specifically,
15411 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15412 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15413 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15414 false-positive rate of at most 2^-80 for random input.
15418 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15422 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15423 in the 0.9.5 release), this returns the chain
15424 from an X509_CTX structure with a dup of the stack and all
15425 the X509 reference counts upped: so the stack will exist
15426 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15429 Also make SSL_SESSION_print() print out the verify return
15434 * Add manpage for the pkcs12 command. Also change the default
15435 behaviour so MAC iteration counts are used unless the new
15436 -nomaciter option is used. This improves file security and
15437 only older versions of MSIE (4.0 for example) need it.
15441 * Honor the no-xxx Configure options when creating .DEF files.
15445 * Add PKCS#10 attributes to field table: challengePassword,
15446 unstructuredName and unstructuredAddress. These are taken from
15447 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15448 international characters are used.
15450 More changes to X509_ATTRIBUTE code: allow the setting of types
15451 based on strings. Remove the 'loc' parameter when adding
15452 attributes because these will be a SET OF encoding which is sorted
15457 * Initial changes to the 'req' utility to allow request generation
15458 automation. This will allow an application to just generate a template
15459 file containing all the field values and have req construct the
15462 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15463 used all over the place including certificate requests and PKCS#7
15464 structures. They are currently handled manually where necessary with
15465 some primitive wrappers for PKCS#7. The new functions behave in a
15466 manner analogous to the X509 extension functions: they allow
15467 attributes to be looked up by NID and added.
15469 Later something similar to the X509V3 code would be desirable to
15470 automatically handle the encoding, decoding and printing of the
15471 more complex types. The string types like challengePassword can
15472 be handled by the string table functions.
15474 Also modified the multi byte string table handling. Now there is
15475 a 'global mask' which masks out certain types. The table itself
15476 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15477 is useful when for example there is only one permissible type
15478 (as in countryName) and using the mask might result in no valid
15483 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15484 SSL_get_peer_finished to allow applications to obtain the latest
15485 Finished messages sent to the peer or expected from the peer,
15486 respectively. (SSL_get_peer_finished is usually the Finished message
15487 actually received from the peer, otherwise the protocol will be aborted.)
15489 As the Finished message are message digests of the complete handshake
15490 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15491 be used for external authentication procedures when the authentication
15492 provided by SSL/TLS is not desired or is not enough.
15496 * Enhanced support for Alpha Linux is added. Now ./config checks if
15497 the host supports BWX extension and if Compaq C is present on the
15498 $PATH. Just exploiting of the BWX extension results in 20-30%
15499 performance kick for some algorithms, e.g. DES and RC4 to mention
15500 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15505 * Add support for MS "fast SGC". This is arguably a violation of the
15506 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15507 weak crypto and after checking the certificate is SGC a second one
15508 with strong crypto. MS SGC stops the first handshake after receiving
15509 the server certificate message and sends a second client hello. Since
15510 a server will typically do all the time consuming operations before
15511 expecting any further messages from the client (server key exchange
15512 is the most expensive) there is little difference between the two.
15514 To get OpenSSL to support MS SGC we have to permit a second client
15515 hello message after we have sent server done. In addition we have to
15516 reset the MAC if we do get this second client hello.
15520 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15521 if a DER encoded private key is RSA or DSA traditional format. Changed
15522 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15523 format DER encoded private key. Newer code should use PKCS#8 format which
15524 has the key type encoded in the ASN1 structure. Added DER private key
15525 support to pkcs8 application.
15529 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15530 ciphersuites has been selected (as required by the SSL 3/TLS 1
15531 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15532 is set, we interpret this as a request to violate the specification
15533 (the worst that can happen is a handshake failure, and 'correct'
15534 behaviour would result in a handshake failure anyway).
15538 * In SSL_CTX_add_session, take into account that there might be multiple
15539 SSL_SESSION structures with the same session ID (e.g. when two threads
15540 concurrently obtain them from an external cache).
15541 The internal cache can handle only one SSL_SESSION with a given ID,
15542 so if there's a conflict, we now throw out the old one to achieve
15547 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15548 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15549 some routines that use cipher OIDs: some ciphers do not have OIDs
15550 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15555 * Simplify the trust setting structure and code. Now we just have
15556 two sequences of OIDs for trusted and rejected settings. These will
15557 typically have values the same as the extended key usage extension
15558 and any application specific purposes.
15560 The trust checking code now has a default behaviour: it will just
15561 check for an object with the same NID as the passed id. Functions can
15562 be provided to override either the default behaviour or the behaviour
15563 for a given id. SSL client, server and email already have functions
15564 in place for compatibility: they check the NID and also return "trusted"
15565 if the certificate is self signed.
15569 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15570 traditional format into an EVP_PKEY structure.
15574 * Add a password callback function PEM_cb() which either prompts for
15575 a password if usr_data is NULL or otherwise assumes it is a null
15576 terminated password. Allow passwords to be passed on command line
15577 environment or config files in a few more utilities.
15581 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15582 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15583 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15584 Update documentation.
15588 * Support for ASN1 "NULL" type. This could be handled before by using
15589 ASN1_TYPE but there wasn't any function that would try to read a NULL
15590 and produce an error if it couldn't. For compatibility we also have
15591 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15592 don't allocate anything because they don't need to.
15596 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15599 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15601 * Rebuild of the memory allocation routines used by OpenSSL code and
15602 possibly others as well. The purpose is to make an interface that
15603 provide hooks so anyone can build a separate set of allocation and
15604 deallocation routines to be used by OpenSSL, for example memory
15605 pool implementations, or something else, which was previously hard
15606 since Malloc(), Realloc() and Free() were defined as macros having
15607 the values malloc, realloc and free, respectively (except for Win32
15608 compilations). The same is provided for memory debugging code.
15609 OpenSSL already comes with functionality to find memory leaks, but
15610 this gives people a chance to debug other memory problems.
15612 With these changes, a new set of functions and macros have appeared:
15614 CRYPTO_set_mem_debug_functions() [F]
15615 CRYPTO_get_mem_debug_functions() [F]
15616 CRYPTO_dbg_set_options() [F]
15617 CRYPTO_dbg_get_options() [F]
15618 CRYPTO_malloc_debug_init() [M]
15620 The memory debug functions are NULL by default, unless the library
15621 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15622 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15623 gives the standard debugging functions that come with OpenSSL) or
15624 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15625 provided by the library user) must be used. When the standard
15626 debugging functions are used, CRYPTO_dbg_set_options can be used to
15627 request additional information:
15628 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15629 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15631 Also, things like CRYPTO_set_mem_functions will always give the
15632 expected result (the new set of functions is used for allocation
15633 and deallocation) at all times, regardless of platform and compiler
15636 To finish it up, some functions that were never use in any other
15637 way than through macros have a new API and new semantic:
15639 CRYPTO_dbg_malloc()
15640 CRYPTO_dbg_realloc()
15643 All macros of value have retained their old syntax.
15645 *Richard Levitte and Bodo Moeller*
15647 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15648 ordering of SMIMECapabilities wasn't in "strength order" and there
15649 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15654 * Some ASN1 types with illegal zero length encoding (INTEGER,
15655 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15657 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15659 * Merge in my S/MIME library for OpenSSL. This provides a simple
15660 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15661 functionality to handle multipart/signed properly) and a utility
15662 called 'smime' to call all this stuff. This is based on code I
15663 originally wrote for Celo who have kindly allowed it to be
15664 included in OpenSSL.
15668 * Add variants des_set_key_checked and des_set_key_unchecked of
15669 des_set_key (aka des_key_sched). Global variable des_check_key
15670 decides which of these is called by des_set_key; this way
15671 des_check_key behaves as it always did, but applications and
15672 the library itself, which was buggy for des_check_key == 1,
15673 have a cleaner way to pick the version they need.
15677 * New function PKCS12_newpass() which changes the password of a
15682 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15683 dynamic mix. In both cases the ids can be used as an index into the
15684 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15685 functions so they accept a list of the field values and the
15686 application doesn't need to directly manipulate the X509_TRUST
15691 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15696 * Modify the way the V3 extension code looks up extensions. This now
15697 works in a similar way to the object code: we have some "standard"
15698 extensions in a static table which is searched with OBJ_bsearch()
15699 and the application can add dynamic ones if needed. The file
15700 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15701 updated whenever a new extension is added to the core code and kept
15702 in ext_nid order. There is a simple program 'tabtest.c' which checks
15703 this. New extensions are not added too often so this file can readily
15704 be maintained manually.
15706 There are two big advantages in doing things this way. The extensions
15707 can be looked up immediately and no longer need to be "added" using
15708 X509V3_add_standard_extensions(): this function now does nothing.
15709 *Side note: I get *lots* of email saying the extension code doesn't
15710 work because people forget to call this function*
15711 Also no dynamic allocation is done unless new extensions are added:
15712 so if we don't add custom extensions there is no need to call
15713 X509V3_EXT_cleanup().
15717 * Modify enc utility's salting as follows: make salting the default. Add a
15718 magic header, so unsalted files fail gracefully instead of just decrypting
15719 to garbage. This is because not salting is a big security hole, so people
15720 should be discouraged from doing it.
