5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.1.1 and 3.0.0 [xx XXX xxxx]
12 *) X509 certificates signed using SHA1 are no longer allowed at security
14 In TLS/SSL the default security level is 1. It can be set either
15 using the cipher string with @SECLEVEL, or calling
16 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
17 a call to SSL_CTX_use_certificate() will fail if the security level is not
19 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
20 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
24 *) The command line utilities dhparam, dsa, gendsa and dsaparam have been
25 deprecated. Instead use the pkeyparam, pkey, genpkey and pkeyparam
26 programs respectively.
29 *) All of the low level DSA functions have been deprecated including:
31 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
32 DSA_get_default_method, DSA_set_method, DSA_get_method, DSA_new_method,
33 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
34 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
35 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
36 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
37 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
38 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
39 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
40 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
41 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
42 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
43 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
45 Use of these low level functions has been informally discouraged for a long
46 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
47 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
50 *) Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
51 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
52 This means that applications don't have to look at the curve NID and
53 'EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)' to get SM2 computations.
54 However, they still can, that EVP_PKEY_set_alias_type() call acts as
55 a no-op when the EVP_PKEY is already of the given type.
57 Parameter and key generation is also reworked to make it possible
58 to generate EVP_PKEY_SM2 parameters and keys without having to go
59 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
60 However, code that does the latter will still work as before.
63 *) Deprecated low level ECDH and ECDSA functions. These include:
65 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
66 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
69 Use of these low level functions has been informally discouraged for a long
70 time. Instead applications should use the EVP_PKEY_derive(3),
71 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
74 *) Deprecated the EC_KEY_METHOD functions. These include:
76 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
77 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
78 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
79 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
80 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
81 EC_KEY_METHOD_get_verify.
83 Instead applications and extension writers should use the OSSL_PROVIDER
87 *) Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
88 and EVP_PKEY_decrypt() instead.
89 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
90 and EVP_PKEY_encrypt() instead.
93 *) Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
94 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
95 a new formulation to include all the things it can be used for,
96 as well as words of caution.
99 *) The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
100 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
103 *) All of the low level HMAC functions have been deprecated including:
104 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
105 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
107 Use of these low level functions has been informally discouraged for a long
108 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
109 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
110 and L<EVP_MAC_final(3)>.
113 *) All of the low level CMAC functions have been deprecated including:
114 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
115 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
116 Use of these low level functions has been informally discouraged for a long
117 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
118 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
119 and L<EVP_MAC_final(3)>.
122 *) Over two thousand fixes were made to the documentation, including:
123 - Common options (such as -rand/-writerand, TLS version control, etc)
124 were refactored and point to newly-enhanced descriptions in openssl.pod.
125 - Added style conformance for all options (with help from Richard Levitte),
126 documented all reported missing options, added a CI build to check
127 that all options are documented and that no unimplemented options
129 - Documented some internals, such as all use of environment variables.
130 - Addressed all internal broken L<> references.
133 *) All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
134 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
137 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
138 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
139 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
140 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
141 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
142 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
143 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
144 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
145 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
146 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
148 Use of these low level functions has been informally discouraged
149 for a long time. Applications should use the EVP_DigestInit_ex(3),
150 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
153 *) Corrected the documentation of the return values from the EVP_DigestSign*
154 set of functions. The documentation mentioned negative values for some
155 errors, but this was never the case, so the mention of negative values
158 Code that followed the documentation and thereby check with something
159 like 'EVP_DigestSignInit(...) <= 0' will continue to work undisturbed.
162 *) All of the low level cipher functions have been deprecated including:
164 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
165 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
166 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
167 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
168 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
169 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
170 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
171 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
172 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
173 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
174 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
175 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
176 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
177 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
178 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
179 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
180 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
181 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
182 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
183 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
184 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
185 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
186 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
187 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
188 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
189 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
190 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
191 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
192 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
194 Use of these low level functions has been informally discouraged for
195 a long time. Applications should use the high level EVP APIs, e.g.
196 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
197 equivalently named decrypt functions instead.
198 [Matt Caswell and Paul Dale]
200 *) Removed include/openssl/opensslconf.h.in and replaced it with
201 include/openssl/configuration.h.in, which differs in not including
202 <openssl/macros.h>. A short header include/openssl/opensslconf.h
203 was added to include both.
205 This allows internal hacks where one might need to modify the set
206 of configured macros, for example this if deprecated symbols are
207 still supposed to be available internally:
209 #include <openssl/configuration.h>
211 #undef OPENSSL_NO_DEPRECATED
212 #define OPENSSL_SUPPRESS_DEPRECATED
214 #include <openssl/macros.h>
216 This should not be used by applications that use the exported
217 symbols, as that will lead to linking errors.
220 *) Fixed an an overflow bug in the x64_64 Montgomery squaring procedure
221 used in exponentiation with 512-bit moduli. No EC algorithms are
222 affected. Analysis suggests that attacks against 2-prime RSA1024,
223 3-prime RSA1536, and DSA1024 as a result of this defect would be very
224 difficult to perform and are not believed likely. Attacks against DH512
225 are considered just feasible. However, for an attack the target would
226 have to re-use the DH512 private key, which is not recommended anyway.
227 Also applications directly using the low level API BN_mod_exp may be
228 affected if they use BN_FLG_CONSTTIME.
232 *) Most memory-debug features have been deprecated, and the functionality
233 replaced with no-ops.
236 *) Introduced a new method type and API, OSSL_SERIALIZER, to
237 represent generic serializers. An implementation is expected to
238 be able to serialize an object associated with a given name (such
239 as an algorithm name for an asymmetric key) into forms given by
240 implementation properties.
242 Serializers are primarily used from inside libcrypto, through
243 calls to functions like EVP_PKEY_print_private(),
244 PEM_write_bio_PrivateKey() and similar.
246 Serializers are specified in such a way that they can be made to
247 directly handle the provider side portion of an object, if this
248 provider side part comes from the same provider as the serializer
249 itself, but can also be made to handle objects in parametrized
250 form (as an OSSL_PARAM array of data). This allows a provider to
251 offer generic serializers as a service for any other provider.
254 *) Added a .pragma directive to the syntax of configuration files, to
255 allow varying behavior in a supported and predictable manner.
256 Currently added pragma:
260 This allows dollar signs to be a keyword character unless it's
261 followed by a opening brace or parenthesis. This is useful for
262 platforms where dollar signs are commonly used in names, such as
263 volume names and system directory names on VMS.
266 *) Added functionality to create an EVP_PKEY from user data. This
267 is effectively the same as creating a RSA, DH or DSA object and
268 then assigning them to an EVP_PKEY, but directly using algorithm
269 agnostic EVP functions. A benefit is that this should be future
270 proof for public key algorithms to come.
273 *) Change the interpretation of the '--api' configuration option to
274 mean that this is a desired API compatibility level with no
275 further meaning. The previous interpretation, that this would
276 also mean to remove all deprecated symbols up to and including
277 the given version, no requires that 'no-deprecated' is also used
278 in the configuration.
280 When building applications, the desired API compatibility level
281 can be set with the OPENSSL_API_COMPAT macro like before. For
282 API compatibility version below 3.0, the old style numerical
283 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
284 For version 3.0 and on, the value is expected to be the decimal
285 value calculated from the major and minor version like this:
287 MAJOR * 10000 + MINOR * 100
291 -DOPENSSL_API_COMPAT=30000 For 3.0
292 -DOPENSSL_API_COMPAT=30200 For 3.2
294 To hide declarations that are deprecated up to and including the
295 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
296 given when building the application as well.
299 *) Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
300 access to certificate and CRL stores via URIs and OSSL_STORE
303 This adds the following functions:
306 X509_STORE_load_file()
307 X509_STORE_load_path()
308 X509_STORE_load_store()
309 SSL_add_store_cert_subjects_to_stack()
310 SSL_CTX_set_default_verify_store()
311 SSL_CTX_load_verify_file()
312 SSL_CTX_load_verify_dir()
313 SSL_CTX_load_verify_store()
315 Also, the following functions are now deprecated:
317 - X509_STORE_load_locations() (use X509_STORE_load_file(),
318 X509_STORE_load_path() or X509_STORE_load_store() instead)
319 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
320 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
323 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
324 The presence of this system service is determined at run-time.
327 *) Added functionality to create an EVP_PKEY context based on data
328 for methods from providers. This takes an algorithm name and a
329 property query string and simply stores them, with the intent
330 that any operation that uses this context will use those strings
331 to fetch the needed methods implicitly, thereby making the port
332 of application written for pre-3.0 OpenSSL easier.
335 *) The undocumented function NCONF_WIN32() has been deprecated; for
336 conversion details see the HISTORY section of doc/man5/config.pod
339 *) Introduced the new functions EVP_DigestSignInit_ex() and
340 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
341 EVP_DigestVerifyUpdate() have been converted to functions. See the man
342 pages for further details.
345 *) s390x assembly pack: add hardware-support for P-256, P-384, P-521,
346 X25519, X448, Ed25519 and Ed448.
349 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
353 *) Deprecated the public definition of ERR_STATE as well as the function
354 ERR_get_state(). This is done in preparation of making ERR_STATE an
358 *) Added ERR functionality to give callers access to the stored function
359 names that have replaced the older function code based functions.
361 New functions are ERR_get_error_func(), ERR_peek_error_func(),
362 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
363 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
364 and ERR_peek_last_error_all().
366 These functions have become deprecated: ERR_get_error_line_data(),
367 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
368 ERR_func_error_string().
371 *) Extended testing to be verbose for failing tests only. The make variables
372 VERBOSE_FAILURE or VF can be used to enable this:
374 $ make VF=1 test # Unix
375 $ mms /macro=(VF=1) test ! OpenVMS
376 $ nmake VF=1 test # Windows
380 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
381 used even when parsing explicit parameters, when loading a serialized key
382 or calling `EC_GROUP_new_from_ecpkparameters()`/
383 `EC_GROUP_new_from_ecparameters()`.
384 This prevents bypass of security hardening and performance gains,
385 especially for curves with specialized EC_METHODs.
386 By default, if a key encoded with explicit parameters is loaded and later
387 serialized, the output is still encoded with explicit parameters, even if
388 internally a "named" EC_GROUP is used for computation.
391 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
392 this change, EC_GROUP_set_generator would accept order and/or cofactor as
393 NULL. After this change, only the cofactor parameter can be NULL. It also
394 does some minimal sanity checks on the passed order.
398 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
399 An attack is simple, if the first CMS_recipientInfo is valid but the
400 second CMS_recipientInfo is chosen ciphertext. If the second
401 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
402 encryption key will be replaced by garbage, and the message cannot be
403 decoded, but if the RSA decryption fails, the correct encryption key is
404 used and the recipient will not notice the attack.
405 As a work around for this potential attack the length of the decrypted
406 key must be equal to the cipher default key length, in case the
407 certifiate is not given and all recipientInfo are tried out.
408 The old behaviour can be re-enabled in the CMS code by setting the
409 CMS_DEBUG_DECRYPT flag.
412 *) Early start up entropy quality from the DEVRANDOM seed source has been
413 improved for older Linux systems. The RAND subsystem will wait for
414 /dev/random to be producing output before seeding from /dev/urandom.
415 The seeded state is stored for future library initialisations using
416 a system global shared memory segment. The shared memory identifier
417 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
418 the desired value. The default identifier is 114.
421 *) Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
422 when primes for RSA keys are computed.
423 Since we previously always generated primes == 2 (mod 3) for RSA keys,
424 the 2-prime and 3-prime RSA modules were easy to distinguish, since
425 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
426 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
427 This avoids possible fingerprinting of newly generated RSA modules.
430 *) Correct the extended master secret constant on EBCDIC systems. Without this
431 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
432 negotiate EMS will fail. Unfortunately this also means that TLS connections
433 between EBCDIC systems with this fix, and EBCDIC systems without this
434 fix will fail if they negotiate EMS.
437 *) Changed the library initialisation so that the config file is now loaded
438 by default. This was already the case for libssl. It now occurs for both
439 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
440 OPENSSL_init_crypto() to suppress automatic loading of a config file.
443 *) Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
444 where the former acts as a replacement for ERR_put_error(), and the
445 latter replaces the combination ERR_put_error()+ERR_add_error_data().
446 ERR_raise_data() adds more flexibility by taking a format string and
447 an arbitrary number of arguments following it, to be processed with
451 *) Introduced a new function, OSSL_PROVIDER_available(), which can be used
452 to check if a named provider is loaded and available. When called, it
453 will also activate all fallback providers if such are still present.
456 *) Enforce a minimum DH modulus size of 512 bits.
459 *) Changed DH parameters to generate the order q subgroup instead of 2q.
460 Previously generated DH parameters are still accepted by DH_check
461 but DH_generate_key works around that by clearing bit 0 of the
462 private key for those. This avoids leaking bit 0 of the private key.
465 *) Significantly reduce secure memory usage by the randomness pools.
468 *) {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
472 *) A new type, EVP_KEYEXCH, has been introduced to represent key exchange
473 algorithms. An implementation of a key exchange algorithm can be obtained
474 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
475 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
476 the older EVP_PKEY_derive_init() function. See the man pages for the new
477 functions for further details.
480 *) The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
483 *) Removed the function names from error messages and deprecated the
486 *) Removed NextStep support and the macro OPENSSL_UNISTD
489 *) Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
490 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
491 Also removed "export var as function" capability; we do not export
492 variables, only functions.
495 *) RC5_32_set_key has been changed to return an int type, with 0 indicating
496 an error and 1 indicating success. In previous versions of OpenSSL this
497 was a void type. If a key was set longer than the maximum possible this
501 *) Support SM2 signing and verification schemes with X509 certificate.
504 *) Use SHA256 as the default digest for TS query in the ts app.
507 *) Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
508 This checks that the salt length is at least 128 bits, the derived key
509 length is at least 112 bits, and that the iteration count is at least 1000.
510 For backwards compatibility these checks are disabled by default in the
511 default provider, but are enabled by default in the fips provider.
512 To enable or disable these checks use the control
513 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
516 *) Default cipher lists/suites are now available via a function, the
517 #defines are deprecated.
520 *) Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
521 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
522 for Windows Store apps easier. Also, the "no-uplink" option has been added.
525 *) Join the directories crypto/x509 and crypto/x509v3
528 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
529 This changes the size when using the genpkey app when no size is given. It
530 fixes an omission in earlier changes that changed all RSA, DSA and DH
531 generation apps to use 2048 bits by default.
534 *) Added command 'openssl kdf' that uses the EVP_KDF API.
537 *) Added command 'openssl mac' that uses the EVP_MAC API.
540 *) Added OPENSSL_info() to get diverse built-in OpenSSL data, such
541 as default directories. Also added the command 'openssl info'
542 for scripting purposes.
545 *) The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
546 deprecated. These undocumented functions were never integrated into the EVP
547 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
548 Bi-directional IGE mode. These modes were never formally standardised and
549 usage of these functions is believed to be very small. In particular
550 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
551 is ever used. The security implications are believed to be minimal, but
552 this issue was never fixed for backwards compatibility reasons. New code
553 should not use these modes.
556 *) Add prediction resistance to the DRBG reseeding process.
559 *) Limit the number of blocks in a data unit for AES-XTS to 2^20 as
560 mandated by IEEE Std 1619-2018.
563 *) Added newline escaping functionality to a filename when using openssl dgst.
564 This output format is to replicate the output format found in the '*sum'
565 checksum programs. This aims to preserve backward compatibility.
566 [Matt Eaton, Richard Levitte, and Paul Dale]
568 *) Removed the heartbeat message in DTLS feature, as it has very
569 little usage and doesn't seem to fulfill a valuable purpose.
570 The configuration option is now deprecated.
573 *) Changed the output of 'openssl {digestname} < file' to display the
574 digest name in its output.
577 *) Added a new generic trace API which provides support for enabling
578 instrumentation through trace output. This feature is mainly intended
579 as an aid for developers and is disabled by default. To utilize it,
580 OpenSSL needs to be configured with the `enable-trace` option.
582 If the tracing API is enabled, the application can activate trace output
583 by registering BIOs as trace channels for a number of tracing and debugging
586 The 'openssl' application has been expanded to enable any of the types
587 available via environment variables defined by the user, and serves as
588 one possible example on how to use this functionality.
589 [Richard Levitte & Matthias St. Pierre]
591 *) Added build tests for C++. These are generated files that only do one
592 thing, to include one public OpenSSL head file each. This tests that
593 the public header files can be usefully included in a C++ application.
595 This test isn't enabled by default. It can be enabled with the option
596 'enable-buildtest-c++'.
599 *) Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
602 *) Add KMAC to EVP_MAC.
605 *) Added property based algorithm implementation selection framework to
609 *) Added SCA hardening for modular field inversion in EC_GROUP through
610 a new dedicated field_inv() pointer in EC_METHOD.
611 This also addresses a leakage affecting conversions from projective
612 to affine coordinates.
613 [Billy Bob Brumley, Nicola Tuveri]
615 *) Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
616 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
617 those algorithms that were already supported through the EVP_PKEY API
618 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
619 and scrypt are now wrappers that call EVP_KDF.
622 *) Build devcrypto engine as a dynamic engine.
625 *) Add keyed BLAKE2 to EVP_MAC.
628 *) Fix a bug in the computation of the endpoint-pair shared secret used
629 by DTLS over SCTP. This breaks interoperability with older versions
630 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
631 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
632 interoperability with such broken implementations. However, enabling
633 this switch breaks interoperability with correct implementations.
635 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
636 re-used X509_PUBKEY object if the second PUBKEY is malformed.
639 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
642 *) Change the license to the Apache License v2.0.
645 *) Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
647 o Major releases (indicated by incrementing the MAJOR release number)
648 may introduce incompatible API/ABI changes.
649 o Minor releases (indicated by incrementing the MINOR release number)
650 may introduce new features but retain API/ABI compatibility.
651 o Patch releases (indicated by incrementing the PATCH number)
652 are intended for bug fixes and other improvements of existing
653 features only (like improving performance or adding documentation)
654 and retain API/ABI compatibility.
657 *) Add support for RFC5297 SIV mode (siv128), including AES-SIV.
660 *) Remove the 'dist' target and add a tarball building script. The
661 'dist' target has fallen out of use, and it shouldn't be
662 necessary to configure just to create a source distribution.
665 *) Recreate the OS390-Unix config target. It no longer relies on a
666 special script like it did for OpenSSL pre-1.1.0.
669 *) Instead of having the source directories listed in Configure, add
670 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
674 *) Add GMAC to EVP_MAC.
677 *) Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
680 *) Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
681 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
682 to facilitate the continued use of MACs through raw private keys in
683 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
686 *) Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
687 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
690 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
691 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
692 are retained for backwards compatibility.
695 *) AES-XTS mode now enforces that its two keys are different to mitigate
696 the attacked described in "Efficient Instantiations of Tweakable
697 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
698 Details of this attack can be obtained from:
699 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
702 *) Rename the object files, i.e. give them other names than in previous
703 versions. Their names now include the name of the final product, as
704 well as its type mnemonic (bin, lib, shlib).
707 *) Added new option for 'openssl list', '-objects', which will display the
708 list of built in objects, i.e. OIDs with names.
711 *) Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
712 improves application performance by removing data copies and providing
713 applications with zero-copy system calls such as sendfile and splice.
716 Changes between 1.1.1a and 1.1.1b [xx XXX xxxx]
718 *) Change the info callback signals for the start and end of a post-handshake
719 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
720 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
721 confused by this and assume that a TLSv1.2 renegotiation has started. This
722 can break KeyUpdate handling. Instead we no longer signal the start and end
723 of a post handshake message exchange (although the messages themselves are
724 still signalled). This could break some applications that were expecting
725 the old signals. However without this KeyUpdate is not usable for many
729 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
731 *) Timing vulnerability in DSA signature generation
733 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
734 timing side channel attack. An attacker could use variations in the signing
735 algorithm to recover the private key.
737 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
741 *) Timing vulnerability in ECDSA signature generation
743 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
744 timing side channel attack. An attacker could use variations in the signing
745 algorithm to recover the private key.
747 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
751 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
752 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
753 of two gigabytes and the error handling improved.
755 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
756 categorized as a normal bug, not a security issue, because the DRBG reseeds
757 automatically and is fully functional even without additional randomness
758 provided by the application.
760 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
762 *) Add a new ClientHello callback. Provides a callback interface that gives
763 the application the ability to adjust the nascent SSL object at the
764 earliest stage of ClientHello processing, immediately after extensions have
765 been collected but before they have been processed. In particular, this
766 callback can adjust the supported TLS versions in response to the contents
770 *) Add SM2 base algorithm support.
773 *) s390x assembly pack: add (improved) hardware-support for the following
774 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
775 aes-cfb/cfb8, aes-ecb.