15724 * Fixes and enhancements to the 'x509' utility. It allowed a message
15725 digest to be passed on the command line but it only used this
15726 parameter when signing a certificate. Modified so all relevant
15727 operations are affected by the digest parameter including the
15728 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15729 DSA key was used because it didn't fix the digest.
15733 * Initial certificate chain verify code. Currently tests the untrusted
15734 certificates for consistency with the verify purpose (which is set
15735 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15737 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15738 this is because it will reject chains with invalid extensions whereas
15739 every previous version of OpenSSL and SSLeay made no checks at all.
15741 Trust code: checks the root CA for the relevant trust settings. Trust
15742 settings have an initial value consistent with the verify purpose: e.g.
15743 if the verify purpose is for SSL client use it expects the CA to be
15744 trusted for SSL client use. However the default value can be changed to
15745 permit custom trust settings: one example of this would be to only trust
15746 certificates from a specific "secure" set of CAs.
15748 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15749 which should be used for version portability: especially since the
15750 verify structure is likely to change more often now.
15752 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15753 to set them. If not set then assume SSL clients will verify SSL servers
15756 Two new options to the verify program: -untrusted allows a set of
15757 untrusted certificates to be passed in and -purpose which sets the
15758 intended purpose of the certificate. If a purpose is set then the
15759 new chain verify code is used to check extension consistency.
15763 * Support for the authority information access extension.
15767 * Modify RSA and DSA PEM read routines to transparently handle
15768 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15769 public keys in a format compatible with certificate
15770 SubjectPublicKeyInfo structures. Unfortunately there were already
15771 functions called *_PublicKey_* which used various odd formats so
15772 these are retained for compatibility: however the DSA variants were
15773 never in a public release so they have been deleted. Changed dsa/rsa
15774 utilities to handle the new format: note no releases ever handled public
15775 keys so we should be OK.
15777 The primary motivation for this change is to avoid the same fiasco
15778 that dogs private keys: there are several incompatible private key
15779 formats some of which are standard and some OpenSSL specific and
15780 require various evil hacks to allow partial transparent handling and
15781 even then it doesn't work with DER formats. Given the option anything
15782 other than PKCS#8 should be dumped: but the other formats have to
15783 stay in the name of compatibility.
15785 With public keys and the benefit of hindsight one standard format
15786 is used which works with EVP_PKEY, RSA or DSA structures: though
15787 it clearly returns an error if you try to read the wrong kind of key.
15789 Added a -pubkey option to the 'x509' utility to output the public key.
15790 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
15791 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
15792 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
15793 that do the same as the EVP_PKEY_assign_*() except they up the
15794 reference count of the added key (they don't "swallow" the
15799 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15800 CRLs would fail if the file contained no certificates or no CRLs:
15801 added a new function to read in both types and return the number
15802 read: this means that if none are read it will be an error. The
15803 DER versions of the certificate and CRL reader would always fail
15804 because it isn't possible to mix certificates and CRLs in DER format
15805 without choking one or the other routine. Changed this to just read
15806 a certificate: this is the best we can do. Also modified the code
15807 in apps/verify.c to take notice of return codes: it was previously
15808 attempting to read in certificates from NULL pointers and ignoring
15809 any errors: this is one reason why the cert and CRL reader seemed
15810 to work. It doesn't check return codes from the default certificate
15811 routines: these may well fail if the certificates aren't installed.
15815 * Code to support otherName option in GeneralName.
15819 * First update to verify code. Change the verify utility
15820 so it warns if it is passed a self signed certificate:
15821 for consistency with the normal behaviour. X509_verify
15822 has been modified to it will now verify a self signed
15823 certificate if *exactly* the same certificate appears
15824 in the store: it was previously impossible to trust a
15825 single self signed certificate. This means that:
15826 openssl verify ss.pem
15827 now gives a warning about a self signed certificate but
15828 openssl verify -CAfile ss.pem ss.pem
15833 * For servers, store verify_result in SSL_SESSION data structure
15834 (and add it to external session representation).
15835 This is needed when client certificate verifications fails,
15836 but an application-provided verification callback (set by
15837 SSL_CTX_set_cert_verify_callback) allows accepting the session
15838 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15839 but returns 1): When the session is reused, we have to set
15840 ssl->verify_result to the appropriate error code to avoid
15843 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15845 * Fix a bug in the new PKCS#7 code: it didn't consider the
15846 case in PKCS7_dataInit() where the signed PKCS7 structure
15847 didn't contain any existing data because it was being created.
15849 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15851 * Add a salt to the key derivation routines in enc.c. This
15852 forms the first 8 bytes of the encrypted file. Also add a
15853 -S option to allow a salt to be input on the command line.
15857 * New function X509_cmp(). Oddly enough there wasn't a function
15858 to compare two certificates. We do this by working out the SHA1
15859 hash and comparing that. X509_cmp() will be needed by the trust
15864 * SSL_get1_session() is like SSL_get_session(), but increments
15865 the reference count in the SSL_SESSION returned.
15867 *Geoff Thorpe <geoff@eu.c2.net>*
15869 * Fix for 'req': it was adding a null to request attributes.
15870 Also change the X509_LOOKUP and X509_INFO code to handle
15871 certificate auxiliary information.
15875 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15880 * Add the possibility to add extra information to the memory leak
15881 detecting output, to form tracebacks, showing from where each
15882 allocation was originated: CRYPTO_push_info("constant string") adds
15883 the string plus current file name and line number to a per-thread
15884 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15885 is like calling CYRPTO_pop_info() until the stack is empty.
15886 Also updated memory leak detection code to be multi-thread-safe.
15890 * Add options -text and -noout to pkcs7 utility and delete the
15891 encryption options which never did anything. Update docs.
15895 * Add options to some of the utilities to allow the pass phrase
15896 to be included on either the command line (not recommended on
15897 OSes like Unix) or read from the environment. Update the
15898 manpages and fix a few bugs.
15902 * Add a few manpages for some of the openssl commands.
15906 * Fix the -revoke option in ca. It was freeing up memory twice,
15907 leaking and not finding already revoked certificates.
15911 * Extensive changes to support certificate auxiliary information.
15912 This involves the use of X509_CERT_AUX structure and X509_AUX
15913 functions. An X509_AUX function such as PEM_read_X509_AUX()
15914 can still read in a certificate file in the usual way but it
15915 will also read in any additional "auxiliary information". By
15916 doing things this way a fair degree of compatibility can be
15917 retained: existing certificates can have this information added
15918 using the new 'x509' options.
15920 Current auxiliary information includes an "alias" and some trust
15921 settings. The trust settings will ultimately be used in enhanced
15922 certificate chain verification routines: currently a certificate
15923 can only be trusted if it is self signed and then it is trusted
15928 * Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
15929 The problem was that one of the replacement routines had not been working
15930 since SSLeay releases. For now the offending routine has been replaced
15931 with non-optimised assembler. Even so, this now gives around 95%
15932 performance improvement for 1024 bit RSA signs.
15936 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
15937 handling. Most clients have the effective key size in bits equal to
15938 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
15939 A few however don't do this and instead use the size of the decrypted key
15940 to determine the RC2 key length and the AlgorithmIdentifier to determine
15941 the effective key length. In this case the effective key length can still
15942 be 40 bits but the key length can be 168 bits for example. This is fixed
15943 by manually forcing an RC2 key into the EVP_PKEY structure because the
15944 EVP code can't currently handle unusual RC2 key sizes: it always assumes
15945 the key length and effective key length are equal.
15949 * Add a bunch of functions that should simplify the creation of
15950 X509_NAME structures. Now you should be able to do:
15951 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
15952 and have it automatically work out the correct field type and fill in
15953 the structures. The more adventurous can try:
15954 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
15955 and it will (hopefully) work out the correct multibyte encoding.
15959 * Change the 'req' utility to use the new field handling and multibyte
15960 copy routines. Before the DN field creation was handled in an ad hoc
15961 way in req, ca, and x509 which was rather broken and didn't support
15962 BMPStrings or UTF8Strings. Since some software doesn't implement
15963 BMPStrings or UTF8Strings yet, they can be enabled using the config file
15964 using the dirstring_type option. See the new comment in the default
15965 openssl.cnf for more info.
15969 * Make crypto/rand/md_rand.c more robust:
15970 - Assure unique random numbers after fork().
15971 - Make sure that concurrent threads access the global counter and
15972 md serializably so that we never lose entropy in them
15973 or use exactly the same state in multiple threads.
15974 Access to the large state is not always serializable because
15975 the additional locking could be a performance killer, and
15976 md should be large enough anyway.
15980 * New file apps/app_rand.c with commonly needed functionality
15981 for handling the random seed file.
15983 Use the random seed file in some applications that previously did not:
15985 dsaparam -genkey (which also ignored its '-rand' option),
15988 x509 (when signing).
15989 Except on systems with /dev/urandom, it is crucial to have a random
15990 seed file at least for key creation, DSA signing, and for DH exchanges;
15991 for RSA signatures we could do without one.
15993 gendh and gendsa (unlike genrsa) used to read only the first byte
15994 of each file listed in the '-rand' option. The function as previously
15995 found in genrsa is now in app_rand.c and is used by all programs
15996 that support '-rand'.
16000 * In RAND_write_file, use mode 0600 for creating files;
16001 don't just chmod when it may be too late.
16005 * Report an error from X509_STORE_load_locations
16006 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16010 * New function ASN1_mbstring_copy() this copies a string in either
16011 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16012 into an ASN1_STRING type. A mask of permissible types is passed
16013 and it chooses the "minimal" type to use or an error if not type
16018 * Add function equivalents to the various macros in asn1.h. The old
16019 macros are retained with an M_ prefix. Code inside the library can
16020 use the M_ macros. External code (including the openssl utility)
16021 should *NOT* in order to be "shared library friendly".
16025 * Add various functions that can check a certificate's extensions
16026 to see if it usable for various purposes such as SSL client,
16027 server or S/MIME and CAs of these types. This is currently
16028 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16029 verification. Also added a -purpose flag to x509 utility to
16030 print out all the purposes.
16034 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16039 * New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
16040 for, obtain and decode and extension and obtain its critical flag.
16041 This allows all the necessary extension code to be handled in a
16042 single function call.
16046 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16047 platforms. See crypto/rc4/rc4_enc.c for further details.
16051 * New -noout option to asn1parse. This causes no output to be produced
16052 its main use is when combined with -strparse and -out to extract data
16053 from a file (which may not be in ASN.1 format).
16057 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16058 when producing the local key id.
16060 *Richard Levitte <levitte@stacken.kth.se>*
16062 * New option -dhparam in s_server. This allows a DH parameter file to be
16063 stated explicitly. If it is not stated then it tries the first server
16064 certificate file. The previous behaviour hard coded the filename
16069 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16070 a public key to be input or output. For example:
16071 openssl rsa -in key.pem -pubout -out pubkey.pem
16072 Also added necessary DSA public key functions to handle this.
16076 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16077 in the message. This was handled by allowing
16078 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16080 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16082 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16083 to the end of the strings whereas this didn't. This would cause problems
16084 if strings read with d2i_ASN1_bytes() were later modified.
16086 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16088 * Fix for base64 decode bug. When a base64 bio reads only one line of
16089 data and it contains EOF it will end up returning an error. This is
16090 caused by input 46 bytes long. The cause is due to the way base64
16091 BIOs find the start of base64 encoded data. They do this by trying a
16092 trial decode on each line until they find one that works. When they
16093 do a flag is set and it starts again knowing it can pass all the
16094 data directly through the decoder. Unfortunately it doesn't reset
16095 the context it uses. This means that if EOF is reached an attempt
16096 is made to pass two EOFs through the context and this causes the
16097 resulting error. This can also cause other problems as well. As is
16098 usual with these problems it takes *ages* to find and the fix is
16099 trivial: move one line.
16101 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) *
16103 * Ugly workaround to get s_client and s_server working under Windows. The
16104 old code wouldn't work because it needed to select() on sockets and the
16105 tty (for keypresses and to see if data could be written). Win32 only
16106 supports select() on sockets so we select() with a 1s timeout on the
16107 sockets and then see if any characters are waiting to be read, if none
16108 are present then we retry, we also assume we can always write data to
16109 the tty. This isn't nice because the code then blocks until we've
16110 received a complete line of data and it is effectively polling the
16111 keyboard at 1s intervals: however it's quite a bit better than not
16112 working at all :-) A dedicated Windows application might handle this
16113 with an event loop for example.
16117 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16118 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16119 will be called when RSA_sign() and RSA_verify() are used. This is useful
16120 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16121 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16122 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16123 This necessitated the support of an extra signature type NID_md5_sha1
16124 for SSL signatures and modifications to the SSL library to use it instead
16125 of calling RSA_public_decrypt() and RSA_private_encrypt().
16129 * Add new -verify -CAfile and -CApath options to the crl program, these
16130 will lookup a CRL issuers certificate and verify the signature in a
16131 similar way to the verify program. Tidy up the crl program so it
16132 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16133 less strict. It will now permit CRL extensions even if it is not
16134 a V2 CRL: this will allow it to tolerate some broken CRLs.
16138 * Initialize all non-automatic variables each time one of the openssl
16139 sub-programs is started (this is necessary as they may be started
16140 multiple times from the "OpenSSL>" prompt).
16142 *Lennart Bang, Bodo Moeller*
16144 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16145 removing all other RSA functionality (this is what NO_RSA does). This
16146 is so (for example) those in the US can disable those operations covered
16147 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16152 * Non-copying interface to BIO pairs.
16153 (still largely untested)
16157 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16158 ASCII string. This was handled independently in various places before.
16162 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16163 UTF8 strings a character at a time.
16167 * Use client_version from client hello to select the protocol
16168 (s23_srvr.c) and for RSA client key exchange verification
16169 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16173 * Add various utility functions to handle SPKACs, these were previously
16174 handled by poking round in the structure internals. Added new function
16175 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16176 print, verify and generate SPKACs. Based on an original idea from
16177 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16181 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16185 * Allow the config file extension section to be overwritten on the
16186 command line. Based on an original idea from Massimiliano Pala
16187 <madwolf@comune.modena.it>. The new option is called -extensions
16188 and can be applied to ca, req and x509. Also -reqexts to override
16189 the request extensions in req and -crlexts to override the crl extensions
16194 * Add new feature to the SPKAC handling in ca. Now you can include
16195 the same field multiple times by preceding it by "XXXX." for example:
16198 this is the same syntax as used in the req config file.
16202 * Allow certificate extensions to be added to certificate requests. These
16203 are specified in a 'req_extensions' option of the req section of the
16204 config file. They can be printed out with the -text option to req but
16205 are otherwise ignored at present.
16209 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16210 data read consists of only the final block it would not decrypted because
16211 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16212 A misplaced 'break' also meant the decrypted final block might not be
16213 copied until the next read.
16217 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16218 a few extra parameters to the DH structure: these will be useful if
16219 for example we want the value of 'q' or implement X9.42 DH.
16223 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16224 provides hooks that allow the default DSA functions or functions on a
16225 "per key" basis to be replaced. This allows hardware acceleration and
16226 hardware key storage to be handled without major modification to the
16227 library. Also added low level modexp hooks and CRYPTO_EX structure and
16228 associated functions.
16232 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16233 as "read only": it can't be written to and the buffer it points to will
16234 not be freed. Reading from a read only BIO is much more efficient than
16235 a normal memory BIO. This was added because there are several times when
16236 an area of memory needs to be read from a BIO. The previous method was
16237 to create a memory BIO and write the data to it, this results in two
16238 copies of the data and an O(n^2) reading algorithm. There is a new
16239 function BIO_new_mem_buf() which creates a read only memory BIO from
16240 an area of memory. Also modified the PKCS#7 routines to use read only
16245 * Bugfix: ssl23_get_client_hello did not work properly when called in
16246 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16247 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16248 but a retry condition occurred while trying to read the rest.
16252 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16253 NID_pkcs7_encrypted by default: this was wrong since this should almost
16254 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16255 the encrypted data type: this is a more sensible place to put it and it
16256 allows the PKCS#12 code to be tidied up that duplicated this
16261 * Changed obj_dat.pl script so it takes its input and output files on
16262 the command line. This should avoid shell escape redirection problems
16267 * Initial support for certificate extension requests, these are included
16268 in things like Xenroll certificate requests. Included functions to allow
16269 extensions to be obtained and added.
16273 * -crlf option to s_client and s_server for sending newlines as
16274 CRLF (as required by many protocols).
16278 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999] ###
16280 * Install libRSAglue.a when OpenSSL is built with RSAref.
16282 *Ralf S. Engelschall*
16284 * A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
16286 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16288 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16293 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16294 DH parameters/keys (q is lost during that conversion, but the resulting
16295 DH parameters contain its length).
16297 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16298 much faster than DH_generate_parameters (which creates parameters
16299 where p = 2*q + 1), and also the smaller q makes DH computations
16300 much more efficient (160-bit exponentiation instead of 1024-bit
16301 exponentiation); so this provides a convenient way to support DHE
16302 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16303 utter importance to use
16304 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16306 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16307 when such DH parameters are used, because otherwise small subgroup
16308 attacks may become possible!