778 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
779 parameter is no longer accepted, as it leads to a corrupt table. NULL
780 pem_str is reserved for alias entries only.
783 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
784 step for prime curves. The new implementation is based on formulae from
785 differential addition-and-doubling in homogeneous projective coordinates
786 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
787 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
788 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
789 to work in projective coordinates.
790 [Billy Bob Brumley, Nicola Tuveri]
792 *) Change generating and checking of primes so that the error rate of not
793 being prime depends on the intended use based on the size of the input.
794 For larger primes this will result in more rounds of Miller-Rabin.
795 The maximal error rate for primes with more than 1080 bits is lowered
797 [Kurt Roeckx, Annie Yousar]
799 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
802 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
803 moving between systems, and to avoid confusion when a Windows build is
804 done with mingw vs with MSVC. For POSIX installs, there's still a
805 symlink or copy named 'tsget' to avoid that confusion as well.
808 *) Revert blinding in ECDSA sign and instead make problematic addition
809 length-invariant. Switch even to fixed-length Montgomery multiplication.
812 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
813 step for binary curves. The new implementation is based on formulae from
814 differential addition-and-doubling in mixed Lopez-Dahab projective
815 coordinates, modified to independently blind the operands.
816 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
818 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
819 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
820 EC_METHODs to implement their own specialized "ladder step", to take
821 advantage of more favorable coordinate systems or more efficient
822 differential addition-and-doubling algorithms.
823 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
825 *) Modified the random device based seed sources to keep the relevant
826 file descriptors open rather than reopening them on each access.
827 This allows such sources to operate in a chroot() jail without
828 the associated device nodes being available. This behaviour can be
829 controlled using RAND_keep_random_devices_open().
832 *) Numerous side-channel attack mitigations have been applied. This may have
833 performance impacts for some algorithms for the benefit of improved
834 security. Specific changes are noted in this change log by their respective
838 *) AIX shared library support overhaul. Switch to AIX "natural" way of
839 handling shared libraries, which means collecting shared objects of
840 different versions and bitnesses in one common archive. This allows to
841 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
842 doesn't affect the way 3rd party applications are linked, only how
843 multi-version installation is managed.
846 *) Make ec_group_do_inverse_ord() more robust and available to other
847 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
848 mitigations are applied to the fallback BN_mod_inverse().
849 When using this function rather than BN_mod_inverse() directly, new
850 EC cryptosystem implementations are then safer-by-default.
853 *) Add coordinate blinding for EC_POINT and implement projective
854 coordinate blinding for generic prime curves as a countermeasure to
855 chosen point SCA attacks.
856 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
858 *) Add blinding to ECDSA and DSA signatures to protect against side channel
859 attacks discovered by Keegan Ryan (NCC Group).
862 *) Enforce checking in the pkeyutl command line app to ensure that the input
863 length does not exceed the maximum supported digest length when performing
864 a sign, verify or verifyrecover operation.
867 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
868 I/O in combination with something like select() or poll() will hang. This
869 can be turned off again using SSL_CTX_clear_mode().
870 Many applications do not properly handle non-application data records, and
871 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
872 around the problems in those applications, but can also break some.
873 It's recommended to read the manpages about SSL_read(), SSL_write(),
874 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
875 SSL_CTX_set_read_ahead() again.
878 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
879 now allow empty (zero character) pass phrases.
882 *) Apply blinding to binary field modular inversion and remove patent
883 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
886 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
887 binary and prime elliptic curves.
890 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
891 constant time fixed point multiplication.
894 *) Revise elliptic curve scalar multiplication with timing attack
895 defenses: ec_wNAF_mul redirects to a constant time implementation
896 when computing fixed point and variable point multiplication (which
897 in OpenSSL are mostly used with secret scalars in keygen, sign,
898 ECDH derive operations).
899 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
902 *) Updated CONTRIBUTING
905 *) Updated DRBG / RAND to request nonce and additional low entropy
906 randomness from the system.
907 [Matthias St. Pierre]
909 *) Updated 'openssl rehash' to use OpenSSL consistent default.
912 *) Moved the load of the ssl_conf module to libcrypto, which helps
913 loading engines that libssl uses before libssl is initialised.
916 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
919 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
920 [Ingo Schwarze, Rich Salz]
922 *) Added output of accepting IP address and port for 'openssl s_server'
925 *) Added a new API for TLSv1.3 ciphersuites:
926 SSL_CTX_set_ciphersuites()
927 SSL_set_ciphersuites()
930 *) Memory allocation failures consistently add an error to the error
934 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
935 in libcrypto when run as setuid/setgid.
938 *) Load any config file by default when libssl is used.
941 *) Added new public header file <openssl/rand_drbg.h> and documentation
942 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
943 [Matthias St. Pierre]
945 *) QNX support removed (cannot find contributors to get their approval
946 for the license change).
949 *) TLSv1.3 replay protection for early data has been implemented. See the
950 SSL_read_early_data() man page for further details.
953 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
954 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
955 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
956 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
957 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
958 configuration has been separated out. See the ciphers man page or the
959 SSL_CTX_set_ciphersuites() man page for more information.
962 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
963 in responder mode now supports the new "-multi" option, which
964 spawns the specified number of child processes to handle OCSP
965 requests. The "-timeout" option now also limits the OCSP
966 responder's patience to wait to receive the full client request
967 on a newly accepted connection. Child processes are respawned
968 as needed, and the CA index file is automatically reloaded
969 when changed. This makes it possible to run the "ocsp" responder
970 as a long-running service, making the OpenSSL CA somewhat more
971 feature-complete. In this mode, most diagnostic messages logged
972 after entering the event loop are logged via syslog(3) rather than
976 *) Added support for X448 and Ed448. Heavily based on original work by
980 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
981 objects loaded. This adds the functions OSSL_STORE_expect() and
982 OSSL_STORE_find() as well as needed tools to construct searches and
983 get the search data out of them.
986 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
987 version of OpenSSL should review their configuration settings to ensure
988 that they are still appropriate for TLSv1.3. For further information see:
989 https://wiki.openssl.org/index.php/TLS1.3
992 *) Grand redesign of the OpenSSL random generator
994 The default RAND method now utilizes an AES-CTR DRBG according to
995 NIST standard SP 800-90Ar1. The new random generator is essentially
996 a port of the default random generator from the OpenSSL FIPS 2.0
997 object module. It is a hybrid deterministic random bit generator
998 using an AES-CTR bit stream and which seeds and reseeds itself
999 automatically using trusted system entropy sources.
1001 Some of its new features are:
1002 o Support for multiple DRBG instances with seed chaining.
1003 o The default RAND method makes use of a DRBG.
1004 o There is a public and private DRBG instance.
1005 o The DRBG instances are fork-safe.
1006 o Keep all global DRBG instances on the secure heap if it is enabled.
1007 o The public and private DRBG instance are per thread for lock free
1009 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
1011 *) Changed Configure so it only says what it does and doesn't dump
1012 so much data. Instead, ./configdata.pm should be used as a script
1013 to display all sorts of configuration data.
1016 *) Added processing of "make variables" to Configure.
1019 *) Added SHA512/224 and SHA512/256 algorithm support.
1022 *) The last traces of Netware support, first removed in 1.1.0, have
1026 *) Get rid of Makefile.shared, and in the process, make the processing
1027 of certain files (rc.obj, or the .def/.map/.opt files produced from
1028 the ordinal files) more visible and hopefully easier to trace and
1029 debug (or make silent).
1032 *) Make it possible to have environment variable assignments as
1033 arguments to config / Configure.
1036 *) Add multi-prime RSA (RFC 8017) support.
1039 *) Add SM3 implemented according to GB/T 32905-2016
1040 [ Jack Lloyd <jack.lloyd@ribose.com>,
1041 Ronald Tse <ronald.tse@ribose.com>,
1042 Erick Borsboom <erick.borsboom@ribose.com> ]
1044 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
1045 as documented in RFC6066.
1046 Based on a patch from Tomasz Moń
1047 [Filipe Raimundo da Silva]
1049 *) Add SM4 implemented according to GB/T 32907-2016.
1050 [ Jack Lloyd <jack.lloyd@ribose.com>,
1051 Ronald Tse <ronald.tse@ribose.com>,
1052 Erick Borsboom <erick.borsboom@ribose.com> ]
1054 *) Reimplement -newreq-nodes and ERR_error_string_n; the
1055 original author does not agree with the license change.
1058 *) Add ARIA AEAD TLS support.
1061 *) Some macro definitions to support VS6 have been removed. Visual
1062 Studio 6 has not worked since 1.1.0
1065 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
1066 without clearing the errors.
1069 *) Add "atfork" functions. If building on a system that without
1070 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1071 requirements. The RAND facility now uses/requires this.
1077 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
1078 not possible to disable entirely. However, it's still possible to
1079 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1082 To disable, configure with 'no-ui-console'. 'no-ui' is still
1083 possible to use as an alias. Check at compile time with the
1084 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1085 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1088 *) Add a STORE module, which implements a uniform and URI based reader of
1089 stores that can contain keys, certificates, CRLs and numerous other
1090 objects. The main API is loosely based on a few stdio functions,
1091 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1092 OSSL_STORE_error and OSSL_STORE_close.
1093 The implementation uses backends called "loaders" to implement arbitrary
1094 URI schemes. There is one built in "loader" for the 'file' scheme.
1097 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
1098 then adjusted to work on FreeBSD 8.4 as well.
1099 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1100 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1103 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
1104 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1105 error code calls like this:
1107 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1109 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1110 that can be encoded in C. For the foreseeable future, this will only
1112 [Richard Levitte and Tim Hudson]
1114 *) Removed BSD cryptodev engine.
1117 *) Add a build target 'build_all_generated', to build all generated files
1118 and only that. This can be used to prepare everything that requires
1119 things like perl for a system that lacks perl and then move everything
1120 to that system and do the rest of the build there.
1123 *) In the UI interface, make it possible to duplicate the user data. This
1124 can be used by engines that need to retain the data for a longer time
1125 than just the call where this user data is passed.
1128 *) Ignore the '-named_curve auto' value for compatibility of applications
1130 [Tomas Mraz <tmraz@fedoraproject.org>]
1132 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1133 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1134 alerts across multiple records (some of which could be empty). In practice
1135 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1136 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1137 support this at all. Supporting it adds significant complexity to the
1138 record layer, and its removal is unlikely to cause interoperability
1142 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1143 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1144 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1148 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1149 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1150 [Richard Levitte, Andy Polyakov]
1152 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1156 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1157 platform rather than 'mingw'.
1160 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
1161 success if they are asked to add an object which already exists
1162 in the store. This change cascades to other functions which load
1163 certificates and CRLs.
1166 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
1167 facilitate stack unwinding even from assembly subroutines.
1170 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1171 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1174 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1175 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1176 which is the minimum version we support.
1179 *) Certificate time validation (X509_cmp_time) enforces stricter
1180 compliance with RFC 5280. Fractional seconds and timezone offsets
1181 are no longer allowed.
1184 *) Add support for ARIA
1187 *) s_client will now send the Server Name Indication (SNI) extension by
1188 default unless the new "-noservername" option is used. The server name is
1189 based on the host provided to the "-connect" option unless overridden by
1190 using "-servername".
1193 *) Add support for SipHash
1196 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1197 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1198 prevent issues where no progress is being made and the peer continually
1199 sends unrecognised record types, using up resources processing them.
1202 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
1203 using the algorithm defined in
1204 https://www.akkadia.org/drepper/SHA-crypt.txt
1207 *) Heartbeat support has been removed; the ABI is changed for now.
1208 [Richard Levitte, Rich Salz]
1210 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1213 *) The RSA "null" method, which was partially supported to avoid patent
1214 issues, has been replaced to always returns NULL.
1218 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
1220 *) Client DoS due to large DH parameter
1222 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1223 malicious server can send a very large prime value to the client. This will
1224 cause the client to spend an unreasonably long period of time generating a
1225 key for this prime resulting in a hang until the client has finished. This
1226 could be exploited in a Denial Of Service attack.
1228 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1232 *) Cache timing vulnerability in RSA Key Generation
1234 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1235 a cache timing side channel attack. An attacker with sufficient access to
1236 mount cache timing attacks during the RSA key generation process could
1237 recover the private key.
1239 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1240 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1244 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1245 parameter is no longer accepted, as it leads to a corrupt table. NULL
1246 pem_str is reserved for alias entries only.
1249 *) Revert blinding in ECDSA sign and instead make problematic addition
1250 length-invariant. Switch even to fixed-length Montgomery multiplication.
1253 *) Change generating and checking of primes so that the error rate of not
1254 being prime depends on the intended use based on the size of the input.
1255 For larger primes this will result in more rounds of Miller-Rabin.
1256 The maximal error rate for primes with more than 1080 bits is lowered
1258 [Kurt Roeckx, Annie Yousar]
1260 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1263 *) Add blinding to ECDSA and DSA signatures to protect against side channel
1264 attacks discovered by Keegan Ryan (NCC Group).
1267 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1268 now allow empty (zero character) pass phrases.
1271 *) Certificate time validation (X509_cmp_time) enforces stricter
1272 compliance with RFC 5280. Fractional seconds and timezone offsets
1273 are no longer allowed.
1276 *) Fixed a text canonicalisation bug in CMS
1278 Where a CMS detached signature is used with text content the text goes
1279 through a canonicalisation process first prior to signing or verifying a
1280 signature. This process strips trailing space at the end of lines, converts
1281 line terminators to CRLF and removes additional trailing line terminators
1282 at the end of a file. A bug in the canonicalisation process meant that
1283 some characters, such as form-feed, were incorrectly treated as whitespace
1284 and removed. This is contrary to the specification (RFC5485). This fix
1285 could mean that detached text data signed with an earlier version of
1286 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1287 signed with a fixed OpenSSL may fail to verify with an earlier version of
1288 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1289 and use the "-binary" flag (for the "cms" command line application) or set
1290 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1293 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
1295 *) Constructed ASN.1 types with a recursive definition could exceed the stack
1297 Constructed ASN.1 types with a recursive definition (such as can be found
1298 in PKCS7) could eventually exceed the stack given malicious input with
1299 excessive recursion. This could result in a Denial Of Service attack. There
1300 are no such structures used within SSL/TLS that come from untrusted sources
1301 so this is considered safe.
1303 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1308 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1310 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1311 effectively reduced to only comparing the least significant bit of each
1312 byte. This allows an attacker to forge messages that would be considered as
1313 authenticated in an amount of tries lower than that guaranteed by the
1314 security claims of the scheme. The module can only be compiled by the
1315 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1317 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1322 *) Add a build target 'build_all_generated', to build all generated files
1323 and only that. This can be used to prepare everything that requires
1324 things like perl for a system that lacks perl and then move everything
1325 to that system and do the rest of the build there.
1328 *) Backport SSL_OP_NO_RENGOTIATION
1330 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1331 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1332 changes this is no longer possible in 1.1.0. Therefore the new
1333 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1334 1.1.0 to provide equivalent functionality.
1336 Note that if an application built against 1.1.0h headers (or above) is run
1337 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1338 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1341 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1345 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1347 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1348 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1349 Analysis suggests that attacks against RSA and DSA as a result of this
1350 defect would be very difficult to perform and are not believed likely.
1351 Attacks against DH1024 are considered just feasible, because most of the
1352 work necessary to deduce information about a private key may be performed
1353 offline. The amount of resources required for such an attack would be
1354 significant. However, for an attack on TLS to be meaningful, the server
1355 would have to share the DH1024 private key among multiple clients, which is
1356 no longer an option since CVE-2016-0701.
1358 This only affects processors that support the AVX2 but not ADX extensions
1359 like Intel Haswell (4th generation).
1361 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1362 was originally found via the OSS-Fuzz project.
1366 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1368 *) bn_sqrx8x_internal carry bug on x86_64
1370 There is a carry propagating bug in the x86_64 Montgomery squaring
1371 procedure. No EC algorithms are affected. Analysis suggests that attacks
1372 against RSA and DSA as a result of this defect would be very difficult to
1373 perform and are not believed likely. Attacks against DH are considered just
1374 feasible (although very difficult) because most of the work necessary to
1375 deduce information about a private key may be performed offline. The amount
1376 of resources required for such an attack would be very significant and
1377 likely only accessible to a limited number of attackers. An attacker would
1378 additionally need online access to an unpatched system using the target
1379 private key in a scenario with persistent DH parameters and a private
1380 key that is shared between multiple clients.
1382 This only affects processors that support the BMI1, BMI2 and ADX extensions
1383 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1385 This issue was reported to OpenSSL by the OSS-Fuzz project.
1389 *) Malformed X.509 IPAddressFamily could cause OOB read
1391 If an X.509 certificate has a malformed IPAddressFamily extension,
1392 OpenSSL could do a one-byte buffer overread. The most likely result
1393 would be an erroneous display of the certificate in text format.
1395 This issue was reported to OpenSSL by the OSS-Fuzz project.
1399 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1401 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1402 platform rather than 'mingw'.
1405 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1406 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1407 which is the minimum version we support.
1410 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1412 *) Encrypt-Then-Mac renegotiation crash
1414 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1415 negotiated where it was not in the original handshake (or vice-versa) then
1416 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1417 and servers are affected.
1419 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1423 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1425 *) Truncated packet could crash via OOB read
1427 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1428 cipher is being used, then a truncated packet can cause that host to
1429 perform an out-of-bounds read, usually resulting in a crash.
1431 This issue was reported to OpenSSL by Robert Święcki of Google.
1435 *) Bad (EC)DHE parameters cause a client crash
1437 If a malicious server supplies bad parameters for a DHE or ECDHE key
1438 exchange then this can result in the client attempting to dereference a
1439 NULL pointer leading to a client crash. This could be exploited in a Denial
1442 This issue was reported to OpenSSL by Guido Vranken.
1446 *) BN_mod_exp may produce incorrect results on x86_64
1448 There is a carry propagating bug in the x86_64 Montgomery squaring
1449 procedure. No EC algorithms are affected. Analysis suggests that attacks
1450 against RSA and DSA as a result of this defect would be very difficult to
1451 perform and are not believed likely. Attacks against DH are considered just
1452 feasible (although very difficult) because most of the work necessary to
1453 deduce information about a private key may be performed offline. The amount
1454 of resources required for such an attack would be very significant and
1455 likely only accessible to a limited number of attackers. An attacker would
1456 additionally need online access to an unpatched system using the target
1457 private key in a scenario with persistent DH parameters and a private
1458 key that is shared between multiple clients. For example this can occur by
1459 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1460 similar to CVE-2015-3193 but must be treated as a separate problem.
1462 This issue was reported to OpenSSL by the OSS-Fuzz project.
1466 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1468 *) ChaCha20/Poly1305 heap-buffer-overflow
1470 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1471 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1472 crash. This issue is not considered to be exploitable beyond a DoS.
1474 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1478 *) CMS Null dereference
1480 Applications parsing invalid CMS structures can crash with a NULL pointer
1481 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1482 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1483 structure callback if an attempt is made to free certain invalid encodings.
1484 Only CHOICE structures using a callback which do not handle NULL value are
1487 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1491 *) Montgomery multiplication may produce incorrect results
1493 There is a carry propagating bug in the Broadwell-specific Montgomery
1494 multiplication procedure that handles input lengths divisible by, but
1495 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1496 and DH private keys are impossible. This is because the subroutine in
1497 question is not used in operations with the private key itself and an input
1498 of the attacker's direct choice. Otherwise the bug can manifest itself as
1499 transient authentication and key negotiation failures or reproducible
1500 erroneous outcome of public-key operations with specially crafted input.
1501 Among EC algorithms only Brainpool P-512 curves are affected and one
1502 presumably can attack ECDH key negotiation. Impact was not analyzed in
1503 detail, because pre-requisites for attack are considered unlikely. Namely
1504 multiple clients have to choose the curve in question and the server has to
1505 share the private key among them, neither of which is default behaviour.
1506 Even then only clients that chose the curve will be affected.
1508 This issue was publicly reported as transient failures and was not
1509 initially recognized as a security issue. Thanks to Richard Morgan for
1510 providing reproducible case.
1514 *) Removed automatic addition of RPATH in shared libraries and executables,
1515 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1518 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1520 *) Fix Use After Free for large message sizes
1522 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1523 message larger than approx 16k is received then the underlying buffer to
1524 store the incoming message is reallocated and moved. Unfortunately a
1525 dangling pointer to the old location is left which results in an attempt to
1526 write to the previously freed location. This is likely to result in a
1527 crash, however it could potentially lead to execution of arbitrary code.