16312 * Avoid memory leak in i2d_DHparams.
16316 * Allow the -k option to be used more than once in the enc program:
16317 this allows the same encrypted message to be read by multiple recipients.
16321 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16322 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16323 it will always use the numerical form of the OID, even if it has a short
16328 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16329 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16330 otherwise bn_mod_exp was called. In the case of hardware keys for example
16331 no private key components need be present and it might store extra data
16332 in the RSA structure, which cannot be accessed from bn_mod_exp.
16333 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16334 private key operations.
16338 * Added support for SPARC Linux.
16342 * pem_password_cb function type incompatibly changed from
16343 typedef int pem_password_cb(char *buf, int size, int rwflag);
16345 ....(char *buf, int size, int rwflag, void *userdata);
16346 so that applications can pass data to their callbacks:
16347 The PEM[_ASN1]_{read,write}... functions and macros now take an
16348 additional void * argument, which is just handed through whenever
16349 the password callback is called.
16351 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16353 New function SSL_CTX_set_default_passwd_cb_userdata.
16355 Compatibility note: As many C implementations push function arguments
16356 onto the stack in reverse order, the new library version is likely to
16357 interoperate with programs that have been compiled with the old
16358 pem_password_cb definition (PEM_whatever takes some data that
16359 happens to be on the stack as its last argument, and the callback
16360 just ignores this garbage); but there is no guarantee whatsoever that
16363 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16364 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16365 problems not only on Windows, but also on some Unix platforms.
16366 To avoid problematic command lines, these definitions are now in an
16367 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16368 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16372 * MIPS III/IV assembler module is reimplemented.
16376 * More DES library cleanups: remove references to srand/rand and
16377 delete an unused file.
16381 * Add support for the free Netwide assembler (NASM) under Win32,
16382 since not many people have MASM (ml) and it can be hard to obtain.
16383 This is currently experimental but it seems to work OK and pass all
16384 the tests. Check out INSTALL.W32 for info.
16388 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16389 without temporary keys kept an extra copy of the server key,
16390 and connections with temporary keys did not free everything in case
16395 * New function RSA_check_key and new openssl rsa option -check
16396 for verifying the consistency of RSA keys.
16398 *Ulf Moeller, Bodo Moeller*
16400 * Various changes to make Win32 compile work:
16401 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16402 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16403 comparison" warnings.
16404 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
16408 * Add a debugging option to PKCS#5 v2 key generation function: when
16409 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16410 derived keys are printed to stderr.
16414 * Copy the flags in ASN1_STRING_dup().
16416 *Roman E. Pavlov <pre@mo.msk.ru>*
16418 * The x509 application mishandled signing requests containing DSA
16419 keys when the signing key was also DSA and the parameters didn't match.
16421 It was supposed to omit the parameters when they matched the signing key:
16422 the verifying software was then supposed to automatically use the CA's
16423 parameters if they were absent from the end user certificate.
16425 Omitting parameters is no longer recommended. The test was also
16426 the wrong way round! This was probably due to unusual behaviour in
16427 EVP_cmp_parameters() which returns 1 if the parameters match.
16428 This meant that parameters were omitted when they *didn't* match and
16429 the certificate was useless. Certificates signed with 'ca' didn't have
16432 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16434 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16435 The interface is as follows:
16436 Applications can use
16437 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16438 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16439 "off" is now the default.
16440 The library internally uses
16441 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16442 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16443 to disable memory-checking temporarily.
16445 Some inconsistent states that previously were possible (and were
16446 even the default) are now avoided.
16448 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16449 with each memory chunk allocated; this is occasionally more helpful
16450 than just having a counter.
16452 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16454 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16459 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16460 which largely parallels "options", but is for changing API behaviour,
16461 whereas "options" are about protocol behaviour.
16462 Initial "mode" flags are:
16464 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16465 a single record has been written.
16466 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16467 retries use the same buffer location.
16468 (But all of the contents must be
16473 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16476 * Fix problems with no-hmac etc.
16478 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16480 * New functions RSA_get_default_method(), RSA_set_method() and
16481 RSA_get_method(). These allows replacement of RSA_METHODs without having
16482 to mess around with the internals of an RSA structure.
16486 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16487 Also really enable memory leak checks in openssl.c and in some
16490 *Chad C. Mulligan, Bodo Moeller*
16492 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16493 up the length of negative integers. This has now been simplified to just
16494 store the length when it is first determined and use it later, rather
16495 than trying to keep track of where data is copied and updating it to
16497 *Steve Henson, reported by Brien Wheeler
16498 <bwheeler@authentica-security.com>*
16500 * Add a new function PKCS7_signatureVerify. This allows the verification
16501 of a PKCS#7 signature but with the signing certificate passed to the
16502 function itself. This contrasts with PKCS7_dataVerify which assumes the
16503 certificate is present in the PKCS#7 structure. This isn't always the
16504 case: certificates can be omitted from a PKCS#7 structure and be
16505 distributed by "out of band" means (such as a certificate database).
16509 * Complete the PEM_* macros with DECLARE_PEM versions to replace the
16510 function prototypes in pem.h, also change util/mkdef.pl to add the
16511 necessary function names.
16515 * mk1mf.pl (used by Windows builds) did not properly read the
16516 options set by Configure in the top level Makefile, and Configure
16517 was not even able to write more than one option correctly.
16518 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16522 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16523 file to be loaded from a BIO or FILE pointer. The BIO version will
16524 for example allow memory BIOs to contain config info.
16528 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16529 Whoever hopes to achieve shared-library compatibility across versions
16530 must use this, not the compile-time macro.
16531 (Exercise 0.9.4: Which is the minimum library version required by
16533 Note: All this applies only to multi-threaded programs, others don't
16538 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16539 through a BIO pair triggered the default case, i.e.
16540 SSLerr(...,SSL_R_UNKNOWN_STATE).
16544 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16545 can use the SSL library even if none of the specific BIOs is
16550 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16551 for the encoded length.
16553 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16555 * Add initial documentation of the X509V3 functions.
16559 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16560 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16561 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16562 secure PKCS#8 private key format with a high iteration count.
16566 * Fix determination of Perl interpreter: A perl or perl5
16567 _directory_ in $PATH was also accepted as the interpreter.
16569 *Ralf S. Engelschall*
16571 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16572 wrong with it but it was very old and did things like calling
16573 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16574 unusual formatting.
16578 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16579 to use the new extension code.
16583 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16584 with macros. This should make it easier to change their form, add extra
16585 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16590 * Add to configuration table a new entry that can specify an alternative
16591 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16592 according to Mark Crispin <MRC@Panda.COM>.
16597 * DES CBC did not update the IV. Weird.
16601 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16602 Changing the behaviour of the former might break existing programs --
16603 where IV updating is needed, des_ncbc_encrypt can be used.
16606 * When bntest is run from "make test" it drives bc to check its
16607 calculations, as well as internally checking them. If an internal check
16608 fails, it needs to cause bc to give a non-zero result or make test carries
16609 on without noticing the failure. Fixed.
16613 * DES library cleanups.
16617 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16618 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16619 ciphers. NOTE: although the key derivation function has been verified
16620 against some published test vectors it has not been extensively tested
16621 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16626 * Instead of "mkdir -p", which is not fully portable, use new
16627 Perl script "util/mkdir-p.pl".
16631 * Rewrite the way password based encryption (PBE) is handled. It used to
16632 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16633 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16634 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16635 the 'parameter' field of the AlgorithmIdentifier is passed to the
16636 underlying key generation function so it must do its own ASN1 parsing.
16637 This has also changed the EVP_PBE_CipherInit() function which now has a
16638 'parameter' argument instead of literal salt and iteration count values
16639 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16643 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16644 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16645 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16646 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16647 value was just used as a "magic string" and not used directly its
16648 value doesn't matter.
16652 * Introduce some semblance of const correctness to BN. Shame C doesn't
16657 * "linux-sparc64" configuration (ultrapenguin).
16659 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16660 "linux-sparc" configuration.
16662 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16664 * config now generates no-xxx options for missing ciphers.
16668 * Support the EBCDIC character set (work in progress).
16669 File ebcdic.c not yet included because it has a different license.
16671 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16673 * Support BS2000/OSD-POSIX.
16675 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16677 * Make callbacks for key generation use void * instead of char *.
16681 * Make S/MIME samples compile (not yet tested).
16685 * Additional typesafe stacks.
16689 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16694 ### Changes between 0.9.3 and 0.9.3a [29 May 1999] ###
16696 * New configuration variant "sco5-gcc".
16698 * Updated some demos.
16700 *Sean O Riordain, Wade Scholine*
16702 * Add missing BIO_free at exit of pkcs12 application.
16706 * Fix memory leak in conf.c.
16710 * Updates for Win32 to assembler version of MD5.
16714 * Set #! path to perl in apps/der_chop to where we found it
16715 instead of using a fixed path.
16719 * SHA library changes for irix64-mips4-cc.
16723 * Improvements for VMS support.
16728 ### Changes between 0.9.2b and 0.9.3 [24 May 1999] ###
16730 * Bignum library bug fix. IRIX 6 passes "make test" now!
16731 This also avoids the problems with SC4.2 and unpatched SC5.
16733 *Andy Polyakov <appro@fy.chalmers.se>*
16735 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16736 These are required because of the typesafe stack would otherwise break
16737 existing code. If old code used a structure member which used to be STACK
16738 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16739 sk_num or sk_value it would produce an error because the num, data members
16740 are not present in STACK_OF. Now it just produces a warning. sk_set
16741 replaces the old method of assigning a value to sk_value
16742 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16743 that does this will no longer work (and should use sk_set instead) but
16744 this could be regarded as a "questionable" behaviour anyway.
16748 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16749 correctly handle encrypted S/MIME data.
16753 * Change type of various DES function arguments from des_cblock
16754 (which means, in function argument declarations, pointer to char)
16755 to des_cblock * (meaning pointer to array with 8 char elements),
16756 which allows the compiler to do more typechecking; it was like
16757 that back in SSLeay, but with lots of ugly casts.
16759 Introduce new type const_des_cblock.
16763 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16764 problems: find RecipientInfo structure that matches recipient certificate
16765 and initialise the ASN1 structures properly based on passed cipher.
16769 * Belatedly make the BN tests actually check the results.
16773 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16774 to and from BNs: it was completely broken. New compilation option
16775 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16776 key elements as negative integers.
16780 * Reorganize and speed up MD5.
16782 *Andy Polyakov <appro@fy.chalmers.se>*
16786 *Richard Levitte <richard@levitte.org>*
16788 * New option -out to asn1parse to allow the parsed structure to be
16789 output to a file. This is most useful when combined with the -strparse
16790 option to examine the output of things like OCTET STRINGS.
16794 * Make SSL library a little more fool-proof by not requiring any longer
16795 that SSL_set_{accept,connect}_state be called before
16796 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
16797 in many applications because usually everything *appeared* to work as
16798 intended anyway -- now it really works as intended).
16802 * Move openssl.cnf out of lib/.
16806 * Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
16807 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16808 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
16810 *Ralf S. Engelschall*
16812 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16813 handle PKCS#7 enveloped data properly.
16815 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16817 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16818 copying pointers. The cert_st handling is changed by this in
16819 various ways (and thus what used to be known as ctx->default_cert
16820 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
16821 any longer when s->cert does not give us what we need).
16822 ssl_cert_instantiate becomes obsolete by this change.
16823 As soon as we've got the new code right (possibly it already is?),
16824 we have solved a couple of bugs of the earlier code where s->cert
16825 was used as if it could not have been shared with other SSL structures.
16827 Note that using the SSL API in certain dirty ways now will result
16828 in different behaviour than observed with earlier library versions:
16829 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
16830 does not influence s as it used to.
16832 In order to clean up things more thoroughly, inside SSL_SESSION
16833 we don't use CERT any longer, but a new structure SESS_CERT
16834 that holds per-session data (if available); currently, this is
16835 the peer's certificate chain and, for clients, the server's certificate
16836 and temporary key. CERT holds only those values that can have
16837 meaningful defaults in an SSL_CTX.
16841 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16842 from the internal representation. Various PKCS#7 fixes: remove some
16843 evil casts and set the enc_dig_alg field properly based on the signing
16848 * Allow PKCS#12 password to be set from the command line or the
16849 environment. Let 'ca' get its config file name from the environment
16850 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16855 * Allow certificate policies extension to use an IA5STRING for the
16856 organization field. This is contrary to the PKIX definition but
16857 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16862 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16863 without disallowing inline assembler and the like for non-pedantic builds.
16867 * Support Borland C++ builder.
16869 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16875 * SHA-1 cleanups and performance enhancements.
16877 *Andy Polyakov <appro@fy.chalmers.se>*
16879 * Sparc v8plus assembler for the bignum library.
16881 *Andy Polyakov <appro@fy.chalmers.se>*
16883 * Accept any -xxx and +xxx compiler options in Configure.
16887 * Update HPUX configuration.
16891 * Add missing sk_<type>_unshift() function to safestack.h
16893 *Ralf S. Engelschall*
16895 * New function SSL_CTX_use_certificate_chain_file that sets the
16896 "extra_cert"s in addition to the certificate. (This makes sense
16897 only for "PEM" format files, as chains as a whole are not
16902 * Support verify_depth from the SSL API.
16903 x509_vfy.c had what can be considered an off-by-one-error:
16904 Its depth (which was not part of the external interface)
16905 was actually counting the number of certificates in a chain;
16906 now it really counts the depth.
16910 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16911 instead of X509err, which often resulted in confusing error
16912 messages since the error codes are not globally unique
16913 (e.g. an alleged error in ssl3_accept when a certificate
16914 didn't match the private key).
16916 * New function SSL_CTX_set_session_id_context that allows to set a default
16917 value (so that you don't need SSL_set_session_id_context for each
16918 connection using the SSL_CTX).
16922 * OAEP decoding bug fix.
16926 * Support INSTALL_PREFIX for package builders, as proposed by
16931 * New Configure options "threads" and "no-threads". For systems
16932 where the proper compiler options are known (currently Solaris
16933 and Linux), "threads" is the default.
16937 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
16941 * Install various scripts to $(OPENSSLDIR)/misc, not to
16942 $(INSTALLTOP)/bin -- they shouldn't clutter directories
16943 such as /usr/local/bin.
16947 * "make linux-shared" to build shared libraries.
16949 *Niels Poppe <niels@netbox.org>*
16951 * New Configure option no-<cipher> (rsa, idea, rc5, ...).
16955 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
16956 extension adding in x509 utility.
16960 * Remove NOPROTO sections and error code comments.
16964 * Partial rewrite of the DEF file generator to now parse the ANSI
16969 * New Configure options --prefix=DIR and --openssldir=DIR.
16973 * Complete rewrite of the error code script(s). It is all now handled
16974 by one script at the top level which handles error code gathering,
16975 header rewriting and C source file generation. It should be much better
16976 than the old method: it now uses a modified version of Ulf's parser to
16977 read the ANSI prototypes in all header files (thus the old K&R definitions
16978 aren't needed for error creation any more) and do a better job of
16979 translating function codes into names. The old 'ASN1 error code embedded
16980 in a comment' is no longer necessary and it doesn't use .err files which
16981 have now been deleted. Also the error code call doesn't have to appear all
16982 on one line (which resulted in some large lines...).
16986 * Change #include filenames from <foo.h> to <openssl/foo.h>.
16990 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
16991 0 (which usually indicates a closed connection), but continue reading.
16995 * Fix some race conditions.
16999 * Add support for CRL distribution points extension. Add Certificate
17000 Policies and CRL distribution points documentation.
17004 * Move the autogenerated header file parts to crypto/opensslconf.h.
17008 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17009 8 of keying material. Merlin has also confirmed interop with this fix
17010 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17012 *Merlin Hughes <merlin@baltimore.ie>*
17014 * Fix lots of warnings.
17016 *Richard Levitte <levitte@stacken.kth.se>*
17018 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17019 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17021 *Richard Levitte <levitte@stacken.kth.se>*
17023 * Fix problems with sizeof(long) == 8.
17025 *Andy Polyakov <appro@fy.chalmers.se>*
17027 * Change functions to ANSI C.
17031 * Fix typos in error codes.
17033 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17035 * Remove defunct assembler files from Configure.
17039 * SPARC v8 assembler BIGNUM implementation.
17041 *Andy Polyakov <appro@fy.chalmers.se>*
17043 * Support for Certificate Policies extension: both print and set.
17044 Various additions to support the r2i method this uses.
17048 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17049 return a const string when you are expecting an allocated buffer.
17053 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17054 types DirectoryString and DisplayText.
17058 * Add code to allow r2i extensions to access the configuration database,
17059 add an LHASH database driver and add several ctx helper functions.
17063 * Fix an evil bug in bn_expand2() which caused various BN functions to
17064 fail when they extended the size of a BIGNUM.
17068 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17069 support typesafe stack.
17073 * Fix typo in SSL_[gs]et_options().
17075 *Nils Frostberg <nils@medcom.se>*
17077 * Delete various functions and files that belonged to the (now obsolete)
17078 old X509V3 handling code.
17082 * New Configure option "rsaref".
17086 * Don't auto-generate pem.h.
17090 * Introduce type-safe ASN.1 SETs.
17094 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17096 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17098 * Introduce type-safe STACKs. This will almost certainly break lots of code
17099 that links with OpenSSL (well at least cause lots of warnings), but fear
17100 not: the conversion is trivial, and it eliminates loads of evil casts. A
17101 few STACKed things have been converted already. Feel free to convert more.