1529 This issue only affects OpenSSL 1.1.0a.
1531 This issue was reported to OpenSSL by Robert Święcki.
1535 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1537 *) OCSP Status Request extension unbounded memory growth
1539 A malicious client can send an excessively large OCSP Status Request
1540 extension. If that client continually requests renegotiation, sending a
1541 large OCSP Status Request extension each time, then there will be unbounded
1542 memory growth on the server. This will eventually lead to a Denial Of
1543 Service attack through memory exhaustion. Servers with a default
1544 configuration are vulnerable even if they do not support OCSP. Builds using
1545 the "no-ocsp" build time option are not affected.
1547 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1551 *) SSL_peek() hang on empty record
1553 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1554 sends an empty record. This could be exploited by a malicious peer in a
1555 Denial Of Service attack.
1557 This issue was reported to OpenSSL by Alex Gaynor.
1561 *) Excessive allocation of memory in tls_get_message_header() and
1562 dtls1_preprocess_fragment()
1564 A (D)TLS message includes 3 bytes for its length in the header for the
1565 message. This would allow for messages up to 16Mb in length. Messages of
1566 this length are excessive and OpenSSL includes a check to ensure that a
1567 peer is sending reasonably sized messages in order to avoid too much memory
1568 being consumed to service a connection. A flaw in the logic of version
1569 1.1.0 means that memory for the message is allocated too early, prior to
1570 the excessive message length check. Due to way memory is allocated in
1571 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1572 to service a connection. This could lead to a Denial of Service through
1573 memory exhaustion. However, the excessive message length check still takes
1574 place, and this would cause the connection to immediately fail. Assuming
1575 that the application calls SSL_free() on the failed connection in a timely
1576 manner then the 21Mb of allocated memory will then be immediately freed
1577 again. Therefore the excessive memory allocation will be transitory in
1578 nature. This then means that there is only a security impact if:
1580 1) The application does not call SSL_free() in a timely manner in the event
1581 that the connection fails
1583 2) The application is working in a constrained environment where there is
1584 very little free memory
1586 3) The attacker initiates multiple connection attempts such that there are
1587 multiple connections in a state where memory has been allocated for the
1588 connection; SSL_free() has not yet been called; and there is insufficient
1589 memory to service the multiple requests.
1591 Except in the instance of (1) above any Denial Of Service is likely to be
1592 transitory because as soon as the connection fails the memory is
1593 subsequently freed again in the SSL_free() call. However there is an
1594 increased risk during this period of application crashes due to the lack of
1595 memory - which would then mean a more serious Denial of Service.
1597 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1598 (CVE-2016-6307 and CVE-2016-6308)
1601 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1602 had to be removed. Primary reason is that vendor assembler can't
1603 assemble our modules with -KPIC flag. As result it, assembly
1604 support, was not even available as option. But its lack means
1605 lack of side-channel resistant code, which is incompatible with
1606 security by todays standards. Fortunately gcc is readily available
1607 prepackaged option, which we firmly point at...
1610 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1612 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1613 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1614 (to any value) allows Windows user to access PKCS#12 file generated
1615 with Windows CryptoAPI and protected with non-ASCII password, as well
1616 as files generated under UTF-8 locale on Linux also protected with
1620 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1621 have been disabled by default and removed from DEFAULT, just like RC4.
1622 See the RC4 item below to re-enable both.
1625 *) The method for finding the storage location for the Windows RAND seed file
1626 has changed. First we check %RANDFILE%. If that is not set then we check
1627 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1628 all else fails we fall back to C:\.
1631 *) The EVP_EncryptUpdate() function has had its return type changed from void
1632 to int. A return of 0 indicates and error while a return of 1 indicates
1636 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1637 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1638 off the constant time implementation for RSA, DSA and DH have been made
1639 no-ops and deprecated.
1642 *) Windows RAND implementation was simplified to only get entropy by
1643 calling CryptGenRandom(). Various other RAND-related tickets
1645 [Joseph Wylie Yandle, Rich Salz]
1647 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1648 and OPENSSL_LH_, respectively. The old names are available
1649 with API compatibility. They new names are now completely documented.
1652 *) Unify TYPE_up_ref(obj) methods signature.
1653 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1654 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1655 int (instead of void) like all others TYPE_up_ref() methods.
1656 So now these methods also check the return value of CRYPTO_atomic_add(),
1657 and the validity of object reference counter.
1658 [fdasilvayy@gmail.com]
1660 *) With Windows Visual Studio builds, the .pdb files are installed
1661 alongside the installed libraries and executables. For a static
1662 library installation, ossl_static.pdb is the associate compiler
1663 generated .pdb file to be used when linking programs.
1666 *) Remove openssl.spec. Packaging files belong with the packagers.
1669 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1670 recognise x86_64 architectures automatically. You can still decide
1671 to build for a different bitness with the environment variable
1672 KERNEL_BITS (can be 32 or 64), for example:
1674 KERNEL_BITS=32 ./config
1678 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1679 256 bit AES and HMAC with SHA256.
1682 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1685 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1688 *) To enable users to have their own config files and build file templates,
1689 Configure looks in the directory indicated by the environment variable
1690 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1691 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1692 name and is used as is.
1695 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1696 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1697 X509_CERT_FILE_CTX was removed.
1700 *) "shared" builds are now the default. To create only static libraries use
1701 the "no-shared" Configure option.
1704 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1705 All of these option have not worked for some while and are fundamental
1709 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1710 global cleanup functions are no longer required because they are handled
1711 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1712 Explicitly de-initing can cause problems (e.g. where a library that uses
1713 OpenSSL de-inits, but an application is still using it). The affected
1714 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1715 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1716 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1717 COMP_zlib_cleanup().
1720 *) --strict-warnings no longer enables runtime debugging options
1721 such as REF_DEBUG. Instead, debug options are automatically
1722 enabled with '--debug' builds.
1723 [Andy Polyakov, Emilia Käsper]
1725 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1726 have been moved out of the public header files. New functions for managing
1727 these have been added.
1730 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1731 objects have been moved out of the public header files. New
1732 functions for managing these have been added.
1735 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1736 have been moved out of the public header files. New functions for managing
1737 these have been added.
1740 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1741 moved out of the public header files. New functions for managing these
1745 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1748 *) Removed the mk1mf build scripts.
1751 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1752 it is always safe to #include a header now.
1755 *) Removed the aged BC-32 config and all its supporting scripts
1758 *) Removed support for Ultrix, Netware, and OS/2.
1761 *) Add support for HKDF.
1762 [Alessandro Ghedini]
1764 *) Add support for blake2b and blake2s
1767 *) Added support for "pipelining". Ciphers that have the
1768 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1769 encryptions/decryptions simultaneously. There are currently no built-in
1770 ciphers with this property but the expectation is that engines will be able
1771 to offer it to significantly improve throughput. Support has been extended
1772 into libssl so that multiple records for a single connection can be
1773 processed in one go (for >=TLS 1.1).
1776 *) Added the AFALG engine. This is an async capable engine which is able to
1777 offload work to the Linux kernel. In this initial version it only supports
1778 AES128-CBC. The kernel must be version 4.1.0 or greater.
1781 *) OpenSSL now uses a new threading API. It is no longer necessary to
1782 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1783 are two supported threading models: pthreads and windows threads. It is
1784 also possible to configure OpenSSL at compile time for "no-threads". The
1785 old threading API should no longer be used. The functions have been
1786 replaced with "no-op" compatibility macros.
1787 [Alessandro Ghedini, Matt Caswell]
1789 *) Modify behavior of ALPN to invoke callback after SNI/servername
1790 callback, such that updates to the SSL_CTX affect ALPN.
1793 *) Add SSL_CIPHER queries for authentication and key-exchange.
1796 *) Changes to the DEFAULT cipherlist:
1797 - Prefer (EC)DHE handshakes over plain RSA.
1798 - Prefer AEAD ciphers over legacy ciphers.
1799 - Prefer ECDSA over RSA when both certificates are available.
1800 - Prefer TLSv1.2 ciphers/PRF.
1801 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1805 *) Change the ECC default curve list to be this, in order: x25519,
1806 secp256r1, secp521r1, secp384r1.
1809 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1810 disabled by default. They can be re-enabled using the
1811 enable-weak-ssl-ciphers option to Configure.
1814 *) If the server has ALPN configured, but supports no protocols that the
1815 client advertises, send a fatal "no_application_protocol" alert.
1816 This behaviour is SHALL in RFC 7301, though it isn't universally
1817 implemented by other servers.
1820 *) Add X25519 support.
1821 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1822 for public and private key encoding using the format documented in
1823 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1824 key generation and key derivation.
1826 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1830 *) Deprecate SRP_VBASE_get_by_user.
1831 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1832 In order to fix an unavoidable memory leak (CVE-2016-0798),
1833 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1834 seed, even if the seed is configured.
1836 Users should use SRP_VBASE_get1_by_user instead. Note that in
1837 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1838 also that even though configuring the SRP seed attempts to hide
1839 invalid usernames by continuing the handshake with fake
1840 credentials, this behaviour is not constant time and no strong
1841 guarantees are made that the handshake is indistinguishable from
1842 that of a valid user.
1845 *) Configuration change; it's now possible to build dynamic engines
1846 without having to build shared libraries and vice versa. This
1847 only applies to the engines in engines/, those in crypto/engine/
1848 will always be built into libcrypto (i.e. "static").
1850 Building dynamic engines is enabled by default; to disable, use
1851 the configuration option "disable-dynamic-engine".
1853 The only requirements for building dynamic engines are the
1854 presence of the DSO module and building with position independent
1855 code, so they will also automatically be disabled if configuring
1856 with "disable-dso" or "disable-pic".
1858 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1859 are also taken away from openssl/opensslconf.h, as they are
1863 *) Configuration change; if there is a known flag to compile
1864 position independent code, it will always be applied on the
1865 libcrypto and libssl object files, and never on the application
1866 object files. This means other libraries that use routines from
1867 libcrypto / libssl can be made into shared libraries regardless
1868 of how OpenSSL was configured.
1870 If this isn't desirable, the configuration options "disable-pic"
1871 or "no-pic" can be used to disable the use of PIC. This will
1872 also disable building shared libraries and dynamic engines.
1875 *) Removed JPAKE code. It was experimental and has no wide use.
1878 *) The INSTALL_PREFIX Makefile variable has been renamed to
1879 DESTDIR. That makes for less confusion on what this variable
1880 is for. Also, the configuration option --install_prefix is
1884 *) Heartbeat for TLS has been removed and is disabled by default
1885 for DTLS; configure with enable-heartbeats. Code that uses the
1886 old #define's might need to be updated.
1887 [Emilia Käsper, Rich Salz]
1889 *) Rename REF_CHECK to REF_DEBUG.
1892 *) New "unified" build system
1894 The "unified" build system is aimed to be a common system for all
1895 platforms we support. With it comes new support for VMS.
1897 This system builds supports building in a different directory tree
1898 than the source tree. It produces one Makefile (for unix family
1899 or lookalikes), or one descrip.mms (for VMS).
1901 The source of information to make the Makefile / descrip.mms is
1902 small files called 'build.info', holding the necessary
1903 information for each directory with source to compile, and a
1904 template in Configurations, like unix-Makefile.tmpl or
1907 With this change, the library names were also renamed on Windows
1908 and on VMS. They now have names that are closer to the standard
1909 on Unix, and include the major version number, and in certain
1910 cases, the architecture they are built for. See "Notes on shared
1911 libraries" in INSTALL.
1913 We rely heavily on the perl module Text::Template.
1916 *) Added support for auto-initialisation and de-initialisation of the library.
1917 OpenSSL no longer requires explicit init or deinit routines to be called,
1918 except in certain circumstances. See the OPENSSL_init_crypto() and
1919 OPENSSL_init_ssl() man pages for further information.
1922 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1923 "peer" argument is now expected to be a BIO_ADDR object.
1925 *) Rewrite of BIO networking library. The BIO library lacked consistent
1926 support of IPv6, and adding it required some more extensive
1927 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1928 which hold all types of addresses and chains of address information.
1929 It also introduces a new API, with functions like BIO_socket,
1930 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1931 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1932 have been adapted accordingly.
1935 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1939 *) CRIME protection: disable compression by default, even if OpenSSL is
1940 compiled with zlib enabled. Applications can still enable compression
1941 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1942 using the SSL_CONF library to configure compression.
1945 *) The signature of the session callback configured with
1946 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1947 was explicitly marked as 'const unsigned char*' instead of
1951 *) Always DPURIFY. Remove the use of uninitialized memory in the
1952 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1955 *) Removed many obsolete configuration items, including
1956 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1957 MD2_CHAR, MD2_INT, MD2_LONG
1959 IDEA_SHORT, IDEA_LONG
1960 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1961 [Rich Salz, with advice from Andy Polyakov]
1963 *) Many BN internals have been moved to an internal header file.
1964 [Rich Salz with help from Andy Polyakov]
1966 *) Configuration and writing out the results from it has changed.
1967 Files such as Makefile include/openssl/opensslconf.h and are now
1968 produced through general templates, such as Makefile.in and
1969 crypto/opensslconf.h.in and some help from the perl module
1972 Also, the center of configuration information is no longer
1973 Makefile. Instead, Configure produces a perl module in
1974 configdata.pm which holds most of the config data (in the hash
1975 table %config), the target data that comes from the target
1976 configuration in one of the Configurations/*.conf files (in
1980 *) To clarify their intended purposes, the Configure options
1981 --prefix and --openssldir change their semantics, and become more
1982 straightforward and less interdependent.
1984 --prefix shall be used exclusively to give the location INSTALLTOP
1985 where programs, scripts, libraries, include files and manuals are
1986 going to be installed. The default is now /usr/local.
1988 --openssldir shall be used exclusively to give the default
1989 location OPENSSLDIR where certificates, private keys, CRLs are
1990 managed. This is also where the default openssl.cnf gets
1992 If the directory given with this option is a relative path, the
1993 values of both the --prefix value and the --openssldir value will
1994 be combined to become OPENSSLDIR.
1995 The default for --openssldir is INSTALLTOP/ssl.
1997 Anyone who uses --openssldir to specify where OpenSSL is to be
1998 installed MUST change to use --prefix instead.
2001 *) The GOST engine was out of date and therefore it has been removed. An up
2002 to date GOST engine is now being maintained in an external repository.
2003 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2004 support for GOST ciphersuites (these are only activated if a GOST engine
2008 *) EGD is no longer supported by default; use enable-egd when
2010 [Ben Kaduk and Rich Salz]
2012 *) The distribution now has Makefile.in files, which are used to
2013 create Makefile's when Configure is run. *Configure must be run
2014 before trying to build now.*
2017 *) The return value for SSL_CIPHER_description() for error conditions
2021 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2023 Obtaining and performing DNSSEC validation of TLSA records is
2024 the application's responsibility. The application provides
2025 the TLSA records of its choice to OpenSSL, and these are then
2026 used to authenticate the peer.
2028 The TLSA records need not even come from DNS. They can, for
2029 example, be used to implement local end-entity certificate or
2030 trust-anchor "pinning", where the "pin" data takes the form
2031 of TLSA records, which can augment or replace verification
2032 based on the usual WebPKI public certification authorities.
2035 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2036 continues to support deprecated interfaces in default builds.
2037 However, applications are strongly advised to compile their
2038 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2039 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2040 or the 1.1.0 releases.
2042 In environments in which all applications have been ported to
2043 not use any deprecated interfaces OpenSSL's Configure script
2044 should be used with the --api=1.1.0 option to entirely remove
2045 support for the deprecated features from the library and
2046 unconditionally disable them in the installed headers.
2047 Essentially the same effect can be achieved with the "no-deprecated"
2048 argument to Configure, except that this will always restrict
2049 the build to just the latest API, rather than a fixed API
2052 As applications are ported to future revisions of the API,
2053 they should update their compile-time OPENSSL_API_COMPAT define
2054 accordingly, but in most cases should be able to continue to
2055 compile with later releases.
2057 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2058 0x10000000L and 0x00908000L, respectively. However those
2059 versions did not support the OPENSSL_API_COMPAT feature, and
2060 so applications are not typically tested for explicit support
2061 of just the undeprecated features of either release.
2064 *) Add support for setting the minimum and maximum supported protocol.
2065 It can bet set via the SSL_set_min_proto_version() and
2066 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2067 MaxProtocol. It's recommended to use the new APIs to disable
2068 protocols instead of disabling individual protocols using
2069 SSL_set_options() or SSL_CONF's Protocol. This change also
2070 removes support for disabling TLS 1.2 in the OpenSSL TLS
2071 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2074 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2077 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2078 and integrates ECDSA and ECDH functionality into EC. Implementations can
2079 now redirect key generation and no longer need to convert to or from
2082 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2083 include the ec.h header file instead.
2086 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
2087 ciphers who are no longer supported and drops support the ephemeral RSA key
2088 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2091 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2092 opaque. For HMAC_CTX, the following constructors and destructors
2095 HMAC_CTX *HMAC_CTX_new(void);
2096 void HMAC_CTX_free(HMAC_CTX *ctx);
2098 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2099 destroy such methods has been added. See EVP_MD_meth_new(3) and
2100 EVP_CIPHER_meth_new(3) for documentation.
2103 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2104 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2105 EVP_MD_CTX_reset() should be called instead to reinitialise
2106 an already created structure.
2107 2) For consistency with the majority of our object creators and
2108 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2109 EVP_MD_CTX_(new|free). The old names are retained as macros
2110 for deprecated builds.
2113 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
2114 cryptographic operations to be performed asynchronously as long as an
2115 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2116 further details. Libssl has also had this capability integrated with the
2117 introduction of the new mode SSL_MODE_ASYNC and associated error
2118 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2119 pages. This work was developed in partnership with Intel Corp.
2122 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2123 always enabled now. If you want to disable the support you should
2124 exclude it using the list of supported ciphers. This also means that the
2125 "-no_ecdhe" option has been removed from s_server.
2128 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2129 SSL_{CTX_}set1_curves() which can set a list.
2132 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2133 curve you want to support using SSL_{CTX_}set1_curves().
2136 *) State machine rewrite. The state machine code has been significantly
2137 refactored in order to remove much duplication of code and solve issues
2138 with the old code (see ssl/statem/README for further details). This change
2139 does have some associated API changes. Notably the SSL_state() function
2140 has been removed and replaced by SSL_get_state which now returns an
2141 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2142 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2146 *) All instances of the string "ssleay" in the public API were replaced
2147 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2148 Some error codes related to internal RSA_eay API's were renamed.
2151 *) The demo files in crypto/threads were moved to demo/threads.
2154 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2156 [Matt Caswell, Rich Salz]
2158 *) New ASN.1 embed macro.
2160 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2161 structure is not allocated: it is part of the parent. That is instead of
2169 This reduces memory fragmentation and make it impossible to accidentally
2170 set a mandatory field to NULL.
2172 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2173 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2174 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2178 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2181 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2182 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2183 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2184 DES and RC4 ciphersuites.
2187 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2188 This changes the decoding behaviour for some invalid messages,
2189 though the change is mostly in the more lenient direction, and
2190 legacy behaviour is preserved as much as possible.
2193 *) Fix no-stdio build.
2194 [ David Woodhouse <David.Woodhouse@intel.com> and also
2195 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
2197 *) New testing framework
2198 The testing framework has been largely rewritten and is now using
2199 perl and the perl modules Test::Harness and an extended variant of
2200 Test::More called OpenSSL::Test to do its work. All test scripts in
2201 test/ have been rewritten into test recipes, and all direct calls to
2202 executables in test/Makefile have become individual recipes using the
2203 simplified testing OpenSSL::Test::Simple.
2205 For documentation on our testing modules, do:
2207 perldoc test/testlib/OpenSSL/Test/Simple.pm
2208 perldoc test/testlib/OpenSSL/Test.pm
2212 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
2213 are used; the latter aborts on memory leaks (usually checked on exit).
2214 Some undocumented "set malloc, etc., hooks" functions were removed
2215 and others were changed. All are now documented.
2218 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2220 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2222 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
2223 from RFC4279, RFC4785, RFC5487, RFC5489.
2225 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
2226 original RSA_PSK patch.
2229 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
2230 era flag was never set throughout the codebase (only read). Also removed
2231 SSL3_FLAGS_POP_BUFFER which was only used if
2232 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
2235 *) Changed the default name options in the "ca", "crl", "req" and "x509"
2236 to be "oneline" instead of "compat".
2239 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
2240 not aware of clients that still exhibit this bug, and the workaround
2241 hasn't been working properly for a while.
2244 *) The return type of BIO_number_read() and BIO_number_written() as well as
2245 the corresponding num_read and num_write members in the BIO structure has
2246 changed from unsigned long to uint64_t. On platforms where an unsigned
2247 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
2251 *) Given the pervasive nature of TLS extensions it is inadvisable to run
2252 OpenSSL without support for them. It also means that maintaining
2253 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
2254 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
2257 *) Removed support for the two export grade static DH ciphersuites
2258 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
2259 were newly added (along with a number of other static DH ciphersuites) to
2260 1.0.2. However the two export ones have *never* worked since they were
2261 introduced. It seems strange in any case to be adding new export
2262 ciphersuites, and given "logjam" it also does not seem correct to fix them.
2265 *) Version negotiation has been rewritten. In particular SSLv23_method(),
2266 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
2267 and turned into macros which simply call the new preferred function names
2268 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
2269 should use the new names instead. Also as part of this change the ssl23.h
2270 header file has been removed.
2273 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2274 code and the associated standard is no longer considered fit-for-purpose.
2277 *) RT2547 was closed. When generating a private key, try to make the
2278 output file readable only by the owner. This behavior change might
2279 be noticeable when interacting with other software.
2281 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
2285 *) Added HTTP GET support to the ocsp command.
2288 *) Changed default digest for the dgst and enc commands from MD5 to
2292 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2295 *) Added support for TLS extended master secret from
2296 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2297 initial patch which was a great help during development.
2300 *) All libssl internal structures have been removed from the public header
2301 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2302 now redundant). Users should not attempt to access internal structures
2303 directly. Instead they should use the provided API functions.
2306 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2307 Access to deprecated functions can be re-enabled by running config with
2308 "enable-deprecated". In addition applications wishing to use deprecated
2309 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2310 will, by default, disable some transitive includes that previously existed
2311 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2314 *) Added support for OCB mode. OpenSSL has been granted a patent license
2315 compatible with the OpenSSL license for use of OCB. Details are available
2316 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2317 for OCB can be removed by calling config with no-ocb.
2320 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2321 compatible client hello.
2324 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2325 done while fixing the error code for the key-too-small case.
2326 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2328 *) CA.sh has been removed; use CA.pl instead.
2331 *) Removed old DES API.
2334 *) Remove various unsupported platforms:
2340 Sinix/ReliantUNIX RM400
2345 16-bit platforms such as WIN16
2348 *) Clean up OPENSSL_NO_xxx #define's
2349 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2350 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2351 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2352 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2353 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2354 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2355 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2356 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2357 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2358 Remove MS_STATIC; it's a relic from platforms <32 bits.
2361 *) Cleaned up dead code
2362 Remove all but one '#ifdef undef' which is to be looked at.
2365 *) Clean up calling of xxx_free routines.
2366 Just like free(), fix most of the xxx_free routines to accept
2367 NULL. Remove the non-null checks from callers. Save much code.
2370 *) Add secure heap for storage of private keys (when possible).
2371 Add BIO_s_secmem(), CBIGNUM, etc.
2372 Contributed by Akamai Technologies under our Corporate CLA.
2375 *) Experimental support for a new, fast, unbiased prime candidate generator,
2376 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2377 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2379 *) New output format NSS in the sess_id command line tool. This allows
2380 exporting the session id and the master key in NSS keylog format.
2381 [Martin Kaiser <martin@kaiser.cx>]
2383 *) Harmonize version and its documentation. -f flag is used to display
2385 [mancha <mancha1@zoho.com>]
2387 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2388 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2389 [mancha <mancha1@zoho.com>]
2391 *) Fix some double frees. These are not thought to be exploitable.
2392 [mancha <mancha1@zoho.com>]
2394 *) A missing bounds check in the handling of the TLS heartbeat extension
2395 can be used to reveal up to 64k of memory to a connected client or
2398 Thanks for Neel Mehta of Google Security for discovering this bug and to
2399 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2400 preparing the fix (CVE-2014-0160)
2401 [Adam Langley, Bodo Moeller]
2403 *) Fix for the attack described in the paper "Recovering OpenSSL
2404 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2405 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2406 http://eprint.iacr.org/2014/140
2408 Thanks to Yuval Yarom and Naomi Benger for discovering this
2409 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2410 [Yuval Yarom and Naomi Benger]
2412 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2413 this fixes a limitation in previous versions of OpenSSL.
2416 *) Experimental encrypt-then-mac support.
2418 Experimental support for encrypt then mac from
2419 draft-gutmann-tls-encrypt-then-mac-02.txt
2421 To enable it set the appropriate extension number (0x42 for the test
2422 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2424 For non-compliant peers (i.e. just about everything) this should have no
2427 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2431 *) Add EVP support for key wrapping algorithms, to avoid problems with
2432 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2433 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2434 algorithms and include tests cases.
2437 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2441 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2442 MGF1 digest and OAEP label.
2445 *) Make openssl verify return errors.
2446 [Chris Palmer <palmer@google.com> and Ben Laurie]
2448 *) New function ASN1_TIME_diff to calculate the difference between two
2449 ASN1_TIME structures or one structure and the current time.
2452 *) Update fips_test_suite to support multiple command line options. New
2453 test to induce all self test errors in sequence and check expected
2457 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2458 sign or verify all in one operation.
2461 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2462 test programs and fips_test_suite. Includes functionality to parse
2463 the minimal script output of fipsalgest.pl directly.
2466 *) Add authorisation parameter to FIPS_module_mode_set().
2469 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2472 *) Use separate DRBG fields for internal and external flags. New function
2473 FIPS_drbg_health_check() to perform on demand health checking. Add
2474 generation tests to fips_test_suite with reduced health check interval to
2475 demonstrate periodic health checking. Add "nodh" option to
2476 fips_test_suite to skip very slow DH test.
2479 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2483 *) More extensive health check for DRBG checking many more failure modes.
2484 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2485 combination: call this in fips_test_suite.
2488 *) Add support for canonical generation of DSA parameter 'g'. See
2491 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2492 POST to handle HMAC cases.
2495 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2496 to return numerical and string versions of the FIPS module number.
2499 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2500 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2501 outside the validated module in the FIPS capable OpenSSL.
2504 *) Minor change to DRBG entropy callback semantics. In some cases
2505 there is no multiple of the block length between min_len and
2506 max_len. Allow the callback to return more than max_len bytes
2507 of entropy but discard any extra: it is the callback's responsibility
2508 to ensure that the extra data discarded does not impact the
2509 requested amount of entropy.
2512 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2513 information in FIPS186-3, SP800-57 and SP800-131A.
2516 *) CCM support via EVP. Interface is very similar to GCM case except we
2517 must supply all data in one chunk (i.e. no update, final) and the
2518 message length must be supplied if AAD is used. Add algorithm test
2522 *) Initial version of POST overhaul. Add POST callback to allow the status
2523 of POST to be monitored and/or failures induced. Modify fips_test_suite
2524 to use callback. Always run all selftests even if one fails.
2527 *) XTS support including algorithm test driver in the fips_gcmtest program.
2528 Note: this does increase the maximum key length from 32 to 64 bytes but
2529 there should be no binary compatibility issues as existing applications
2530 will never use XTS mode.
2533 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2534 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2535 performs algorithm blocking for unapproved PRNG types. Also do not
2536 set PRNG type in FIPS_mode_set(): leave this to the application.
2537 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2538 the standard OpenSSL PRNG: set additional data to a date time vector.
2541 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2542 This shouldn't present any incompatibility problems because applications
2543 shouldn't be using these directly and any that are will need to rethink
2544 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2547 *) Extensive self tests and health checking required by SP800-90 DRBG.
2548 Remove strength parameter from FIPS_drbg_instantiate and always
2549 instantiate at maximum supported strength.
2552 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2555 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2558 *) New function DH_compute_key_padded() to compute a DH key and pad with
2559 leading zeroes if needed: this complies with SP800-56A et al.
2562 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2563 anything, incomplete, subject to change and largely untested at present.
2566 *) Modify fipscanisteronly build option to only build the necessary object
2567 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2570 *) Add experimental option FIPSSYMS to give all symbols in
2571 fipscanister.o and FIPS or fips prefix. This will avoid
2572 conflicts with future versions of OpenSSL. Add perl script
2573 util/fipsas.pl to preprocess assembly language source files
2574 and rename any affected symbols.
2577 *) Add selftest checks and algorithm block of non-fips algorithms in
2578 FIPS mode. Remove DES2 from selftests.
2581 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2582 return internal method without any ENGINE dependencies. Add new
2583 tiny fips sign and verify functions.
2586 *) New build option no-ec2m to disable characteristic 2 code.
2589 *) New build option "fipscanisteronly". This only builds fipscanister.o
2590 and (currently) associated fips utilities. Uses the file Makefile.fips
2591 instead of Makefile.org as the prototype.
2594 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2595 Update fips_gcmtest to use IV generator.
2598 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2599 setting output buffer to NULL. The *Final function must be
2600 called although it will not retrieve any additional data. The tag
2601 can be set or retrieved with a ctrl. The IV length is by default 12
2602 bytes (96 bits) but can be set to an alternative value. If the IV
2603 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2607 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2608 underlying do_cipher function handles all cipher semantics itself
2609 including padding and finalisation. This is useful if (for example)
2610 an ENGINE cipher handles block padding itself. The behaviour of
2611 do_cipher is subtly changed if this flag is set: the return value
2612 is the number of characters written to the output buffer (zero is
2613 no longer an error code) or a negative error code. Also if the
2614 input buffer is NULL and length 0 finalisation should be performed.
2617 *) If a candidate issuer certificate is already part of the constructed
2618 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2621 *) Improve forward-security support: add functions
2623 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2624 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2626 for use by SSL/TLS servers; the callback function will be called whenever a
2627 new session is created, and gets to decide whether the session may be
2628 cached to make it resumable (return 0) or not (return 1). (As by the
2629 SSL/TLS protocol specifications, the session_id sent by the server will be
2630 empty to indicate that the session is not resumable; also, the server will
2631 not generate RFC 4507 (RFC 5077) session tickets.)
2633 A simple reasonable callback implementation is to return is_forward_secure.
2634 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2635 by the SSL/TLS server library, indicating whether it can provide forward
2637 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2639 *) New -verify_name option in command line utilities to set verification
2643 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2644 Add CMAC pkey methods.
2647 *) Experimental renegotiation in s_server -www mode. If the client
2648 browses /reneg connection is renegotiated. If /renegcert it is
2649 renegotiated requesting a certificate.
2652 *) Add an "external" session cache for debugging purposes to s_server. This
2653 should help trace issues which normally are only apparent in deployed
2654 multi-process servers.
2657 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2658 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2659 BIO_set_cipher() and some obscure PEM functions were changed so they
2660 can now return an error. The RAND changes required a change to the
2661 RAND_METHOD structure.
2664 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2665 a gcc attribute to warn if the result of a function is ignored. This
2666 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2667 whose return value is often ignored.
2670 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2671 These allow SCTs (signed certificate timestamps) to be requested and
2672 validated when establishing a connection.
2673 [Rob Percival <robpercival@google.com>]
2675 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2677 *) Prevent padding oracle in AES-NI CBC MAC check
2679 A MITM attacker can use a padding oracle attack to decrypt traffic
2680 when the connection uses an AES CBC cipher and the server support
2683 This issue was introduced as part of the fix for Lucky 13 padding
2684 attack (CVE-2013-0169). The padding check was rewritten to be in
2685 constant time by making sure that always the same bytes are read and
2686 compared against either the MAC or padding bytes. But it no longer
2687 checked that there was enough data to have both the MAC and padding
2690 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2694 *) Fix EVP_EncodeUpdate overflow
2696 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2697 Base64 encoding of binary data. If an attacker is able to supply very large
2698 amounts of input data then a length check can overflow resulting in a heap
2701 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2702 the PEM_write_bio* family of functions. These are mainly used within the
2703 OpenSSL command line applications, so any application which processes data
2704 from an untrusted source and outputs it as a PEM file should be considered
2705 vulnerable to this issue. User applications that call these APIs directly
2706 with large amounts of untrusted data may also be vulnerable.
2708 This issue was reported by Guido Vranken.
2712 *) Fix EVP_EncryptUpdate overflow
2714 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2715 is able to supply very large amounts of input data after a previous call to
2716 EVP_EncryptUpdate() with a partial block then a length check can overflow
2717 resulting in a heap corruption. Following an analysis of all OpenSSL
2718 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2719 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2720 the first called function after an EVP_EncryptInit(), and therefore that
2721 specific call must be safe. The second form is where the length passed to
2722 EVP_EncryptUpdate() can be seen from the code to be some small value and
2723 therefore there is no possibility of an overflow. Since all instances are
2724 one of these two forms, it is believed that there can be no overflows in
2725 internal code due to this problem. It should be noted that
2726 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2727 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2728 of these calls have also been analysed too and it is believed there are no
2729 instances in internal usage where an overflow could occur.
2731 This issue was reported by Guido Vranken.
2735 *) Prevent ASN.1 BIO excessive memory allocation
2737 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2738 a short invalid encoding can cause allocation of large amounts of memory
2739 potentially consuming excessive resources or exhausting memory.
2741 Any application parsing untrusted data through d2i BIO functions is
2742 affected. The memory based functions such as d2i_X509() are *not* affected.
2743 Since the memory based functions are used by the TLS library, TLS
2744 applications are not affected.
2746 This issue was reported by Brian Carpenter.
2752 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2753 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2754 in arbitrary stack data being returned in the buffer.
2756 This issue was reported by Guido Vranken.
2760 *) Modify behavior of ALPN to invoke callback after SNI/servername
2761 callback, such that updates to the SSL_CTX affect ALPN.
2764 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2768 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2769 methods are enabled and ssl2 is disabled the methods return NULL.
2772 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2774 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2775 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2776 provide any "EXPORT" or "LOW" strength ciphers.
2779 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2780 is by default disabled at build-time. Builds that are not configured with
2781 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2782 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2783 will need to explicitly call either of:
2785 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2787 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2789 as appropriate. Even if either of those is used, or the application
2790 explicitly uses the version-specific SSLv2_method() or its client and
2791 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2792 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2793 ciphers, and SSLv2 56-bit DES are no longer available.
2797 *) Fix a double-free in DSA code
2799 A double free bug was discovered when OpenSSL parses malformed DSA private
2800 keys and could lead to a DoS attack or memory corruption for applications
2801 that receive DSA private keys from untrusted sources. This scenario is
2804 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2809 *) Disable SRP fake user seed to address a server memory leak.
2811 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2813 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2814 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2815 was changed to ignore the "fake user" SRP seed, even if the seed
2818 Users should use SRP_VBASE_get1_by_user instead. Note that in
2819 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2820 also that even though configuring the SRP seed attempts to hide
2821 invalid usernames by continuing the handshake with fake
2822 credentials, this behaviour is not constant time and no strong
2823 guarantees are made that the handshake is indistinguishable from
2824 that of a valid user.
2828 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2830 In the BN_hex2bn function the number of hex digits is calculated using an
2831 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2832 large values of |i| this can result in |bn_expand| not allocating any
2833 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2834 field as NULL leading to a subsequent NULL ptr deref. For very large values
2835 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2836 In this case memory is allocated to the internal BIGNUM data field, but it
2837 is insufficiently sized leading to heap corruption. A similar issue exists
2838 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2839 is ever called by user applications with very large untrusted hex/dec data.
2840 This is anticipated to be a rare occurrence.
2842 All OpenSSL internal usage of these functions use data that is not expected
2843 to be untrusted, e.g. config file data or application command line
2844 arguments. If user developed applications generate config file data based
2845 on untrusted data then it is possible that this could also lead to security
2846 consequences. This is also anticipated to be rare.
2848 This issue was reported to OpenSSL by Guido Vranken.
2852 *) Fix memory issues in BIO_*printf functions
2854 The internal |fmtstr| function used in processing a "%s" format string in
2855 the BIO_*printf functions could overflow while calculating the length of a
2856 string and cause an OOB read when printing very long strings.
2858 Additionally the internal |doapr_outch| function can attempt to write to an
2859 OOB memory location (at an offset from the NULL pointer) in the event of a
2860 memory allocation failure. In 1.0.2 and below this could be caused where
2861 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2862 could be in processing a very long "%s" format string. Memory leaks can
2865 The first issue may mask the second issue dependent on compiler behaviour.
2866 These problems could enable attacks where large amounts of untrusted data
2867 is passed to the BIO_*printf functions. If applications use these functions
2868 in this way then they could be vulnerable. OpenSSL itself uses these
2869 functions when printing out human-readable dumps of ASN.1 data. Therefore
2870 applications that print this data could be vulnerable if the data is from
2871 untrusted sources. OpenSSL command line applications could also be
2872 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2873 as command line arguments.
2875 Libssl is not considered directly vulnerable. Additionally certificates etc
2876 received via remote connections via libssl are also unlikely to be able to
2877 trigger these issues because of message size limits enforced within libssl.
2879 This issue was reported to OpenSSL Guido Vranken.
2883 *) Side channel attack on modular exponentiation
2885 A side-channel attack was found which makes use of cache-bank conflicts on
2886 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2887 of RSA keys. The ability to exploit this issue is limited as it relies on
2888 an attacker who has control of code in a thread running on the same
2889 hyper-threaded core as the victim thread which is performing decryptions.
2891 This issue was reported to OpenSSL by Yuval Yarom, The University of
2892 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2893 Nadia Heninger, University of Pennsylvania with more information at
2894 http://cachebleed.info.
2898 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2899 if no keysize is specified with default_bits. This fixes an
2900 omission in an earlier change that changed all RSA/DSA key generation
2901 apps to use 2048 bits by default.
2904 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2905 *) DH small subgroups
2907 Historically OpenSSL only ever generated DH parameters based on "safe"
2908 primes. More recently (in version 1.0.2) support was provided for
2909 generating X9.42 style parameter files such as those required for RFC 5114
2910 support. The primes used in such files may not be "safe". Where an
2911 application is using DH configured with parameters based on primes that are
2912 not "safe" then an attacker could use this fact to find a peer's private
2913 DH exponent. This attack requires that the attacker complete multiple
2914 handshakes in which the peer uses the same private DH exponent. For example
2915 this could be used to discover a TLS server's private DH exponent if it's
2916 reusing the private DH exponent or it's using a static DH ciphersuite.
2918 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2919 TLS. It is not on by default. If the option is not set then the server
2920 reuses the same private DH exponent for the life of the server process and
2921 would be vulnerable to this attack. It is believed that many popular
2922 applications do set this option and would therefore not be at risk.
2924 The fix for this issue adds an additional check where a "q" parameter is
2925 available (as is the case in X9.42 based parameters). This detects the
2926 only known attack, and is the only possible defense for static DH
2927 ciphersuites. This could have some performance impact.
2929 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2930 default and cannot be disabled. This could have some performance impact.
2932 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2936 *) SSLv2 doesn't block disabled ciphers
2938 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2939 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2940 been disabled, provided that the SSLv2 protocol was not also disabled via
2943 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2944 and Sebastian Schinzel.
2948 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2950 *) BN_mod_exp may produce incorrect results on x86_64
2952 There is a carry propagating bug in the x86_64 Montgomery squaring
2953 procedure. No EC algorithms are affected. Analysis suggests that attacks
2954 against RSA and DSA as a result of this defect would be very difficult to
2955 perform and are not believed likely. Attacks against DH are considered just
2956 feasible (although very difficult) because most of the work necessary to
2957 deduce information about a private key may be performed offline. The amount
2958 of resources required for such an attack would be very significant and
2959 likely only accessible to a limited number of attackers. An attacker would
2960 additionally need online access to an unpatched system using the target
2961 private key in a scenario with persistent DH parameters and a private
2962 key that is shared between multiple clients. For example this can occur by
2963 default in OpenSSL DHE based SSL/TLS ciphersuites.
2965 This issue was reported to OpenSSL by Hanno Böck.
2969 *) Certificate verify crash with missing PSS parameter
2971 The signature verification routines will crash with a NULL pointer
2972 dereference if presented with an ASN.1 signature using the RSA PSS
2973 algorithm and absent mask generation function parameter. Since these
2974 routines are used to verify certificate signature algorithms this can be
2975 used to crash any certificate verification operation and exploited in a
2976 DoS attack. Any application which performs certificate verification is
2977 vulnerable including OpenSSL clients and servers which enable client
2980 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2984 *) X509_ATTRIBUTE memory leak
2986 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2987 memory. This structure is used by the PKCS#7 and CMS routines so any
2988 application which reads PKCS#7 or CMS data from untrusted sources is
2989 affected. SSL/TLS is not affected.
2991 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2996 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2997 This changes the decoding behaviour for some invalid messages,
2998 though the change is mostly in the more lenient direction, and
2999 legacy behaviour is preserved as much as possible.