17102 In the fullness of time, I'll do away with the STACK type altogether.
17106 * Add `openssl ca -revoke <certfile>' facility which revokes a certificate
17107 specified in <certfile> by updating the entry in the index.txt file.
17108 This way one no longer has to edit the index.txt file manually for
17109 revoking a certificate. The -revoke option does the gory details now.
17111 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17113 * Fix `openssl crl -noout -text' combination where `-noout' killed the
17114 `-text' option at all and this way the `-noout -text' combination was
17115 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
17117 *Ralf S. Engelschall*
17119 * Make sure a corresponding plain text error message exists for the
17120 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17121 verify callback function determined that a certificate was revoked.
17123 *Ralf S. Engelschall*
17125 * Bugfix: In test/testenc, don't test "openssl <cipher>" for
17126 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17127 all available ciphers including rc5, which was forgotten until now.
17128 In order to let the testing shell script know which algorithms
17129 are available, a new (up to now undocumented) command
17130 "openssl list-cipher-commands" is used.
17134 * Bugfix: s_client occasionally would sleep in select() when
17135 it should have checked SSL_pending() first.
17139 * New functions DSA_do_sign and DSA_do_verify to provide access to
17140 the raw DSA values prior to ASN.1 encoding.
17144 * Tweaks to Configure
17146 *Niels Poppe <niels@netbox.org>*
17148 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17153 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17157 * New config option to avoid instructions that are illegal on the 80386.
17158 The default code is faster, but requires at least a 486.
17162 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17163 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17164 same as SSL2_VERSION anyway.
17168 * New "-showcerts" option for s_client.
17172 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17173 application. Various cleanups and fixes.
17177 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17178 modify error routines to work internally. Add error codes and PBE init
17179 to library startup routines.
17183 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17184 packing functions to asn1 and evp. Changed function names and error
17185 codes along the way.
17189 * PKCS12 integration: and so it begins... First of several patches to
17190 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17191 objects to objects.h
17195 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17196 and display support for Thawte strong extranet extension.
17200 * Add LinuxPPC support.
17202 *Jeff Dubrule <igor@pobox.org>*
17204 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17205 bn_div_words in alpha.s.
17207 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17209 * Make sure the RSA OAEP test is skipped under -DRSAref because
17210 OAEP isn't supported when OpenSSL is built with RSAref.
17212 *Ulf Moeller <ulf@fitug.de>*
17214 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17215 so they no longer are missing under -DNOPROTO.
17217 *Soren S. Jorvang <soren@t.dk>*
17220 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999] ###
17222 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17223 doesn't work when the session is reused. Coming soon!
17227 * Fix a security hole, that allows sessions to be reused in the wrong
17228 context thus bypassing client cert protection! All software that uses
17229 client certs and session caches in multiple contexts NEEDS PATCHING to
17230 allow session reuse! A fuller solution is in the works.
17232 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17234 * Some more source tree cleanups (removed obsolete files
17235 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17236 permission on "config" script to be executable) and a fix for the INSTALL
17239 *Ulf Moeller <ulf@fitug.de>*
17241 * Remove some legacy and erroneous uses of malloc, free instead of
17244 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17246 * Make rsa_oaep_test return non-zero on error.
17248 *Ulf Moeller <ulf@fitug.de>*
17250 * Add support for native Solaris shared libraries. Configure
17251 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17252 if someone would make that last step automatic.
17254 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17256 * ctx_size was not built with the right compiler during "make links". Fixed.
17260 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17261 except NULL ciphers". This means the default cipher list will no longer
17262 enable NULL ciphers. They need to be specifically enabled e.g. with
17263 the string "DEFAULT:eNULL".
17267 * Fix to RSA private encryption routines: if p < q then it would
17268 occasionally produce an invalid result. This will only happen with
17269 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17273 * Be less restrictive and allow also `perl util/perlpath.pl
17274 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17275 because this way one can also use an interpreter named `perl5' (which is
17276 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17277 installed as `perl').
17279 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17281 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17283 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17285 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17286 advapi32.lib to Win32 build and change the pem test comparison
17287 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17288 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17289 and crypto/des/ede_cbcm_enc.c.
17293 * DES quad checksum was broken on big-endian architectures. Fixed.
17297 * Comment out two functions in bio.h that aren't implemented. Fix up the
17298 Win32 test batch file so it (might) work again. The Win32 test batch file
17299 is horrible: I feel ill....
17303 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17304 in e_os.h. Audit of header files to check ANSI and non ANSI
17305 sections: 10 functions were absent from non ANSI section and not exported
17306 from Windows DLLs. Fixed up libeay.num for new functions.
17310 * Make `openssl version' output lines consistent.
17312 *Ralf S. Engelschall*
17314 * Fix Win32 symbol export lists for BIO functions: Added
17315 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17316 to ms/libeay{16,32}.def.
17318 *Ralf S. Engelschall*
17320 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17321 fine under Unix and passes some trivial tests I've now added. But the
17322 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17323 added to make sure no one expects that this stuff really works in the
17324 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17325 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17328 *Ralf S. Engelschall*
17330 * Fix the generation of two part addresses in perl.
17332 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17334 * Add config entry for Linux on MIPS.
17336 *John Tobey <jtobey@channel1.com>*
17338 * Make links whenever Configure is run, unless we are on Windoze.
17342 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17343 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17348 * Add a useful kludge to allow package maintainers to specify compiler and
17349 other platforms details on the command line without having to patch the
17350 Configure script every time: One now can use ``perl Configure
17351 <id>:<details>'', i.e. platform ids are allowed to have details appended
17352 to them (separated by colons). This is treated as there would be a static
17353 pre-configured entry in Configure's %table under key <id> with value
17354 <details> and ``perl Configure <id>'' is called. So, when you want to
17355 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17356 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
17357 now, which overrides the FreeBSD-elf entry on-the-fly.
17359 *Ralf S. Engelschall*
17361 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17365 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17366 on the `perl Configure ...' command line. This way one can compile
17367 OpenSSL libraries with Position Independent Code (PIC) which is needed
17368 for linking it into DSOs.
17370 *Ralf S. Engelschall*
17372 * Remarkably, export ciphers were totally broken and no-one had noticed!
17377 * Cleaned up the LICENSE document: The official contact for any license
17378 questions now is the OpenSSL core team under openssl-core@openssl.org.
17379 And add a paragraph about the dual-license situation to make sure people
17380 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17381 to the OpenSSL toolkit.
17383 *Ralf S. Engelschall*
17385 * General source tree makefile cleanups: Made `making xxx in yyy...'
17386 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17387 Additionally cleaned up the `make links' target: Remove unnecessary
17388 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17389 to speed processing and no longer clutter the display with confusing
17390 stuff. Instead only the actually done links are displayed.
17392 *Ralf S. Engelschall*
17394 * Permit null encryption ciphersuites, used for authentication only. It used
17395 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17396 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17401 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17402 signed attributes when verifying signatures (this would break them),
17403 the detached data encoding was wrong and public keys obtained using
17404 X509_get_pubkey() weren't freed.
17408 * Add text documentation for the BUFFER functions. Also added a work around
17409 to a Win95 console bug. This was triggered by the password read stuff: the
17410 last character typed gets carried over to the next fread(). If you were
17411 generating a new cert request using 'req' for example then the last
17412 character of the passphrase would be CR which would then enter the first
17417 * Added the new `Includes OpenSSL Cryptography Software' button as
17418 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17419 button and can be used by applications based on OpenSSL to show the
17420 relationship to the OpenSSL project.
17422 *Ralf S. Engelschall*
17424 * Remove confusing variables in function signatures in files
17425 ssl/ssl_lib.c and ssl/ssl.h.
17427 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17429 * Don't install bss_file.c under PREFIX/include/
17431 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17433 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17434 functions that return function pointers and has support for NT specific
17435 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17436 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17437 unsigned to signed types: this was killing the Win32 compile.
17441 * Add new certificate file to stack functions,
17442 SSL_add_dir_cert_subjects_to_stack() and
17443 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17444 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17445 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17446 This means that Apache-SSL and similar packages don't have to mess around
17447 to add as many CAs as they want to the preferred list.
17451 * Experiment with doxygen documentation. Currently only partially applied to
17453 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
17454 openssl.doxy as the configuration file.
17458 * Get rid of remaining C++-style comments which strict C compilers hate.
17460 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17462 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17463 compiled in by default: it has problems with large keys.
17467 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17468 DH private keys and/or callback functions which directly correspond to
17469 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17470 is needed for applications which have to configure certificates on a
17471 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17473 For the RSA certificate situation is makes no difference, but
17474 for the DSA certificate situation this fixes the "no shared cipher"
17475 problem where the OpenSSL cipher selection procedure failed because the
17476 temporary keys were not overtaken from the context and the API provided
17477 no way to reconfigure them.