3002 *) In DSA_generate_parameters_ex, if the provided seed is too short,
3004 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
3006 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
3008 *) Alternate chains certificate forgery
3010 During certificate verification, OpenSSL will attempt to find an
3011 alternative certificate chain if the first attempt to build such a chain
3012 fails. An error in the implementation of this logic can mean that an
3013 attacker could cause certain checks on untrusted certificates to be
3014 bypassed, such as the CA flag, enabling them to use a valid leaf
3015 certificate to act as a CA and "issue" an invalid certificate.
3017 This issue was reported to OpenSSL by Adam Langley/David Benjamin
3021 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
3023 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
3024 incompatibility in the handling of HMAC. The previous ABI has now been
3028 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
3030 *) Malformed ECParameters causes infinite loop
3032 When processing an ECParameters structure OpenSSL enters an infinite loop
3033 if the curve specified is over a specially malformed binary polynomial
3036 This can be used to perform denial of service against any
3037 system which processes public keys, certificate requests or
3038 certificates. This includes TLS clients and TLS servers with
3039 client authentication enabled.
3041 This issue was reported to OpenSSL by Joseph Barr-Pixton.
3045 *) Exploitable out-of-bounds read in X509_cmp_time
3047 X509_cmp_time does not properly check the length of the ASN1_TIME
3048 string and can read a few bytes out of bounds. In addition,
3049 X509_cmp_time accepts an arbitrary number of fractional seconds in the
3052 An attacker can use this to craft malformed certificates and CRLs of
3053 various sizes and potentially cause a segmentation fault, resulting in
3054 a DoS on applications that verify certificates or CRLs. TLS clients
3055 that verify CRLs are affected. TLS clients and servers with client
3056 authentication enabled may be affected if they use custom verification
3059 This issue was reported to OpenSSL by Robert Swiecki (Google), and
3060 independently by Hanno Böck.
3064 *) PKCS7 crash with missing EnvelopedContent
3066 The PKCS#7 parsing code does not handle missing inner EncryptedContent
3067 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
3068 with missing content and trigger a NULL pointer dereference on parsing.
3070 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
3071 structures from untrusted sources are affected. OpenSSL clients and
3072 servers are not affected.
3074 This issue was reported to OpenSSL by Michal Zalewski (Google).
3078 *) CMS verify infinite loop with unknown hash function
3080 When verifying a signedData message the CMS code can enter an infinite loop
3081 if presented with an unknown hash function OID. This can be used to perform
3082 denial of service against any system which verifies signedData messages using
3084 This issue was reported to OpenSSL by Johannes Bauer.
3088 *) Race condition handling NewSessionTicket
3090 If a NewSessionTicket is received by a multi-threaded client when attempting to
3091 reuse a previous ticket then a race condition can occur potentially leading to
3092 a double free of the ticket data.
3096 *) Only support 256-bit or stronger elliptic curves with the
3097 'ecdh_auto' setting (server) or by default (client). Of supported
3098 curves, prefer P-256 (both).
3101 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
3103 *) ClientHello sigalgs DoS fix
3105 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
3106 invalid signature algorithms extension a NULL pointer dereference will
3107 occur. This can be exploited in a DoS attack against the server.
3109 This issue was was reported to OpenSSL by David Ramos of Stanford
3112 [Stephen Henson and Matt Caswell]
3114 *) Multiblock corrupted pointer fix
3116 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
3117 feature only applies on 64 bit x86 architecture platforms that support AES
3118 NI instructions. A defect in the implementation of "multiblock" can cause
3119 OpenSSL's internal write buffer to become incorrectly set to NULL when
3120 using non-blocking IO. Typically, when the user application is using a
3121 socket BIO for writing, this will only result in a failed connection.
3122 However if some other BIO is used then it is likely that a segmentation
3123 fault will be triggered, thus enabling a potential DoS attack.
3125 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
3129 *) Segmentation fault in DTLSv1_listen fix
3131 The DTLSv1_listen function is intended to be stateless and processes the
3132 initial ClientHello from many peers. It is common for user code to loop
3133 over the call to DTLSv1_listen until a valid ClientHello is received with
3134 an associated cookie. A defect in the implementation of DTLSv1_listen means
3135 that state is preserved in the SSL object from one invocation to the next
3136 that can lead to a segmentation fault. Errors processing the initial
3137 ClientHello can trigger this scenario. An example of such an error could be
3138 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
3141 This issue was reported to OpenSSL by Per Allansson.
3145 *) Segmentation fault in ASN1_TYPE_cmp fix
3147 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
3148 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
3149 certificate signature algorithm consistency this can be used to crash any
3150 certificate verification operation and exploited in a DoS attack. Any
3151 application which performs certificate verification is vulnerable including
3152 OpenSSL clients and servers which enable client authentication.
3156 *) Segmentation fault for invalid PSS parameters fix
3158 The signature verification routines will crash with a NULL pointer
3159 dereference if presented with an ASN.1 signature using the RSA PSS
3160 algorithm and invalid parameters. Since these routines are used to verify
3161 certificate signature algorithms this can be used to crash any
3162 certificate verification operation and exploited in a DoS attack. Any
3163 application which performs certificate verification is vulnerable including
3164 OpenSSL clients and servers which enable client authentication.
3166 This issue was was reported to OpenSSL by Brian Carpenter.
3170 *) ASN.1 structure reuse memory corruption fix
3172 Reusing a structure in ASN.1 parsing may allow an attacker to cause
3173 memory corruption via an invalid write. Such reuse is and has been
3174 strongly discouraged and is believed to be rare.
3176 Applications that parse structures containing CHOICE or ANY DEFINED BY
3177 components may be affected. Certificate parsing (d2i_X509 and related
3178 functions) are however not affected. OpenSSL clients and servers are
3183 *) PKCS7 NULL pointer dereferences fix
3185 The PKCS#7 parsing code does not handle missing outer ContentInfo
3186 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
3187 missing content and trigger a NULL pointer dereference on parsing.
3189 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
3190 otherwise parse PKCS#7 structures from untrusted sources are
3191 affected. OpenSSL clients and servers are not affected.
3193 This issue was reported to OpenSSL by Michal Zalewski (Google).
3197 *) DoS via reachable assert in SSLv2 servers fix
3199 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
3200 servers that both support SSLv2 and enable export cipher suites by sending
3201 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
3203 This issue was discovered by Sean Burford (Google) and Emilia Käsper
3204 (OpenSSL development team).
3208 *) Empty CKE with client auth and DHE fix
3210 If client auth is used then a server can seg fault in the event of a DHE
3211 ciphersuite being selected and a zero length ClientKeyExchange message
3212 being sent by the client. This could be exploited in a DoS attack.
3216 *) Handshake with unseeded PRNG fix
3218 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
3219 with an unseeded PRNG. The conditions are:
3220 - The client is on a platform where the PRNG has not been seeded
3221 automatically, and the user has not seeded manually
3222 - A protocol specific client method version has been used (i.e. not
3223 SSL_client_methodv23)
3224 - A ciphersuite is used that does not require additional random data from
3225 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
3227 If the handshake succeeds then the client random that has been used will
3228 have been generated from a PRNG with insufficient entropy and therefore the
3229 output may be predictable.
3231 For example using the following command with an unseeded openssl will
3232 succeed on an unpatched platform:
3234 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
3238 *) Use After Free following d2i_ECPrivatekey error fix
3240 A malformed EC private key file consumed via the d2i_ECPrivateKey function
3241 could cause a use after free condition. This, in turn, could cause a double
3242 free in several private key parsing functions (such as d2i_PrivateKey
3243 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
3244 for applications that receive EC private keys from untrusted
3245 sources. This scenario is considered rare.
3247 This issue was discovered by the BoringSSL project and fixed in their
3252 *) X509_to_X509_REQ NULL pointer deref fix
3254 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
3255 the certificate key is invalid. This function is rarely used in practice.
3257 This issue was discovered by Brian Carpenter.
3261 *) Removed the export ciphers from the DEFAULT ciphers
3264 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
3266 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
3267 ARMv5 through ARMv8, as opposite to "locking" it to single one.
3268 So far those who have to target multiple platforms would compromise
3269 and argue that binary targeting say ARMv5 would still execute on
3270 ARMv8. "Universal" build resolves this compromise by providing
3271 near-optimal performance even on newer platforms.
3274 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
3275 (other platforms pending).
3276 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
3278 *) Add support for the SignedCertificateTimestampList certificate and
3279 OCSP response extensions from RFC6962.
3282 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3283 for corner cases. (Certain input points at infinity could lead to
3284 bogus results, with non-infinity inputs mapped to infinity too.)
3287 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
3288 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3289 common cases are optimized and there still is room for further
3290 improvements. Vector Permutation AES for Altivec is also added.
3293 *) Add support for little-endian ppc64 Linux target.
3294 [Marcelo Cerri (IBM)]
3296 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
3297 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3298 are optimized and there still is room for further improvements.
3299 Both 32- and 64-bit modes are supported.
3300 [Andy Polyakov, Ard Biesheuvel (Linaro)]
3302 *) Improved ARMv7 NEON support.
3305 *) Support for SPARC Architecture 2011 crypto extensions, first
3306 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3307 SHA256/512, MD5, GHASH and modular exponentiation.
3308 [Andy Polyakov, David Miller]
3310 *) Accelerated modular exponentiation for Intel processors, a.k.a.
3312 [Shay Gueron & Vlad Krasnov (Intel Corp)]
3314 *) Support for new and upcoming Intel processors, including AVX2,
3315 BMI and SHA ISA extensions. This includes additional "stitched"
3316 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3319 This work was sponsored by Intel Corp.
3322 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3323 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3324 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3327 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3328 this fixes a limitation in previous versions of OpenSSL.
3331 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3332 MGF1 digest and OAEP label.
3335 *) Add EVP support for key wrapping algorithms, to avoid problems with
3336 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3337 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3338 algorithms and include tests cases.
3341 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3343 [Douglas E. Engert, Steve Henson]
3345 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3346 difference in days and seconds between two tm or ASN1_TIME structures.
3349 *) Add -rev test option to s_server to just reverse order of characters
3350 received by client and send back to server. Also prints an abbreviated
3351 summary of the connection parameters.
3354 *) New option -brief for s_client and s_server to print out a brief summary
3355 of connection parameters.
3358 *) Add callbacks for arbitrary TLS extensions.
3359 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3361 *) New option -crl_download in several openssl utilities to download CRLs
3362 from CRLDP extension in certificates.
3365 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3368 *) New function X509_CRL_diff to generate a delta CRL from the difference
3369 of two full CRLs. Add support to "crl" utility.
3372 *) New functions to set lookup_crls function and to retrieve
3373 X509_STORE from X509_STORE_CTX.
3376 *) Print out deprecated issuer and subject unique ID fields in
3380 *) Extend OCSP I/O functions so they can be used for simple general purpose
3381 HTTP as well as OCSP. New wrapper function which can be used to download
3382 CRLs using the OCSP API.
3385 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3388 *) SSL_CONF* functions. These provide a common framework for application
3389 configuration using configuration files or command lines.
3392 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3393 message callback and prints the results. Needs compile time option
3394 "enable-ssl-trace". New options to s_client and s_server to enable
3398 *) New ctrl and macro to retrieve supported points extensions.
3399 Print out extension in s_server and s_client.
3402 *) New functions to retrieve certificate signature and signature
3406 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3410 *) New Suite B modes for TLS code. These use and enforce the requirements
3411 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3412 only use Suite B curves. The Suite B modes can be set by using the
3413 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3416 *) New chain verification flags for Suite B levels of security. Check
3417 algorithms are acceptable when flags are set in X509_verify_cert.
3420 *) Make tls1_check_chain return a set of flags indicating checks passed
3421 by a certificate chain. Add additional tests to handle client
3422 certificates: checks for matching certificate type and issuer name
3426 *) If an attempt is made to use a signature algorithm not in the peer
3427 preference list abort the handshake. If client has no suitable
3428 signature algorithms in response to a certificate request do not
3429 use the certificate.
3432 *) If server EC tmp key is not in client preference list abort handshake.
3435 *) Add support for certificate stores in CERT structure. This makes it
3436 possible to have different stores per SSL structure or one store in
3437 the parent SSL_CTX. Include distinct stores for certificate chain
3438 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3439 to build and store a certificate chain in CERT structure: returning
3440 an error if the chain cannot be built: this will allow applications
3441 to test if a chain is correctly configured.
3443 Note: if the CERT based stores are not set then the parent SSL_CTX
3444 store is used to retain compatibility with existing behaviour.
3448 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3449 mask based on the current session, check mask when sending client
3450 hello and checking the requested ciphersuite.
3453 *) New ctrls to retrieve and set certificate types in a certificate
3454 request message. Print out received values in s_client. If certificate
3455 types is not set with custom values set sensible values based on
3456 supported signature algorithms.
3459 *) Support for distinct client and server supported signature algorithms.
3462 *) Add certificate callback. If set this is called whenever a certificate
3463 is required by client or server. An application can decide which
3464 certificate chain to present based on arbitrary criteria: for example
3465 supported signature algorithms. Add very simple example to s_server.
3466 This fixes many of the problems and restrictions of the existing client
3467 certificate callback: for example you can now clear an existing
3468 certificate and specify the whole chain.
3471 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3472 the certificate can be used for (if anything). Set valid_flags field
3473 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3474 to have similar checks in it.
3476 Add new "cert_flags" field to CERT structure and include a "strict mode".
3477 This enforces some TLS certificate requirements (such as only permitting
3478 certificate signature algorithms contained in the supported algorithms
3479 extension) which some implementations ignore: this option should be used
3480 with caution as it could cause interoperability issues.
3483 *) Update and tidy signature algorithm extension processing. Work out
3484 shared signature algorithms based on preferences and peer algorithms
3485 and print them out in s_client and s_server. Abort handshake if no
3486 shared signature algorithms.
3489 *) Add new functions to allow customised supported signature algorithms
3490 for SSL and SSL_CTX structures. Add options to s_client and s_server
3494 *) New function SSL_certs_clear() to delete all references to certificates
3495 from an SSL structure. Before this once a certificate had been added
3496 it couldn't be removed.
3499 *) Integrate hostname, email address and IP address checking with certificate
3500 verification. New verify options supporting checking in openssl utility.
3503 *) Fixes and wildcard matching support to hostname and email checking
3504 functions. Add manual page.
3505 [Florian Weimer (Red Hat Product Security Team)]
3507 *) New functions to check a hostname email or IP address against a
3508 certificate. Add options x509 utility to print results of checks against
3512 *) Fix OCSP checking.
3513 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3515 *) Initial experimental support for explicitly trusted non-root CAs.
3516 OpenSSL still tries to build a complete chain to a root but if an
3517 intermediate CA has a trust setting included that is used. The first
3518 setting is used: whether to trust (e.g., -addtrust option to the x509
3522 *) Add -trusted_first option which attempts to find certificates in the
3523 trusted store even if an untrusted chain is also supplied.
3526 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3527 platform support for Linux and Android.
3530 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3533 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3534 When in FIPS mode the approved implementations are used as normal,
3535 when not in FIPS mode the internal unapproved versions are used instead.
3536 This means that the FIPS capable OpenSSL isn't forced to use the
3537 (often lower performance) FIPS implementations outside FIPS mode.
3540 *) Transparently support X9.42 DH parameters when calling
3541 PEM_read_bio_DHparameters. This means existing applications can handle
3542 the new parameter format automatically.
3545 *) Initial experimental support for X9.42 DH parameter format: mainly
3546 to support use of 'q' parameter for RFC5114 parameters.
3549 *) Add DH parameters from RFC5114 including test data to dhtest.
3552 *) Support for automatic EC temporary key parameter selection. If enabled
3553 the most preferred EC parameters are automatically used instead of
3554 hardcoded fixed parameters. Now a server just has to call:
3555 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3556 support ECDH and use the most appropriate parameters.
3559 *) Enhance and tidy EC curve and point format TLS extension code. Use
3560 static structures instead of allocation if default values are used.
3561 New ctrls to set curves we wish to support and to retrieve shared curves.
3562 Print out shared curves in s_server. New options to s_server and s_client
3563 to set list of supported curves.
3566 *) New ctrls to retrieve supported signature algorithms and
3567 supported curve values as an array of NIDs. Extend openssl utility
3568 to print out received values.
3571 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3572 between NIDs and the more common NIST names such as "P-256". Enhance
3573 ecparam utility and ECC method to recognise the NIST names for curves.
3576 *) Enhance SSL/TLS certificate chain handling to support different
3577 chains for each certificate instead of one chain in the parent SSL_CTX.
3580 *) Support for fixed DH ciphersuite client authentication: where both
3581 server and client use DH certificates with common parameters.
3584 *) Support for fixed DH ciphersuites: those requiring DH server
3588 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3590 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3591 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3592 X509_CINF_get_signature were reverted post internal team review.
3594 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3596 *) Build fixes for the Windows and OpenVMS platforms
3597 [Matt Caswell and Richard Levitte]
3599 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3601 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3602 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3603 dereference. This could lead to a Denial Of Service attack. Thanks to
3604 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3608 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3609 dtls1_buffer_record function under certain conditions. In particular this
3610 could occur if an attacker sent repeated DTLS records with the same
3611 sequence number but for the next epoch. The memory leak could be exploited
3612 by an attacker in a Denial of Service attack through memory exhaustion.
3613 Thanks to Chris Mueller for reporting this issue.
3617 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3618 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3619 method would be set to NULL which could later result in a NULL pointer
3620 dereference. Thanks to Frank Schmirler for reporting this issue.
3624 *) Abort handshake if server key exchange message is omitted for ephemeral
3627 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3628 reporting this issue.
3632 *) Remove non-export ephemeral RSA code on client and server. This code
3633 violated the TLS standard by allowing the use of temporary RSA keys in
3634 non-export ciphersuites and could be used by a server to effectively
3635 downgrade the RSA key length used to a value smaller than the server
3636 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3637 INRIA or reporting this issue.
3641 *) Fixed issue where DH client certificates are accepted without verification.
3642 An OpenSSL server will accept a DH certificate for client authentication
3643 without the certificate verify message. This effectively allows a client to
3644 authenticate without the use of a private key. This only affects servers
3645 which trust a client certificate authority which issues certificates
3646 containing DH keys: these are extremely rare and hardly ever encountered.
3647 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3652 *) Ensure that the session ID context of an SSL is updated when its
3653 SSL_CTX is updated via SSL_set_SSL_CTX.
3655 The session ID context is typically set from the parent SSL_CTX,
3656 and can vary with the CTX.
3659 *) Fix various certificate fingerprint issues.
3661 By using non-DER or invalid encodings outside the signed portion of a
3662 certificate the fingerprint can be changed without breaking the signature.
3663 Although no details of the signed portion of the certificate can be changed
3664 this can cause problems with some applications: e.g. those using the
3665 certificate fingerprint for blacklists.
3667 1. Reject signatures with non zero unused bits.
3669 If the BIT STRING containing the signature has non zero unused bits reject
3670 the signature. All current signature algorithms require zero unused bits.
3672 2. Check certificate algorithm consistency.
3674 Check the AlgorithmIdentifier inside TBS matches the one in the
3675 certificate signature. NB: this will result in signature failure
3676 errors for some broken certificates.
3678 Thanks to Konrad Kraszewski from Google for reporting this issue.
3680 3. Check DSA/ECDSA signatures use DER.
3682 Re-encode DSA/ECDSA signatures and compare with the original received
3683 signature. Return an error if there is a mismatch.
3685 This will reject various cases including garbage after signature
3686 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3687 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3688 (negative or with leading zeroes).
3690 Further analysis was conducted and fixes were developed by Stephen Henson
3691 of the OpenSSL core team.
3696 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3697 results on some platforms, including x86_64. This bug occurs at random
3698 with a very low probability, and is not known to be exploitable in any
3699 way, though its exact impact is difficult to determine. Thanks to Pieter
3700 Wuille (Blockstream) who reported this issue and also suggested an initial
3701 fix. Further analysis was conducted by the OpenSSL development team and
3702 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3703 the OpenSSL core team.
3707 *) Do not resume sessions on the server if the negotiated protocol
3708 version does not match the session's version. Resuming with a different
3709 version, while not strictly forbidden by the RFC, is of questionable
3710 sanity and breaks all known clients.
3711 [David Benjamin, Emilia Käsper]
3713 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3714 early CCS messages during renegotiation. (Note that because
3715 renegotiation is encrypted, this early CCS was not exploitable.)
3718 *) Tighten client-side session ticket handling during renegotiation:
3719 ensure that the client only accepts a session ticket if the server sends
3720 the extension anew in the ServerHello. Previously, a TLS client would
3721 reuse the old extension state and thus accept a session ticket if one was
3722 announced in the initial ServerHello.
3724 Similarly, ensure that the client requires a session ticket if one
3725 was advertised in the ServerHello. Previously, a TLS client would
3726 ignore a missing NewSessionTicket message.
3729 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3731 *) SRTP Memory Leak.
3733 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3734 sends a carefully crafted handshake message, to cause OpenSSL to fail
3735 to free up to 64k of memory causing a memory leak. This could be
3736 exploited in a Denial Of Service attack. This issue affects OpenSSL
3737 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3738 whether SRTP is used or configured. Implementations of OpenSSL that
3739 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3741 The fix was developed by the OpenSSL team.
3745 *) Session Ticket Memory Leak.
3747 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3748 integrity of that ticket is first verified. In the event of a session
3749 ticket integrity check failing, OpenSSL will fail to free memory
3750 causing a memory leak. By sending a large number of invalid session
3751 tickets an attacker could exploit this issue in a Denial Of Service
3756 *) Build option no-ssl3 is incomplete.
3758 When OpenSSL is configured with "no-ssl3" as a build option, servers
3759 could accept and complete a SSL 3.0 handshake, and clients could be
3760 configured to send them.
3762 [Akamai and the OpenSSL team]
3764 *) Add support for TLS_FALLBACK_SCSV.
3765 Client applications doing fallback retries should call
3766 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3768 [Adam Langley, Bodo Moeller]
3770 *) Add additional DigestInfo checks.
3772 Re-encode DigestInto in DER and check against the original when
3773 verifying RSA signature: this will reject any improperly encoded
3774 DigestInfo structures.
3776 Note: this is a precautionary measure and no attacks are currently known.
3780 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3782 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3783 SRP code can be overrun an internal buffer. Add sanity check that
3784 g, A, B < N to SRP code.
3786 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3787 Group for discovering this issue.
3791 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3792 TLS 1.0 instead of higher protocol versions when the ClientHello message
3793 is badly fragmented. This allows a man-in-the-middle attacker to force a
3794 downgrade to TLS 1.0 even if both the server and the client support a
3795 higher protocol version, by modifying the client's TLS records.
3797 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3798 researching this issue.
3802 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3803 to a denial of service attack. A malicious server can crash the client
3804 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3805 ciphersuite and sending carefully crafted handshake messages.
3807 Thanks to Felix Gröbert (Google) for discovering and researching this
3812 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3813 to leak memory. This can be exploited through a Denial of Service attack.
3814 Thanks to Adam Langley for discovering and researching this issue.
3818 *) An attacker can force openssl to consume large amounts of memory whilst
3819 processing DTLS handshake messages. This can be exploited through a
3820 Denial of Service attack.
3821 Thanks to Adam Langley for discovering and researching this issue.
3825 *) An attacker can force an error condition which causes openssl to crash
3826 whilst processing DTLS packets due to memory being freed twice. This
3827 can be exploited through a Denial of Service attack.
3828 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3833 *) If a multithreaded client connects to a malicious server using a resumed
3834 session and the server sends an ec point format extension it could write
3835 up to 255 bytes to freed memory.
3837 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3842 *) A malicious server can crash an OpenSSL client with a null pointer
3843 dereference (read) by specifying an SRP ciphersuite even though it was not
3844 properly negotiated with the client. This can be exploited through a
3845 Denial of Service attack.
3847 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3848 discovering and researching this issue.
3852 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3853 X509_name_oneline, X509_name_print_ex et al. to leak some information
3854 from the stack. Applications may be affected if they echo pretty printing
3855 output to the attacker.
3857 Thanks to Ivan Fratric (Google) for discovering this issue.
3859 [Emilia Käsper, and Steve Henson]
3861 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3862 for corner cases. (Certain input points at infinity could lead to
3863 bogus results, with non-infinity inputs mapped to infinity too.)
3866 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3868 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3869 handshake can force the use of weak keying material in OpenSSL
3870 SSL/TLS clients and servers.
3872 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3873 researching this issue. (CVE-2014-0224)
3874 [KIKUCHI Masashi, Steve Henson]
3876 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3877 OpenSSL DTLS client the code can be made to recurse eventually crashing
3880 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3882 [Imre Rad, Steve Henson]
3884 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3885 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3886 client or server. This is potentially exploitable to run arbitrary
3887 code on a vulnerable client or server.
3889 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3890 [Jüri Aedla, Steve Henson]
3892 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3893 are subject to a denial of service attack.
3895 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3896 this issue. (CVE-2014-3470)
3897 [Felix Gröbert, Ivan Fratric, Steve Henson]
3899 *) Harmonize version and its documentation. -f flag is used to display
3901 [mancha <mancha1@zoho.com>]
3903 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3904 in i2d_ECPrivateKey.
3905 [mancha <mancha1@zoho.com>]
3907 *) Fix some double frees. These are not thought to be exploitable.
3908 [mancha <mancha1@zoho.com>]
3910 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3912 *) A missing bounds check in the handling of the TLS heartbeat extension
3913 can be used to reveal up to 64k of memory to a connected client or
3916 Thanks for Neel Mehta of Google Security for discovering this bug and to
3917 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3918 preparing the fix (CVE-2014-0160)
3919 [Adam Langley, Bodo Moeller]
3921 *) Fix for the attack described in the paper "Recovering OpenSSL
3922 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3923 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3924 http://eprint.iacr.org/2014/140
3926 Thanks to Yuval Yarom and Naomi Benger for discovering this
3927 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3928 [Yuval Yarom and Naomi Benger]
3930 *) TLS pad extension: draft-agl-tls-padding-03
3932 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3933 TLS client Hello record length value would otherwise be > 255 and
3934 less that 512 pad with a dummy extension containing zeroes so it
3935 is at least 512 bytes long.
3937 [Adam Langley, Steve Henson]
3939 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3941 *) Fix for TLS record tampering bug. A carefully crafted invalid
3942 handshake could crash OpenSSL with a NULL pointer exception.
3943 Thanks to Anton Johansson for reporting this issues.
3946 *) Keep original DTLS digest and encryption contexts in retransmission
3947 structures so we can use the previous session parameters if they need
3948 to be resent. (CVE-2013-6450)
3951 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3952 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3953 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3954 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3955 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3956 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3957 [Rob Stradling, Adam Langley]
3959 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3961 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3962 supporting platforms or when small records were transferred.
3963 [Andy Polyakov, Steve Henson]
3965 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3967 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3969 This addresses the flaw in CBC record processing discovered by
3970 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3971 at: http://www.isg.rhul.ac.uk/tls/
3973 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3974 Security Group at Royal Holloway, University of London
3975 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3976 Emilia Käsper for the initial patch.
3978 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3980 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3981 ciphersuites which can be exploited in a denial of service attack.
3982 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3983 and detecting this bug and to Wolfgang Ettlinger
3984 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3988 *) Return an error when checking OCSP signatures when key is NULL.
3989 This fixes a DoS attack. (CVE-2013-0166)
3992 *) Make openssl verify return errors.
3993 [Chris Palmer <palmer@google.com> and Ben Laurie]
3995 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3996 the right response is stapled. Also change SSL_get_certificate()
3997 so it returns the certificate actually sent.
3998 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3999 [Rob Stradling <rob.stradling@comodo.com>]
4001 *) Fix possible deadlock when decoding public keys.
4004 *) Don't use TLS 1.0 record version number in initial client hello
4008 Changes between 1.0.1b and 1.0.1c [10 May 2012]
4010 *) Sanity check record length before skipping explicit IV in TLS
4011 1.2, 1.1 and DTLS to fix DoS attack.
4013 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
4014 fuzzing as a service testing platform.
4018 *) Initialise tkeylen properly when encrypting CMS messages.
4019 Thanks to Solar Designer of Openwall for reporting this issue.
4022 *) In FIPS mode don't try to use composite ciphers as they are not
4026 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
4028 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
4029 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
4030 mean any application compiled against OpenSSL 1.0.0 headers setting
4031 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
4032 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
4033 0x10000000L Any application which was previously compiled against
4034 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
4035 will need to be recompiled as a result. Letting be results in
4036 inability to disable specifically TLS 1.1 and in client context,
4037 in unlike event, limit maximum offered version to TLS 1.0 [see below].
4040 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
4041 disable just protocol X, but all protocols above X *if* there are
4042 protocols *below* X still enabled. In more practical terms it means
4043 that if application wants to disable TLS1.0 in favor of TLS1.1 and
4044 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
4045 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
4049 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
4051 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
4052 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
4053 in CRYPTO_realloc_clean.
4055 Thanks to Tavis Ormandy, Google Security Team, for discovering this
4056 issue and to Adam Langley <agl@chromium.org> for fixing it.
4058 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
4060 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
4063 *) Workarounds for some broken servers that "hang" if a client hello
4064 record length exceeds 255 bytes.
4066 1. Do not use record version number > TLS 1.0 in initial client
4067 hello: some (but not all) hanging servers will now work.
4068 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
4069 the number of ciphers sent in the client hello. This should be
4070 set to an even number, such as 50, for example by passing:
4071 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
4072 Most broken servers should now work.
4073 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
4074 TLS 1.2 client support entirely.
4077 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
4080 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
4082 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
4083 STRING form instead of a DigestInfo.
4086 *) The format used for MDC2 RSA signatures is inconsistent between EVP
4087 and the RSA_sign/RSA_verify functions. This was made more apparent when
4088 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
4089 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
4090 the correct format in RSA_verify so both forms transparently work.
4093 *) Some servers which support TLS 1.0 can choke if we initially indicate
4094 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
4095 encrypted premaster secret. As a workaround use the maximum permitted
4096 client version in client hello, this should keep such servers happy
4097 and still work with previous versions of OpenSSL.
4100 *) Add support for TLS/DTLS heartbeats.
4101 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4103 *) Add support for SCTP.
4104 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4106 *) Improved PRNG seeding for VOS.
4107 [Paul Green <Paul.Green@stratus.com>]
4109 *) Extensive assembler packs updates, most notably:
4111 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
4112 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
4113 - x86_64: bit-sliced AES implementation;
4114 - ARM: NEON support, contemporary platforms optimizations;
4115 - s390x: z196 support;
4116 - *: GHASH and GF(2^m) multiplication implementations;
4120 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
4121 (removal of unnecessary code)
4122 [Peter Sylvester <peter.sylvester@edelweb.fr>]
4124 *) Add TLS key material exporter from RFC 5705.
4127 *) Add DTLS-SRTP negotiation from RFC 5764.
4130 *) Add Next Protocol Negotiation,
4131 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
4132 disabled with a no-npn flag to config or Configure. Code donated
4134 [Adam Langley <agl@google.com> and Ben Laurie]
4136 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
4137 NIST-P256, NIST-P521, with constant-time single point multiplication on
4138 typical inputs. Compiler support for the nonstandard type __uint128_t is
4139 required to use this (present in gcc 4.4 and later, for 64-bit builds).
4140 Code made available under Apache License version 2.0.
4142 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
4143 line to include this in your build of OpenSSL, and run "make depend" (or
4144 "make update"). This enables the following EC_METHODs:
4146 EC_GFp_nistp224_method()
4147 EC_GFp_nistp256_method()
4148 EC_GFp_nistp521_method()
4150 EC_GROUP_new_by_curve_name() will automatically use these (while
4151 EC_GROUP_new_curve_GFp() currently prefers the more flexible
4153 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
4155 *) Use type ossl_ssize_t instead of ssize_t which isn't available on
4156 all platforms. Move ssize_t definition from e_os.h to the public
4157 header file e_os2.h as it now appears in public header file cms.h
4160 *) New -sigopt option to the ca, req and x509 utilities. Additional
4161 signature parameters can be passed using this option and in
4165 *) Add RSA PSS signing function. This will generate and set the
4166 appropriate AlgorithmIdentifiers for PSS based on those in the
4167 corresponding EVP_MD_CTX structure. No application support yet.
4170 *) Support for companion algorithm specific ASN1 signing routines.
4171 New function ASN1_item_sign_ctx() signs a pre-initialised
4172 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
4173 the appropriate parameters.
4176 *) Add new algorithm specific ASN1 verification initialisation function
4177 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
4178 handling will be the same no matter what EVP_PKEY_METHOD is used.
4179 Add a PSS handler to support verification of PSS signatures: checked
4180 against a number of sample certificates.
4183 *) Add signature printing for PSS. Add PSS OIDs.
4184 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
4186 *) Add algorithm specific signature printing. An individual ASN1 method
4187 can now print out signatures instead of the standard hex dump.
4189 More complex signatures (e.g. PSS) can print out more meaningful
4190 information. Include DSA version that prints out the signature
4194 *) Password based recipient info support for CMS library: implementing
4198 *) Split password based encryption into PBES2 and PBKDF2 functions. This
4199 neatly separates the code into cipher and PBE sections and is required
4200 for some algorithms that split PBES2 into separate pieces (such as
4201 password based CMS).
4204 *) Session-handling fixes:
4205 - Fix handling of connections that are resuming with a session ID,
4206 but also support Session Tickets.
4207 - Fix a bug that suppressed issuing of a new ticket if the client
4208 presented a ticket with an expired session.
4209 - Try to set the ticket lifetime hint to something reasonable.
4210 - Make tickets shorter by excluding irrelevant information.
4211 - On the client side, don't ignore renewed tickets.
4212 [Adam Langley, Bodo Moeller (Google)]
4214 *) Fix PSK session representation.
4217 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
4219 This work was sponsored by Intel.
4222 *) Add GCM support to TLS library. Some custom code is needed to split
4223 the IV between the fixed (from PRF) and explicit (from TLS record)
4224 portions. This adds all GCM ciphersuites supported by RFC5288 and
4225 RFC5289. Generalise some AES* cipherstrings to include GCM and
4226 add a special AESGCM string for GCM only.
4229 *) Expand range of ctrls for AES GCM. Permit setting invocation
4230 field on decrypt and retrieval of invocation field only on encrypt.
4233 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
4234 As required by RFC5289 these ciphersuites cannot be used if for
4235 versions of TLS earlier than 1.2.
4238 *) For FIPS capable OpenSSL interpret a NULL default public key method
4239 as unset and return the appropriate default but do *not* set the default.
4240 This means we can return the appropriate method in applications that
4241 switch between FIPS and non-FIPS modes.
4244 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
4245 ENGINE is used then we cannot handle that in the FIPS module so we
4246 keep original code iff non-FIPS operations are allowed.
4249 *) Add -attime option to openssl utilities.
4250 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
4252 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
4255 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
4256 FIPS EC methods unconditionally for now.
4259 *) New build option no-ec2m to disable characteristic 2 code.
4262 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
4263 all cases can be covered as some introduce binary incompatibilities.
4266 *) Redirect RSA operations to FIPS module including keygen,
4267 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
4270 *) Add similar low level API blocking to ciphers.
4273 *) Low level digest APIs are not approved in FIPS mode: any attempt
4274 to use these will cause a fatal error. Applications that *really* want
4275 to use them can use the private_* version instead.
4278 *) Redirect cipher operations to FIPS module for FIPS builds.
4281 *) Redirect digest operations to FIPS module for FIPS builds.
4284 *) Update build system to add "fips" flag which will link in fipscanister.o
4285 for static and shared library builds embedding a signature if needed.
4288 *) Output TLS supported curves in preference order instead of numerical
4289 order. This is currently hardcoded for the highest order curves first.
4290 This should be configurable so applications can judge speed vs strength.
4293 *) Add TLS v1.2 server support for client authentication.
4296 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4300 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4301 FIPS modules versions.
4304 *) Add TLS v1.2 client side support for client authentication. Keep cache
4305 of handshake records longer as we don't know the hash algorithm to use
4306 until after the certificate request message is received.
4309 *) Initial TLS v1.2 client support. Add a default signature algorithms
4310 extension including all the algorithms we support. Parse new signature
4311 format in client key exchange. Relax some ECC signing restrictions for
4312 TLS v1.2 as indicated in RFC5246.
4315 *) Add server support for TLS v1.2 signature algorithms extension. Switch
4316 to new signature format when needed using client digest preference.
4317 All server ciphersuites should now work correctly in TLS v1.2. No client
4318 support yet and no support for client certificates.
4321 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4322 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4323 ciphersuites. At present only RSA key exchange ciphersuites work with
4324 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4325 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4326 and version checking.
4329 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4330 with this defined it will not be affected by any changes to ssl internal
4331 structures. Add several utility functions to allow openssl application
4332 to work with OPENSSL_NO_SSL_INTERN defined.
4335 *) A long standing patch to add support for SRP from EdelWeb (Peter
4336 Sylvester and Christophe Renou) was integrated.
4337 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4338 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4341 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4344 *) Permit abbreviated handshakes when renegotiating using the function
4345 SSL_renegotiate_abbreviated().
4346 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4348 *) Add call to ENGINE_register_all_complete() to
4349 ENGINE_load_builtin_engines(), so some implementations get used
4350 automatically instead of needing explicit application support.
4353 *) Add support for TLS key exporter as described in RFC5705.
4354 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4356 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4357 a few changes are required:
4359 Add SSL_OP_NO_TLSv1_1 flag.
4360 Add TLSv1_1 methods.
4361 Update version checking logic to handle version 1.1.
4362 Add explicit IV handling (ported from DTLS code).
4363 Add command line options to s_client/s_server.
4366 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4368 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4369 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4370 content decryption and always return the same error. Note: this attack
4371 needs on average 2^20 messages so it only affects automated senders. The
4372 old behaviour can be re-enabled in the CMS code by setting the
4373 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4374 an MMA defence is not necessary.
4375 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4376 this issue. (CVE-2012-0884)
4379 *) Fix CVE-2011-4619: make sure we really are receiving a
4380 client hello before rejecting multiple SGC restarts. Thanks to
4381 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4384 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4386 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4387 Thanks to Antonio Martin, Enterprise Secure Access Research and
4388 Development, Cisco Systems, Inc. for discovering this bug and
4389 preparing a fix. (CVE-2012-0050)
4392 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4394 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4395 of the Vaudenay padding oracle attack on CBC mode encryption
4396 which enables an efficient plaintext recovery attack against
4397 the OpenSSL implementation of DTLS. Their attack exploits timing
4398 differences arising during decryption processing. A research
4399 paper describing this attack can be found at:
4400 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4401 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4402 Security Group at Royal Holloway, University of London
4403 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4404 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4405 for preparing the fix. (CVE-2011-4108)
4406 [Robin Seggelmann, Michael Tuexen]
4408 *) Clear bytes used for block padding of SSL 3.0 records.
4410 [Adam Langley (Google)]
4412 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4413 Kadianakis <desnacked@gmail.com> for discovering this issue and
4414 Adam Langley for preparing the fix. (CVE-2011-4619)
4415 [Adam Langley (Google)]
4417 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4418 [Andrey Kulikov <amdeich@gmail.com>]
4420 *) Prevent malformed RFC3779 data triggering an assertion failure.
4421 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4422 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4423 [Rob Austein <sra@hactrn.net>]
4425 *) Improved PRNG seeding for VOS.
4426 [Paul Green <Paul.Green@stratus.com>]
4428 *) Fix ssl_ciph.c set-up race.
4429 [Adam Langley (Google)]
4431 *) Fix spurious failures in ecdsatest.c.
4432 [Emilia Käsper (Google)]
4434 *) Fix the BIO_f_buffer() implementation (which was mixing different
4435 interpretations of the '..._len' fields).
4436 [Adam Langley (Google)]
4438 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4439 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4440 threads won't reuse the same blinding coefficients.
4442 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4443 lock to call BN_BLINDING_invert_ex, and avoids one use of
4444 BN_BLINDING_update for each BN_BLINDING structure (previously,
4445 the last update always remained unused).
4446 [Emilia Käsper (Google)]
4448 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4449 [Bob Buckholz (Google)]
4451 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4453 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4454 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4455 [Kaspar Brand <ossl@velox.ch>]
4457 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4458 for multi-threaded use of ECDH. (CVE-2011-3210)
4459 [Adam Langley (Google)]
4461 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4464 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4465 signature public key algorithm by using OID xref utilities instead.
4466 Before this you could only use some ECC ciphersuites with SHA1 only.
4469 *) Add protection against ECDSA timing attacks as mentioned in the paper
4470 by Billy Bob Brumley and Nicola Tuveri, see:
4472 http://eprint.iacr.org/2011/232.pdf
4474 [Billy Bob Brumley and Nicola Tuveri]
4476 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4478 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4479 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4481 *) Fix bug in string printing code: if *any* escaping is enabled we must
4482 escape the escape character (backslash) or the resulting string is
4486 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4488 *) Disable code workaround for ancient and obsolete Netscape browsers
4489 and servers: an attacker can use it in a ciphersuite downgrade attack.