17478 The new functions now let applications reconfigure the stuff and they
17479 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17480 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17481 non-public-API function ssl_cert_instantiate() is used as a helper
17482 function and also to reduce code redundancy inside ssl_rsa.c.
17484 *Ralf S. Engelschall*
17486 * Move s_server -dcert and -dkey options out of the undocumented feature
17487 area because they are useful for the DSA situation and should be
17488 recognized by the users.
17490 *Ralf S. Engelschall*
17492 * Fix the cipher decision scheme for export ciphers: the export bits are
17493 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17494 SSL_EXP_MASK. So, the original variable has to be used instead of the
17495 already masked variable.
17497 *Richard Levitte <levitte@stacken.kth.se>*
17499 * Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
17501 *Richard Levitte <levitte@stacken.kth.se>*
17503 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17504 from `int' to `unsigned int' because it's a length and initialized by
17505 EVP_DigestFinal() which expects an `unsigned int *'.
17507 *Richard Levitte <levitte@stacken.kth.se>*
17509 * Don't hard-code path to Perl interpreter on shebang line of Configure
17510 script. Instead use the usual Shell->Perl transition trick.
17512 *Ralf S. Engelschall*
17514 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17515 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17516 -noout -modulus' as it's already the case for `openssl rsa -noout
17517 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17518 currently the public key is printed (a decision which was already done by
17519 `openssl dsa -modulus' in the past) which serves a similar purpose.
17520 Additionally the NO_RSA no longer completely removes the whole -modulus
17521 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17524 *Ralf S. Engelschall*
17526 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17527 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17529 *Arne Ansper <arne@ats.cyber.ee>*
17531 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17532 to be added. Now both 'req' and 'ca' can use new objects defined in the
17537 * Add cool BIO that does syslog (or event log on NT).
17539 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17541 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17542 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17543 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17544 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17548 * Add preliminary config info for new extension code.
17552 * Make RSA_NO_PADDING really use no padding.
17554 *Ulf Moeller <ulf@fitug.de>*
17556 * Generate errors when private/public key check is done.
17560 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17561 for some CRL extensions and new objects added.
17565 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17566 key usage extension and fuller support for authority key id.
17570 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17571 padding method for RSA, which is recommended for new applications in PKCS
17572 #1 v2.0 (RFC 2437, October 1998).
17573 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17574 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17575 against Bleichbacher's attack on RSA.
17576 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17579 * Updates to the new SSL compression code
17581 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17583 * Fix so that the version number in the master secret, when passed
17584 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17585 (because the server will not accept higher), that the version number
17586 is 0x03,0x01, not 0x03,0x00
17588 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17590 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17591 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17592 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17596 * Support for RAW extensions where an arbitrary extension can be
17597 created by including its DER encoding. See apps/openssl.cnf for
17602 * Make sure latest Perl versions don't interpret some generated C array
17603 code as Perl array code in the crypto/err/err_genc.pl script.
17605 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17607 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17608 not many people have the assembler. Various Win32 compilation fixes and
17609 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17610 build instructions.
17614 * Modify configure script 'Configure' to automatically create crypto/date.h
17615 file under Win32 and also build pem.h from pem.org. New script
17616 util/mkfiles.pl to create the MINFO file on environments that can't do a
17617 'make files': perl util/mkfiles.pl >MINFO should work.
17621 * Major rework of DES function declarations, in the pursuit of correctness
17622 and purity. As a result, many evil casts evaporated, and some weirdness,
17623 too. You may find this causes warnings in your code. Zapping your evil
17624 casts will probably fix them. Mostly.
17628 * Fix for a typo in asn1.h. Bug fix to object creation script
17629 obj_dat.pl. It considered a zero in an object definition to mean
17630 "end of object": none of the objects in objects.h have any zeros
17631 so it wasn't spotted.
17633 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17635 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17636 Masking (CBCM). In the absence of test vectors, the best I have been able
17637 to do is check that the decrypt undoes the encrypt, so far. Send me test
17638 vectors if you have them.
17642 * Correct calculation of key length for export ciphers (too much space was
17643 allocated for null ciphers). This has not been tested!
17647 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17648 message is now correct (it understands "crypto" and "ssl" on its
17649 command line). There is also now an "update" option. This will update
17650 the util/ssleay.num and util/libeay.num files with any new functions.
17652 perl util/mkdef.pl crypto ssl update
17653 it will update them.
17657 * Overhauled the Perl interface (perl/*):
17658 - ported BN stuff to OpenSSL's different BN library
17659 - made the perl/ source tree CVS-aware
17660 - renamed the package from SSLeay to OpenSSL (the files still contain
17661 their history because I've copied them in the repository)
17662 - removed obsolete files (the test scripts will be replaced
17663 by better Test::Harness variants in the future)
17665 *Ralf S. Engelschall*
17667 * First cut for a very conservative source tree cleanup:
17668 1. merge various obsolete readme texts into doc/ssleay.txt
17669 where we collect the old documents and readme texts.
17670 2. remove the first part of files where I'm already sure that we no
17671 longer need them because of three reasons: either they are just temporary
17672 files which were left by Eric or they are preserved original files where
17673 I've verified that the diff is also available in the CVS via "cvs diff
17674 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17675 the crypto/md/ stuff).
17677 *Ralf S. Engelschall*
17679 * More extension code. Incomplete support for subject and issuer alt
17680 name, issuer and authority key id. Change the i2v function parameters
17681 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17682 what that's for :-) Fix to ASN1 macro which messed up
17683 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17687 * Preliminary support for ENUMERATED type. This is largely copied from the
17692 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17694 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17696 * Make sure `make rehash' target really finds the `openssl' program.
17698 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17700 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17701 like to hear about it if this slows down other processors.
17705 * Add CygWin32 platform information to Configure script.
17707 *Alan Batie <batie@aahz.jf.intel.com>*
17709 * Fixed ms/32all.bat script: `no_asm' -> `no-asm'
17711 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17713 * New program nseq to manipulate netscape certificate sequences
17717 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17722 * Fixes to BN code. Previously the default was to define BN_RECURSION
17723 but the BN code had some problems that would cause failures when
17724 doing certificate verification and some other functions.
17726 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17728 * Add ASN1 and PEM code to support netscape certificate sequences.
17732 * Add ASN1 and PEM code to support netscape certificate sequences.
17736 * Add several PKIX and private extended key usage OIDs.
17740 * Modify the 'ca' program to handle the new extension code. Modify
17741 openssl.cnf for new extension format, add comments.
17745 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17746 and add a sample to openssl.cnf so req -x509 now adds appropriate
17751 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17752 error code, add initial support to X509_print() and x509 application.
17756 * Takes a deep breath and start adding X509 V3 extension support code. Add
17757 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17758 stuff is currently isolated and isn't even compiled yet.
17762 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17763 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17764 Removed the versions check from X509 routines when loading extensions:
17765 this allows certain broken certificates that don't set the version
17766 properly to be processed.
17770 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17771 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17772 can still be regenerated with "make depend".
17776 * Spelling mistake in C version of CAST-128.
17778 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17780 * Changes to the error generation code. The perl script err-code.pl
17781 now reads in the old error codes and retains the old numbers, only
17782 adding new ones if necessary. It also only changes the .err files if new
17783 codes are added. The makefiles have been modified to only insert errors
17784 when needed (to avoid needlessly modifying header files). This is done
17785 by only inserting errors if the .err file is newer than the auto generated
17786 C file. To rebuild all the error codes from scratch (the old behaviour)
17787 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17788 or delete all the .err files.
17792 * CAST-128 was incorrectly implemented for short keys. The C version has
17793 been fixed, but is untested. The assembler versions are also fixed, but
17794 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17795 to regenerate it if needed.
17796 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17797 Hagino <itojun@kame.net>*
17799 * File was opened incorrectly in randfile.c.
17801 *Ulf Möller <ulf@fitug.de>*
17803 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17804 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17805 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17806 al: it's just almost always a UTCTime. Note this patch adds new error
17807 codes so do a "make errors" if there are problems.
17811 * Correct Linux 1 recognition in config.
17813 *Ulf Möller <ulf@fitug.de>*
17815 * Remove pointless MD5 hash when using DSA keys in ca.
17817 *Anonymous <nobody@replay.com>*
17819 * Generate an error if given an empty string as a cert directory. Also
17820 generate an error if handed NULL (previously returned 0 to indicate an
17821 error, but didn't set one).
17823 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17825 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17829 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17830 parameters. This was causing a warning which killed off the Win32 compile.
17834 * Remove C++ style comments from crypto/bn/bn_local.h.
17836 *Neil Costigan <neil.costigan@celocom.com>*
17838 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17839 based on a text string, looking up short and long names and finally
17840 "dot" format. The "dot" format stuff didn't work. Added new function
17841 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17842 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17843 OID is not part of the table.