4490 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4493 *) Fixed J-PAKE implementation error, originally discovered by
4494 Sebastien Martini, further info and confirmation from Stefan
4495 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4498 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4500 *) Fix extension code to avoid race conditions which can result in a buffer
4501 overrun vulnerability: resumed sessions must not be modified as they can
4502 be shared by multiple threads. CVE-2010-3864
4505 *) Fix WIN32 build system to correctly link an ENGINE directory into
4509 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4511 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4513 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4515 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4517 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4518 context. The operation can be customised via the ctrl mechanism in
4519 case ENGINEs want to include additional functionality.
4522 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4525 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4526 output hashes compatible with older versions of OpenSSL.
4527 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4529 *) Fix compression algorithm handling: if resuming a session use the
4530 compression algorithm of the resumed session instead of determining
4531 it from client hello again. Don't allow server to change algorithm.
4534 *) Add load_crls() function to apps tidying load_certs() too. Add option
4535 to verify utility to allow additional CRLs to be included.
4538 *) Update OCSP request code to permit adding custom headers to the request:
4539 some responders need this.
4542 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4544 [Julia Lawall <julia@diku.dk>]
4546 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4547 needlessly dereferenced structures, used obsolete functions and
4548 didn't handle all updated verify codes correctly.
4551 *) Disable MD2 in the default configuration.
4554 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4555 indicate the initial BIO being pushed or popped. This makes it possible
4556 to determine whether the BIO is the one explicitly called or as a result
4557 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4558 it handles reference counts correctly and doesn't zero out the I/O bio
4559 when it is not being explicitly popped. WARNING: applications which
4560 included workarounds for the old buggy behaviour will need to be modified
4561 or they could free up already freed BIOs.
4564 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4565 renaming to all platforms (within the 0.9.8 branch, this was
4566 done conditionally on Netware platforms to avoid a name clash).
4567 [Guenter <lists@gknw.net>]
4569 *) Add ECDHE and PSK support to DTLS.
4570 [Michael Tuexen <tuexen@fh-muenster.de>]
4572 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4576 *) Add "missing" function EVP_MD_flags() (without this the only way to
4577 retrieve a digest flags is by accessing the structure directly. Update
4578 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4579 or cipher is registered as in the "from" argument. Print out all
4580 registered digests in the dgst usage message instead of manually
4581 attempting to work them out.
4584 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4585 this allows the use of compression and extensions. Change default cipher
4586 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4587 by default unless an application cipher string requests it.
4590 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4591 key ids to find matching certificates and keys but some PKCS#12 files
4592 don't follow the (somewhat unwritten) rules and this strategy fails.
4593 Now just gather all certificates together and the first private key
4594 then look for the first certificate that matches the key.
4597 *) Support use of registered digest and cipher names for dgst and cipher
4598 commands instead of having to add each one as a special case. So now
4605 openssl dgst -sha256 foo
4607 and this works for ENGINE based algorithms too.
4611 *) Update Gost ENGINE to support parameter files.
4612 [Victor B. Wagner <vitus@cryptocom.ru>]
4614 *) Support GeneralizedTime in ca utility.
4615 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4617 *) Enhance the hash format used for certificate directory links. The new
4618 form uses the canonical encoding (meaning equivalent names will work
4619 even if they aren't identical) and uses SHA1 instead of MD5. This form
4620 is incompatible with the older format and as a result c_rehash should
4621 be used to rebuild symbolic links.
4624 *) Make PKCS#8 the default write format for private keys, replacing the
4625 traditional format. This form is standardised, more secure and doesn't
4626 include an implicit MD5 dependency.
4629 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4630 committed to OpenSSL should pass this lot as a minimum.
4633 *) Add session ticket override functionality for use by EAP-FAST.
4634 [Jouni Malinen <j@w1.fi>]
4636 *) Modify HMAC functions to return a value. Since these can be implemented
4637 in an ENGINE errors can occur.
4640 *) Type-checked OBJ_bsearch_ex.
4643 *) Type-checked OBJ_bsearch. Also some constification necessitated
4644 by type-checking. Still to come: TXT_DB, bsearch(?),
4645 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4649 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4650 seconds to a tm structure directly, instead of going through OS
4651 specific date routines. This avoids any issues with OS routines such
4652 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4653 and X509_time_adj_ex() to cover the extended range. The existing
4654 X509_time_adj() is still usable and will no longer have any date issues.
4657 *) Delta CRL support. New use deltas option which will attempt to locate
4658 and search any appropriate delta CRLs available.
4660 This work was sponsored by Google.
4663 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4664 code and add additional score elements. Validate alternate CRL paths
4665 as part of the CRL checking and indicate a new error "CRL path validation
4666 error" in this case. Applications wanting additional details can use
4667 the verify callback and check the new "parent" field. If this is not
4668 NULL CRL path validation is taking place. Existing applications won't
4669 see this because it requires extended CRL support which is off by
4672 This work was sponsored by Google.
4675 *) Support for freshest CRL extension.
4677 This work was sponsored by Google.
4680 *) Initial indirect CRL support. Currently only supported in the CRLs
4681 passed directly and not via lookup. Process certificate issuer
4682 CRL entry extension and lookup CRL entries by bother issuer name
4683 and serial number. Check and process CRL issuer entry in IDP extension.
4685 This work was sponsored by Google.
4688 *) Add support for distinct certificate and CRL paths. The CRL issuer
4689 certificate is validated separately in this case. Only enabled if
4690 an extended CRL support flag is set: this flag will enable additional
4691 CRL functionality in future.
4693 This work was sponsored by Google.
4696 *) Add support for policy mappings extension.
4698 This work was sponsored by Google.
4701 *) Fixes to pathlength constraint, self issued certificate handling,
4702 policy processing to align with RFC3280 and PKITS tests.
4704 This work was sponsored by Google.
4707 *) Support for name constraints certificate extension. DN, email, DNS
4708 and URI types are currently supported.
4710 This work was sponsored by Google.
4713 *) To cater for systems that provide a pointer-based thread ID rather
4714 than numeric, deprecate the current numeric thread ID mechanism and
4715 replace it with a structure and associated callback type. This
4716 mechanism allows a numeric "hash" to be extracted from a thread ID in
4717 either case, and on platforms where pointers are larger than 'long',
4718 mixing is done to help ensure the numeric 'hash' is usable even if it
4719 can't be guaranteed unique. The default mechanism is to use "&errno"
4720 as a pointer-based thread ID to distinguish between threads.
4722 Applications that want to provide their own thread IDs should now use
4723 CRYPTO_THREADID_set_callback() to register a callback that will call
4724 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4726 Note that ERR_remove_state() is now deprecated, because it is tied
4727 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4728 to free the current thread's error state should be replaced by
4729 ERR_remove_thread_state(NULL).
4731 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4732 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4733 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4734 application was previously providing a numeric thread callback that
4735 was inappropriate for distinguishing threads, then uniqueness might
4736 have been obtained with &errno that happened immediately in the
4737 intermediate development versions of OpenSSL; this is no longer the
4738 case, the numeric thread callback will now override the automatic use
4740 [Geoff Thorpe, with help from Bodo Moeller]
4742 *) Initial support for different CRL issuing certificates. This covers a
4743 simple case where the self issued certificates in the chain exist and
4744 the real CRL issuer is higher in the existing chain.
4746 This work was sponsored by Google.
4749 *) Removed effectively defunct crypto/store from the build.
4752 *) Revamp of STACK to provide stronger type-checking. Still to come:
4753 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4754 ASN1_STRING, CONF_VALUE.
4757 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4758 RAM on SSL connections. This option can save about 34k per idle SSL.
4761 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4762 STACK, TXT_DB, bsearch, qsort.
4765 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4766 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4767 support for data, signedData, compressedData, digestedData and
4768 encryptedData, envelopedData types included. Scripts to check against
4769 RFC4134 examples draft and interop and consistency checks of many
4770 content types and variants.
4773 *) Add options to enc utility to support use of zlib compression BIO.
4776 *) Extend mk1mf to support importing of options and assembly language
4777 files from Configure script, currently only included in VC-WIN32.
4778 The assembly language rules can now optionally generate the source
4779 files from the associated perl scripts.
4782 *) Implement remaining functionality needed to support GOST ciphersuites.
4783 Interop testing has been performed using CryptoPro implementations.
4784 [Victor B. Wagner <vitus@cryptocom.ru>]
4786 *) s390x assembler pack.
4789 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4793 *) Implement Opaque PRF Input TLS extension as specified in
4794 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4795 official specification yet and no extension type assignment by
4796 IANA exists, this extension (for now) will have to be explicitly
4797 enabled when building OpenSSL by providing the extension number
4798 to use. For example, specify an option
4800 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4802 to the "config" or "Configure" script to enable the extension,
4803 assuming extension number 0x9527 (which is a completely arbitrary
4804 and unofficial assignment based on the MD5 hash of the Internet
4805 Draft). Note that by doing so, you potentially lose
4806 interoperability with other TLS implementations since these might
4807 be using the same extension number for other purposes.
4809 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4810 opaque PRF input value to use in the handshake. This will create
4811 an internal copy of the length-'len' string at 'src', and will
4812 return non-zero for success.
4814 To get more control and flexibility, provide a callback function
4817 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4818 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4822 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4825 Callback function 'cb' will be called in handshakes, and is
4826 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4827 Argument 'arg' is for application purposes (the value as given to
4828 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4829 be provided to the callback function). The callback function
4830 has to return non-zero to report success: usually 1 to use opaque
4831 PRF input just if possible, or 2 to enforce use of the opaque PRF
4832 input. In the latter case, the library will abort the handshake
4833 if opaque PRF input is not successfully negotiated.
4835 Arguments 'peerinput' and 'len' given to the callback function
4836 will always be NULL and 0 in the case of a client. A server will
4837 see the client's opaque PRF input through these variables if
4838 available (NULL and 0 otherwise). Note that if the server
4839 provides an opaque PRF input, the length must be the same as the
4840 length of the client's opaque PRF input.
4842 Note that the callback function will only be called when creating
4843 a new session (session resumption can resume whatever was
4844 previously negotiated), and will not be called in SSL 2.0
4845 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4846 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4847 for applications that need to enforce opaque PRF input.
4851 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4854 [Victor B. Wagner <vitus@cryptocom.ru>]
4856 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4857 RFC4507bis. The encrypted ticket format is an encrypted encoded
4858 SSL_SESSION structure, that way new session features are automatically
4861 If a client application caches session in an SSL_SESSION structure
4862 support is transparent because tickets are now stored in the encoded
4865 The SSL_CTX structure automatically generates keys for ticket
4866 protection in servers so again support should be possible
4867 with no application modification.
4869 If a client or server wishes to disable RFC4507 support then the option
4870 SSL_OP_NO_TICKET can be set.
4872 Add a TLS extension debugging callback to allow the contents of any client
4873 or server extensions to be examined.
4875 This work was sponsored by Google.
4878 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4879 OpenSSL should now compile cleanly on gcc 4.2
4880 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4882 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4883 support including streaming MAC support: this is required for GOST
4884 ciphersuite support.
4885 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4887 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4888 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4889 to output in BER and PEM format.
4892 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4893 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4894 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4895 ENGINE support for HMAC keys which are unextractable. New -mac and
4896 -macopt options to dgst utility.
4899 *) New option -sigopt to dgst utility. Update dgst to use
4900 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4901 alternative signing parameters such as X9.31 or PSS in the dgst
4905 *) Change ssl_cipher_apply_rule(), the internal function that does
4906 the work each time a ciphersuite string requests enabling
4907 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4908 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4909 the order of disabled ciphersuites such that those ciphersuites
4910 that most recently went from enabled to disabled not only stay
4911 in order with respect to each other, but also have higher priority
4912 than other disabled ciphersuites the next time ciphersuites are
4915 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4916 the same ciphersuites as with "HIGH" alone, but in a specific
4917 order where the PSK ciphersuites come first (since they are the
4918 most recently disabled ciphersuites when "HIGH" is parsed).
4920 Also, change ssl_create_cipher_list() (using this new
4921 functionality) such that between otherwise identical
4922 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4926 *) Change ssl_create_cipher_list() so that it automatically
4927 arranges the ciphersuites in reasonable order before starting
4928 to process the rule string. Thus, the definition for "DEFAULT"
4929 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4930 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4931 This makes it much easier to arrive at a reasonable default order
4932 in applications for which anonymous ciphers are OK (meaning
4933 that you can't actually use DEFAULT).
4934 [Bodo Moeller; suggested by Victor Duchovni]
4936 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4937 processing) into multiple integers instead of setting
4938 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4939 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4940 (These masks as well as the individual bit definitions are hidden
4941 away into the non-exported interface ssl/ssl_locl.h, so this
4942 change to the definition of the SSL_CIPHER structure shouldn't
4943 affect applications.) This give us more bits for each of these
4944 categories, so there is no longer a need to coagulate AES128 and
4945 AES256 into a single algorithm bit, and to coagulate Camellia128
4946 and Camellia256 into a single algorithm bit, which has led to all
4949 Thus, among other things, the kludge introduced in 0.9.7m and
4950 0.9.8e for masking out AES256 independently of AES128 or masking
4951 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4953 With the change, we also introduce new ciphersuite aliases that
4954 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4958 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4959 Use the leftmost N bytes of the signature input if the input is
4960 larger than the prime q (with N being the size in bytes of q).
4963 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4964 it yet and it is largely untested.
4967 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4970 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4971 some compilers (gcc 4.2 and later) reject their use. Safestack is
4972 reimplemented. Update ASN1 to avoid use of legacy functions.
4975 *) Win32/64 targets are linked with Winsock2.
4978 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4979 to external functions. This can be used to increase CRL handling
4980 efficiency especially when CRLs are very large by (for example) storing
4981 the CRL revoked certificates in a database.
4984 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4985 new CRLs added to a directory can be used. New command line option
4986 -verify_return_error to s_client and s_server. This causes real errors
4987 to be returned by the verify callback instead of carrying on no matter
4988 what. This reflects the way a "real world" verify callback would behave.
4991 *) GOST engine, supporting several GOST algorithms and public key formats.
4992 Kindly donated by Cryptocom.
4995 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4996 partitioned by DP are handled but no indirect CRL or reason partitioning
4997 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4998 selected via a scoring technique which handles IDP and AKID in CRLs.
5001 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
5002 will ultimately be used for all verify operations: this will remove the
5003 X509_STORE dependency on certificate verification and allow alternative
5004 lookup methods. X509_STORE based implementations of these two callbacks.
5007 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
5008 Modify get_crl() to find a valid (unexpired) CRL if possible.
5011 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
5012 this would be called X509_CRL_cmp() but that name is already used by
5013 a function that just compares CRL issuer names. Cache several CRL
5014 extensions in X509_CRL structure and cache CRLDP in X509.
5017 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
5018 this maps equivalent X509_NAME structures into a consistent structure.
5019 Name comparison can then be performed rapidly using memcmp().
5022 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
5026 *) Allow digests to supply their own micalg string for S/MIME type using
5027 the ctrl EVP_MD_CTRL_MICALG.
5030 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
5031 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
5032 ctrl. It can then customise the structure before and/or after signing
5036 *) New function OBJ_add_sigid() to allow application defined signature OIDs
5037 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
5038 to free up any added signature OIDs.
5041 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
5042 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
5043 digest and cipher tables. New options added to openssl utility:
5044 list-message-digest-algorithms and list-cipher-algorithms.
5047 *) Change the array representation of binary polynomials: the list
5048 of degrees of non-zero coefficients is now terminated with -1.
5049 Previously it was terminated with 0, which was also part of the
5050 value; thus, the array representation was not applicable to
5051 polynomials where t^0 has coefficient zero. This change makes
5052 the array representation useful in a more general context.
5055 *) Various modifications and fixes to SSL/TLS cipher string
5056 handling. For ECC, the code now distinguishes between fixed ECDH
5057 with RSA certificates on the one hand and with ECDSA certificates
5058 on the other hand, since these are separate ciphersuites. The
5059 unused code for Fortezza ciphersuites has been removed.
5061 For consistency with EDH, ephemeral ECDH is now called "EECDH"
5062 (not "ECDHE"). For consistency with the code for DH
5063 certificates, use of ECDH certificates is now considered ECDH
5064 authentication, not RSA or ECDSA authentication (the latter is
5065 merely the CA's signing algorithm and not actively used in the
5068 The temporary ciphersuite alias "ECCdraft" is no longer
5069 available, and ECC ciphersuites are no longer excluded from "ALL"
5070 and "DEFAULT". The following aliases now exist for RFC 4492
5071 ciphersuites, most of these by analogy with the DH case:
5073 kECDHr - ECDH cert, signed with RSA
5074 kECDHe - ECDH cert, signed with ECDSA
5075 kECDH - ECDH cert (signed with either RSA or ECDSA)
5076 kEECDH - ephemeral ECDH
5077 ECDH - ECDH cert or ephemeral ECDH
5083 AECDH - anonymous ECDH
5084 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
5088 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
5089 Use correct micalg parameters depending on digest(s) in signed message.
5092 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
5093 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
5096 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
5097 an engine to register a method. Add ENGINE lookups for methods and
5098 functional reference processing.
5101 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
5102 EVP_{Sign,Verify}* which allow an application to customise the signature
5106 *) New -resign option to smime utility. This adds one or more signers
5107 to an existing PKCS#7 signedData structure. Also -md option to use an
5108 alternative message digest algorithm for signing.
5111 *) Tidy up PKCS#7 routines and add new functions to make it easier to
5112 create PKCS7 structures containing multiple signers. Update smime
5113 application to support multiple signers.
5116 *) New -macalg option to pkcs12 utility to allow setting of an alternative
5120 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
5121 Reorganize PBE internals to lookup from a static table using NIDs,
5122 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
5123 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
5124 PRF which will be automatically used with PBES2.
5127 *) Replace the algorithm specific calls to generate keys in "req" with the
5131 *) Update PKCS#7 enveloped data routines to use new API. This is now
5132 supported by any public key method supporting the encrypt operation. A
5133 ctrl is added to allow the public key algorithm to examine or modify
5134 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
5138 *) Add a ctrl to asn1 method to allow a public key algorithm to express
5139 a default digest type to use. In most cases this will be SHA1 but some
5140 algorithms (such as GOST) need to specify an alternative digest. The
5141 return value indicates how strong the preference is 1 means optional and
5142 2 is mandatory (that is it is the only supported type). Modify
5143 ASN1_item_sign() to accept a NULL digest argument to indicate it should
5144 use the default md. Update openssl utilities to use the default digest
5145 type for signing if it is not explicitly indicated.
5148 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
5149 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
5150 signing method from the key type. This effectively removes the link
5151 between digests and public key types.
5154 *) Add an OID cross reference table and utility functions. Its purpose is to
5155 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
5156 rsaEncryption. This will allow some of the algorithm specific hackery
5157 needed to use the correct OID to be removed.
5160 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
5161 structures for PKCS7_sign(). They are now set up by the relevant public
5165 *) Add provisional EC pkey method with support for ECDSA and ECDH.
5168 *) Add support for key derivation (agreement) in the API, DH method and
5172 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
5173 public and private key formats. As a side effect these add additional
5174 command line functionality not previously available: DSA signatures can be
5175 generated and verified using pkeyutl and DH key support and generation in
5180 [Oliver Tappe <zooey@hirschkaefer.de>]
5182 *) New make target "install_html_docs" installs HTML renditions of the
5184 [Oliver Tappe <zooey@hirschkaefer.de>]
5186 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
5187 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
5188 support key and parameter generation and add initial key generation
5189 functionality for RSA.
5192 *) Add functions for main EVP_PKEY_method operations. The undocumented
5193 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
5194 EVP_PKEY_{encrypt,decrypt}_old.
5197 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
5198 key API, doesn't do much yet.
5201 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
5202 public key algorithms. New option to openssl utility:
5203 "list-public-key-algorithms" to print out info.
5206 *) Implement the Supported Elliptic Curves Extension for
5207 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5210 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
5211 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
5214 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
5215 utilities such as rsa, dsa, dsaparam etc except they process any key
5219 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
5220 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
5221 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
5225 *) Initial support for pluggable public key ASN1.
5226 De-spaghettify the public key ASN1 handling. Move public and private
5227 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
5228 algorithm specific handling to a single module within the relevant
5229 algorithm directory. Add functions to allow (near) opaque processing
5230 of public and private key structures.