17847 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17848 X509_LOOKUP_by_alias().
17852 * Sort openssl functions by name.
17856 * Get the gendsa program working (hopefully) and add it to app list. Remove
17857 encryption from sample DSA keys (in case anyone is interested the password
17862 * Make _all_ *_free functions accept a NULL pointer.
17864 *Frans Heymans <fheymans@isaserver.be>*
17866 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17869 *Anonymous <nobody@replay.com>*
17871 * s_server should send the CAfile as acceptable CAs, not its own cert.
17873 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17875 * Don't blow it for numeric -newkey arguments to apps/req.
17877 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17879 * Temp key "for export" tests were wrong in s3_srvr.c.
17881 *Anonymous <nobody@replay.com>*
17883 * Add prototype for temp key callback functions
17884 SSL_CTX_set_tmp_{rsa,dh}_callback().
17888 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17889 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17893 * X509_name_add_entry() freed the wrong thing after an error.
17895 *Arne Ansper <arne@ats.cyber.ee>*
17897 * rsa_eay.c would attempt to free a NULL context.
17899 *Arne Ansper <arne@ats.cyber.ee>*
17901 * BIO_s_socket() had a broken should_retry() on Windoze.
17903 *Arne Ansper <arne@ats.cyber.ee>*
17905 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17907 *Arne Ansper <arne@ats.cyber.ee>*
17909 * Make sure the already existing X509_STORE->depth variable is initialized
17910 in X509_STORE_new(), but document the fact that this variable is still
17911 unused in the certificate verification process.
17913 *Ralf S. Engelschall*
17915 * Fix the various library and apps files to free up pkeys obtained from
17916 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17920 * Fix reference counting in X509_PUBKEY_get(). This makes
17921 demos/maurice/example2.c work, amongst others, probably.
17923 *Steve Henson and Ben Laurie*
17925 * First cut of a cleanup for apps/. First the `ssleay' program is now named
17926 `openssl' and second, the shortcut symlinks for the `openssl <command>'
17927 are no longer created. This way we have a single and consistent command
17928 line interface `openssl <command>', similar to `cvs <command>'.
17930 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
17932 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
17933 BIT STRING wrapper always have zero unused bits.
17937 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
17941 * Make the top-level INSTALL documentation easier to understand.
17945 * Makefiles updated to exit if an error occurs in a sub-directory
17946 make (including if user presses ^C) [Paul Sutton]
17948 * Make Montgomery context stuff explicit in RSA data structure.
17952 * Fix build order of pem and err to allow for generated pem.h.
17956 * Fix renumbering bug in X509_NAME_delete_entry().
17960 * Enhanced the err-ins.pl script so it makes the error library number
17961 global and can add a library name. This is needed for external ASN1 and
17962 other error libraries.
17966 * Fixed sk_insert which never worked properly.
17970 * Fix ASN1 macros so they can handle indefinite length constructed
17971 EXPLICIT tags. Some non standard certificates use these: they can now
17976 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
17977 into a single doc/ssleay.txt bundle. This way the information is still
17978 preserved but no longer messes up this directory. Now it's new room for
17979 the new set of documentation files.
17981 *Ralf S. Engelschall*
17983 * SETs were incorrectly DER encoded. This was a major pain, because they
17984 shared code with SEQUENCEs, which aren't coded the same. This means that
17985 almost everything to do with SETs or SEQUENCEs has either changed name or
17986 number of arguments.
17988 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
17990 * Fix test data to work with the above.
17994 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
17995 was already fixed by Eric for 0.9.1 it seems.
17997 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
17999 * Autodetect FreeBSD3.
18003 * Fix various bugs in Configure. This affects the following platforms:
18007 unixware-2.0-pentium
18012 * Eliminate generated files from CVS. Reorder tests to regenerate files
18013 before they are needed.
18017 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18022 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998] ###
18024 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18025 changed SSLeay to OpenSSL in version strings.
18027 *Ralf S. Engelschall*
18029 * Some fixups to the top-level documents.
18033 * Fixed the nasty bug where rsaref.h was not found under compile-time
18034 because the symlink to include/ was missing.
18036 *Ralf S. Engelschall*
18038 * Incorporated the popular no-RSA/DSA-only patches
18039 which allow to compile a RSA-free SSLeay.
18041 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18043 * Fixed nasty rehash problem under `make -f Makefile.ssl links'
18044 when "ssleay" is still not found.
18046 *Ralf S. Engelschall*
18048 * Added more platforms to Configure: Cray T3E, HPUX 11,
18050 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18052 * Updated the README file.
18054 *Ralf S. Engelschall*
18056 * Added various .cvsignore files in the CVS repository subdirs
18057 to make a "cvs update" really silent.
18059 *Ralf S. Engelschall*
18061 * Recompiled the error-definition header files and added
18062 missing symbols to the Win32 linker tables.
18064 *Ralf S. Engelschall*
18066 * Cleaned up the top-level documents;
18067 o new files: CHANGES and LICENSE
18068 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18069 o merged COPYRIGHT into LICENSE
18070 o removed obsolete TODO file
18071 o renamed MICROSOFT to INSTALL.W32
18073 *Ralf S. Engelschall*
18075 * Removed dummy files from the 0.9.1b source tree:
18076 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18077 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18078 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18079 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18080 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18082 *Ralf S. Engelschall*
18084 * Added various platform portability fixes.
18088 * The Genesis of the OpenSSL rpject:
18089 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18090 Young and Tim J. Hudson created while they were working for C2Net until
18093 *The OpenSSL Project*
18096 ### Changes between 0.9.0b and 0.9.1b [not released] ###
18098 * Updated a few CA certificates under certs/
18102 * Changed some BIGNUM api stuff.
18106 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18107 DGUX x86, Linux Alpha, etc.
18111 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18112 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18117 * Add -strparse option to asn1pars program which parses nested
18120 *Dr Stephen Henson <shenson@bigfoot.com>*
18122 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18126 * DSA fix for "ca" program.
18130 * Added "-genkey" option to "dsaparam" program.
18134 * Added RIPE MD160 (rmd160) message digest.
18138 * Added -a (all) option to "ssleay version" command.
18142 * Added PLATFORM define which is the id given to Configure.
18146 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18150 * Extended the ASN.1 parser routines.
18154 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18158 * Added a BN_CTX to the BN library.
18162 * Fixed the weak key values in DES library
18166 * Changed API in EVP library for cipher aliases.
18170 * Added support for RC2/64bit cipher.
18174 * Converted the lhash library to the crypto/mem.c functions.
18178 * Added more recognized ASN.1 object ids.
18182 * Added more RSA padding checks for SSL/TLS.
18186 * Added BIO proxy/filter functionality.
18190 * Added extra_certs to SSL_CTX which can be used
18191 send extra CA certificates to the client in the CA cert chain sending
18192 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18196 * Now Fortezza is denied in the authentication phase because
18197 this is key exchange mechanism is not supported by SSLeay at all.
18201 * Additional PKCS1 checks.
18205 * Support the string "TLSv1" for all TLS v1 ciphers.
18209 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18210 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18214 * Fixed a few memory leaks.
18218 * Fixed various code and comment typos.
18222 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18223 bytes sent in the client random.
18225 *Edward Bishop <ebishop@spyglass.com>*
18230 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18231 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18232 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18233 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18234 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18235 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18236 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18237 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18238 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18239 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18240 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18241 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18242 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18243 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18244 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18245 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18246 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18247 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18248 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18249 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18250 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18251 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18252 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18253 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18254 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18255 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18256 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18257 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18258 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18259 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18260 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18261 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18262 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18263 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18264 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18265 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18266 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18267 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18268 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18269 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18270 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18271 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18272 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18273 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18274 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18275 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18276 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18277 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18278 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18279 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18280 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18281 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18282 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18283 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18284 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18285 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18286 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18287 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18288 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18289 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18290 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18291 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18292 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18293 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18294 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18295 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18296 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18297 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18298 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18299 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18300 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18301 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18302 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18303 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18304 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18305 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18306 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18307 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18308 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18309 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18310 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18311 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18312 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18313 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18314 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18315 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18316 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18317 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18318 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18319 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18320 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18321 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18322 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18323 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18324 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18325 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18326 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18327 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18328 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18329 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18330 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18331 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18332 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18333 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18334 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18335 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18336 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18337 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18338 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18339 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18340 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18341 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18342 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18343 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18344 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18345 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18346 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18347 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18348 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18349 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18350 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18351 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18352 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18353 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18354 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18355 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18356 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18357 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18358 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18359 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18360 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18361 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18362 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18363 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18364 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18365 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18366 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18367 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18368 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18369 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18370 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18371 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18372 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18373 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18374 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18375 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18376 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18377 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18378 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18379 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18380 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18381 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18382 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18383 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18384 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18385 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18386 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18387 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18388 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18389 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655