5233 *) Implement the Supported Point Formats Extension for
5234 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
5237 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
5238 for the psk identity [hint] and the psk callback functions to the
5239 SSL_SESSION, SSL and SSL_CTX structure.
5242 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
5246 SSL_CTX_use_psk_identity_hint
5247 SSL_get_psk_identity_hint
5248 SSL_get_psk_identity
5249 SSL_use_psk_identity_hint
5251 [Mika Kousa and Pasi Eronen of Nokia Corporation]
5253 *) Add RFC 3161 compliant time stamp request creation, response generation
5254 and response verification functionality.
5255 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
5257 *) Add initial support for TLS extensions, specifically for the server_name
5258 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5259 have new members for a host name. The SSL data structure has an
5260 additional member SSL_CTX *initial_ctx so that new sessions can be
5261 stored in that context to allow for session resumption, even after the
5262 SSL has been switched to a new SSL_CTX in reaction to a client's
5263 server_name extension.
5265 New functions (subject to change):
5267 SSL_get_servername()
5268 SSL_get_servername_type()
5271 New CTRL codes and macros (subject to change):
5273 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5274 - SSL_CTX_set_tlsext_servername_callback()
5275 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5276 - SSL_CTX_set_tlsext_servername_arg()
5277 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5279 openssl s_client has a new '-servername ...' option.
5281 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5282 '-key2 ...', '-servername_fatal' (subject to change). This allows
5283 testing the HostName extension for a specific single host name ('-cert'
5284 and '-key' remain fallbacks for handshakes without HostName
5285 negotiation). If the unrecognized_name alert has to be sent, this by
5286 default is a warning; it becomes fatal with the '-servername_fatal'
5289 [Peter Sylvester, Remy Allais, Christophe Renou]
5291 *) Whirlpool hash implementation is added.
5294 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5295 bn(64,32). Because of instruction set limitations it doesn't have
5296 any negative impact on performance. This was done mostly in order
5297 to make it possible to share assembler modules, such as bn_mul_mont
5298 implementations, between 32- and 64-bit builds without hassle.
5301 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
5302 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5306 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
5307 dedicated Montgomery multiplication procedure, is introduced.
5308 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5309 "64-bit" performance on certain 32-bit targets.
5312 *) New option SSL_OP_NO_COMP to disable use of compression selectively
5313 in SSL structures. New SSL ctrl to set maximum send fragment size.
5314 Save memory by setting the I/O buffer sizes dynamically instead of
5315 using the maximum available value.
5318 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5319 in addition to the text details.
5322 *) Very, very preliminary EXPERIMENTAL support for printing of general
5323 ASN1 structures. This currently produces rather ugly output and doesn't
5324 handle several customised structures at all.
5327 *) Integrated support for PVK file format and some related formats such
5328 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5329 these in the 'rsa' and 'dsa' utilities.
5332 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5335 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5336 place for the (very old) "NETSCAPE" format certificates which are now
5337 handled using new ASN1 code equivalents.
5340 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5341 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5342 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5345 *) Modify CRL distribution points extension code to print out previously
5346 unsupported fields. Enhance extension setting code to allow setting of
5350 *) Add print and set support for Issuing Distribution Point CRL extension.
5353 *) Change 'Configure' script to enable Camellia by default.
5356 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5358 *) When rejecting SSL/TLS records due to an incorrect version number, never
5359 update s->server with a new major version number. As of
5360 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5361 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5362 the previous behavior could result in a read attempt at NULL when
5363 receiving specific incorrect SSL/TLS records once record payload
5364 protection is active. (CVE-2010-0740)
5365 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5367 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5368 could be crashed if the relevant tables were not present (e.g. chrooted).
5369 [Tomas Hoger <thoger@redhat.com>]
5371 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5373 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5374 [Martin Olsson, Neel Mehta]
5376 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5377 accommodate for stack sorting, always a write lock!).
5380 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5381 excessive delays in the RAND_poll(): over a minute. As a workaround
5382 include a time check in the inner Heap32Next loop too.
5385 *) The code that handled flushing of data in SSL/TLS originally used the
5386 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5387 the problem outlined in PR#1949. The fix suggested there however can
5388 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5389 of Apache). So instead simplify the code to flush unconditionally.
5390 This should be fine since flushing with no data to flush is a no op.
5393 *) Handle TLS versions 2.0 and later properly and correctly use the
5394 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5395 off ancient servers have a habit of sticking around for a while...
5398 *) Modify compression code so it frees up structures without using the
5399 ex_data callbacks. This works around a problem where some applications
5400 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5401 restarting) then use compression (e.g. SSL with compression) later.
5402 This results in significant per-connection memory leaks and
5403 has caused some security issues including CVE-2008-1678 and
5407 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5408 change when encrypting or decrypting.
5411 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5412 connect and renegotiate with servers which do not support RI.
5413 Until RI is more widely deployed this option is enabled by default.
5416 *) Add "missing" ssl ctrls to clear options and mode.
5419 *) If client attempts to renegotiate and doesn't support RI respond with
5420 a no_renegotiation alert as required by RFC5746. Some renegotiating
5421 TLS clients will continue a connection gracefully when they receive
5422 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5423 waiting for a server hello which it will never receive. Now we treat a
5424 received no_renegotiation alert as a fatal error. This is because
5425 applications requesting a renegotiation might well expect it to succeed
5426 and would have no code in place to handle the server denying it so the
5427 only safe thing to do is to terminate the connection.
5430 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5431 peer supports secure renegotiation and 0 otherwise. Print out peer
5432 renegotiation support in s_client/s_server.
5435 *) Replace the highly broken and deprecated SPKAC certification method with
5436 the updated NID creation version. This should correctly handle UTF8.
5439 *) Implement RFC5746. Re-enable renegotiation but require the extension
5440 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5441 turns out to be a bad idea. It has been replaced by
5442 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5443 SSL_CTX_set_options(). This is really not recommended unless you
5444 know what you are doing.
5445 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5447 *) Fixes to stateless session resumption handling. Use initial_ctx when
5448 issuing and attempting to decrypt tickets in case it has changed during
5449 servername handling. Use a non-zero length session ID when attempting
5450 stateless session resumption: this makes it possible to determine if
5451 a resumption has occurred immediately after receiving server hello
5452 (several places in OpenSSL subtly assume this) instead of later in
5456 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5457 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5458 fixes for a few places where the return code is not checked
5460 [Julia Lawall <julia@diku.dk>]
5462 *) Add --strict-warnings option to Configure script to include devteam
5463 warnings in other configurations.
5466 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5467 makes it possible to install openssl libraries in locations which
5468 have names other than "lib", for example "/usr/lib64" which some
5470 [Steve Henson, based on patch from Jeremy Utley]
5472 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5473 X690 8.9.12 and can produce some misleading textual output of OIDs.
5474 [Steve Henson, reported by Dan Kaminsky]
5476 *) Delete MD2 from algorithm tables. This follows the recommendation in
5477 several standards that it is not used in new applications due to
5478 several cryptographic weaknesses. For binary compatibility reasons
5479 the MD2 API is still compiled in by default.
5482 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5486 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5487 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5489 [Guenter <lists@gknw.net>]
5491 *) Fix the server certificate chain building code to use X509_verify_cert(),
5492 it used to have an ad-hoc builder which was unable to cope with anything
5493 other than a simple chain.
5494 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5496 *) Don't check self signed certificate signatures in X509_verify_cert()
5497 by default (a flag can override this): it just wastes time without
5498 adding any security. As a useful side effect self signed root CAs
5499 with non-FIPS digests are now usable in FIPS mode.
5502 *) In dtls1_process_out_of_seq_message() the check if the current message
5503 is already buffered was missing. For every new message was memory
5504 allocated, allowing an attacker to perform an denial of service attack
5505 with sending out of seq handshake messages until there is no memory
5506 left. Additionally every future message was buffered, even if the
5507 sequence number made no sense and would be part of another handshake.
5508 So only messages with sequence numbers less than 10 in advance will be
5509 buffered. (CVE-2009-1378)
5510 [Robin Seggelmann, discovered by Daniel Mentz]
5512 *) Records are buffered if they arrive with a future epoch to be
5513 processed after finishing the corresponding handshake. There is
5514 currently no limitation to this buffer allowing an attacker to perform
5515 a DOS attack with sending records with future epochs until there is no
5516 memory left. This patch adds the pqueue_size() function to determine
5517 the size of a buffer and limits the record buffer to 100 entries.
5519 [Robin Seggelmann, discovered by Daniel Mentz]
5521 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5522 parent structure is freed. (CVE-2009-1379)
5525 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5526 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5529 [Ilya O. <vrghost@gmail.com>]
5531 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5533 *) Disable renegotiation completely - this fixes a severe security
5534 problem (CVE-2009-3555) at the cost of breaking all
5535 renegotiation. Renegotiation can be re-enabled by setting
5536 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5537 run-time. This is really not recommended unless you know what
5541 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5543 *) Don't set val to NULL when freeing up structures, it is freed up by
5544 underlying code. If sizeof(void *) > sizeof(long) this can result in
5545 zeroing past the valid field. (CVE-2009-0789)
5546 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5548 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5549 checked correctly. This would allow some invalid signed attributes to
5550 appear to verify correctly. (CVE-2009-0591)
5551 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5553 *) Reject UniversalString and BMPString types with invalid lengths. This
5554 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5555 a legal length. (CVE-2009-0590)
5558 *) Set S/MIME signing as the default purpose rather than setting it
5559 unconditionally. This allows applications to override it at the store
5563 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5564 to handle some structures.
5567 *) Improve efficiency of mem_gets: don't search whole buffer each time
5569 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5571 *) New -hex option for openssl rand.
5574 *) Print out UTF8String and NumericString when parsing ASN1.
5577 *) Support NumericString type for name components.
5580 *) Allow CC in the environment to override the automatically chosen
5581 compiler. Note that nothing is done to ensure flags work with the
5585 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5587 *) Properly check EVP_VerifyFinal() and similar return values
5589 [Ben Laurie, Bodo Moeller, Google Security Team]
5591 *) Enable TLS extensions by default.
5594 *) Allow the CHIL engine to be loaded, whether the application is
5595 multithreaded or not. (This does not release the developer from the
5596 obligation to set up the dynamic locking callbacks.)
5597 [Sander Temme <sander@temme.net>]
5599 *) Use correct exit code if there is an error in dgst command.
5600 [Steve Henson; problem pointed out by Roland Dirlewanger]
5602 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5603 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5606 *) Add experimental JPAKE support, including demo authentication in
5607 s_client and s_server.
5610 *) Set the comparison function in v3_addr_canonize().
5611 [Rob Austein <sra@hactrn.net>]
5613 *) Add support for XMPP STARTTLS in s_client.
5614 [Philip Paeps <philip@freebsd.org>]
5616 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5617 to ensure that even with this option, only ciphersuites in the
5618 server's preference list will be accepted. (Note that the option
5619 applies only when resuming a session, so the earlier behavior was
5620 just about the algorithm choice for symmetric cryptography.)
5623 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5625 *) Fix NULL pointer dereference if a DTLS server received
5626 ChangeCipherSpec as first record (CVE-2009-1386).
5629 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5630 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5633 *) The fix in 0.9.8c that supposedly got rid of unsafe
5634 double-checked locking was incomplete for RSA blinding,
5635 addressing just one layer of what turns out to have been
5636 doubly unsafe triple-checked locking.
5638 So now fix this for real by retiring the MONT_HELPER macro
5639 in crypto/rsa/rsa_eay.c.
5641 [Bodo Moeller; problem pointed out by Marius Schilder]
5643 *) Various precautionary measures:
5645 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5647 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5648 (NB: This would require knowledge of the secret session ticket key
5649 to exploit, in which case you'd be SOL either way.)
5651 - Change bn_nist.c so that it will properly handle input BIGNUMs
5652 outside the expected range.
5654 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5657 [Neel Mehta, Bodo Moeller]
5659 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5660 the load fails. Useful for distros.
5661 [Ben Laurie and the FreeBSD team]
5663 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5666 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5669 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5671 This work was sponsored by Logica.
5674 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5675 keystores. Support for SSL/TLS client authentication too.
5676 Not compiled unless enable-capieng specified to Configure.
5678 This work was sponsored by Logica.
5681 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5682 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5683 attribute creation routines such as certificate requests and PKCS#12
5687 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5689 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5690 handshake which could lead to a client crash as found using the
5691 Codenomicon TLS test suite (CVE-2008-1672)
5692 [Steve Henson, Mark Cox]
5694 *) Fix double free in TLS server name extensions which could lead to
5695 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5698 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5700 Clear the error queue to ensure that error entries left from
5701 older function calls do not interfere with the correct operation.
5702 [Lutz Jaenicke, Erik de Castro Lopo]
5704 *) Remove root CA certificates of commercial CAs:
5706 The OpenSSL project does not recommend any specific CA and does not
5707 have any policy with respect to including or excluding any CA.
5708 Therefore it does not make any sense to ship an arbitrary selection
5709 of root CA certificates with the OpenSSL software.
5712 *) RSA OAEP patches to fix two separate invalid memory reads.
5713 The first one involves inputs when 'lzero' is greater than
5714 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5715 before the beginning of from). The second one involves inputs where
5716 the 'db' section contains nothing but zeroes (there is a one-byte
5717 invalid read after the end of 'db').
5718 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5720 *) Partial backport from 0.9.9-dev:
5722 Introduce bn_mul_mont (dedicated Montgomery multiplication
5723 procedure) as a candidate for BIGNUM assembler implementation.
5724 While 0.9.9-dev uses assembler for various architectures, only
5725 x86_64 is available by default here in the 0.9.8 branch, and
5726 32-bit x86 is available through a compile-time setting.
5728 To try the 32-bit x86 assembler implementation, use Configure
5729 option "enable-montasm" (which exists only for this backport).
5731 As "enable-montasm" for 32-bit x86 disclaims code stability
5732 anyway, in this constellation we activate additional code
5733 backported from 0.9.9-dev for further performance improvements,
5734 namely BN_from_montgomery_word. (To enable this otherwise,
5735 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5737 [Andy Polyakov (backport partially by Bodo Moeller)]
5739 *) Add TLS session ticket callback. This allows an application to set
5740 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5741 values. This is useful for key rollover for example where several key
5742 sets may exist with different names.
5745 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5746 This was broken until now in 0.9.8 releases, such that the only way
5747 a registered ENGINE could be used (assuming it initialises
5748 successfully on the host) was to explicitly set it as the default
5749 for the relevant algorithms. This is in contradiction with 0.9.7
5750 behaviour and the documentation. With this fix, when an ENGINE is
5751 registered into a given algorithm's table of implementations, the
5752 'uptodate' flag is reset so that auto-discovery will be used next
5753 time a new context for that algorithm attempts to select an
5755 [Ian Lister (tweaked by Geoff Thorpe)]
5757 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5758 implementation in the following ways:
5760 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5763 Lack of BER streaming support means one pass streaming processing is
5764 only supported if data is detached: setting the streaming flag is
5765 ignored for embedded content.
5767 CMS support is disabled by default and must be explicitly enabled
5768 with the enable-cms configuration option.
5771 *) Update the GMP engine glue to do direct copies between BIGNUM and
5772 mpz_t when openssl and GMP use the same limb size. Otherwise the
5773 existing "conversion via a text string export" trick is still used.
5774 [Paul Sheer <paulsheer@gmail.com>]
5776 *) Zlib compression BIO. This is a filter BIO which compressed and
5777 uncompresses any data passed through it.
5780 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5781 RFC3394 compatible AES key wrapping.
5784 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5785 sets string data without copying. X509_ALGOR_set0() and
5786 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5787 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5788 from an X509_ATTRIBUTE structure optionally checking it occurs only
5789 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5793 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5794 to get the expected BN_FLG_CONSTTIME behavior.
5795 [Bodo Moeller (Google)]
5799 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5800 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5801 - added some more tests to do_tests.pl
5802 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5803 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5804 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5805 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5806 - various changes to netware.pl to enable gcc-cross builds on Win32
5808 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5809 - various changes to fix missing prototype warnings
5810 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5811 - added AES, WHIRLPOOL and CPUID assembler code to build files
5812 - added missing AES assembler make rules to mk1mf.pl
5813 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5814 [Guenter Knauf <eflash@gmx.net>]
5816 *) Implement certificate status request TLS extension defined in RFC3546.
5817 A client can set the appropriate parameters and receive the encoded
5818 OCSP response via a callback. A server can query the supplied parameters
5819 and set the encoded OCSP response in the callback. Add simplified examples
5820 to s_client and s_server.
5823 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5825 *) Fix various bugs:
5826 + Binary incompatibility of ssl_ctx_st structure
5827 + DTLS interoperation with non-compliant servers
5828 + Don't call get_session_cb() without proposed session
5829 + Fix ia64 assembler code
5830 [Andy Polyakov, Steve Henson]
5832 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5834 *) DTLS Handshake overhaul. There were longstanding issues with
5835 OpenSSL DTLS implementation, which were making it impossible for
5836 RFC 4347 compliant client to communicate with OpenSSL server.
5837 Unfortunately just fixing these incompatibilities would "cut off"
5838 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5839 server keeps tolerating non RFC compliant syntax. The opposite is
5840 not true, 0.9.8f client can not communicate with earlier server.
5841 This update even addresses CVE-2007-4995.
5844 *) Changes to avoid need for function casts in OpenSSL: some compilers
5845 (gcc 4.2 and later) reject their use.
5846 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5849 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5850 RFC4507bis. The encrypted ticket format is an encrypted encoded
5851 SSL_SESSION structure, that way new session features are automatically
5854 If a client application caches session in an SSL_SESSION structure
5855 support is transparent because tickets are now stored in the encoded
5858 The SSL_CTX structure automatically generates keys for ticket
5859 protection in servers so again support should be possible
5860 with no application modification.
5862 If a client or server wishes to disable RFC4507 support then the option
5863 SSL_OP_NO_TICKET can be set.
5865 Add a TLS extension debugging callback to allow the contents of any client
5866 or server extensions to be examined.
5868 This work was sponsored by Google.
5871 *) Add initial support for TLS extensions, specifically for the server_name
5872 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5873 have new members for a host name. The SSL data structure has an
5874 additional member SSL_CTX *initial_ctx so that new sessions can be
5875 stored in that context to allow for session resumption, even after the
5876 SSL has been switched to a new SSL_CTX in reaction to a client's
5877 server_name extension.
5879 New functions (subject to change):
5881 SSL_get_servername()
5882 SSL_get_servername_type()
5885 New CTRL codes and macros (subject to change):
5887 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5888 - SSL_CTX_set_tlsext_servername_callback()
5889 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5890 - SSL_CTX_set_tlsext_servername_arg()
5891 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5893 openssl s_client has a new '-servername ...' option.
5895 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5896 '-key2 ...', '-servername_fatal' (subject to change). This allows
5897 testing the HostName extension for a specific single host name ('-cert'
5898 and '-key' remain fallbacks for handshakes without HostName
5899 negotiation). If the unrecognized_name alert has to be sent, this by
5900 default is a warning; it becomes fatal with the '-servername_fatal'
5903 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5905 *) Add AES and SSE2 assembly language support to VC++ build.
5908 *) Mitigate attack on final subtraction in Montgomery reduction.
5911 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5912 (which previously caused an internal error).
5915 *) Squeeze another 10% out of IGE mode when in != out.
5918 *) AES IGE mode speedup.
5919 [Dean Gaudet (Google)]
5921 *) Add the Korean symmetric 128-bit cipher SEED (see
5922 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5923 add SEED ciphersuites from RFC 4162:
5925 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5926 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5927 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5928 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5930 To minimize changes between patchlevels in the OpenSSL 0.9.8
5931 series, SEED remains excluded from compilation unless OpenSSL
5932 is configured with 'enable-seed'.
5933 [KISA, Bodo Moeller]
5935 *) Mitigate branch prediction attacks, which can be practical if a
5936 single processor is shared, allowing a spy process to extract
5937 information. For detailed background information, see
5938 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5939 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5940 and Necessary Software Countermeasures"). The core of the change
5941 are new versions BN_div_no_branch() and
5942 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5943 respectively, which are slower, but avoid the security-relevant
5944 conditional branches. These are automatically called by BN_div()
5945 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5946 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5947 remove a conditional branch.
5949 BN_FLG_CONSTTIME is the new name for the previous
5950 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5951 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5952 in the exponent causes BN_mod_exp_mont() to use the alternative
5953 implementation in BN_mod_exp_mont_consttime().) The old name
5954 remains as a deprecated alias.
5956 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5957 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5958 constant-time implementations for more than just exponentiation.
5959 Here too the old name is kept as a deprecated alias.
5961 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5962 the BN_BLINDING structure gets an independent copy of the
5963 modulus. This means that the previous "BIGNUM *m" argument to
5964 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5965 essentially becomes "const BIGNUM *m", although we can't actually
5966 change this in the header file before 0.9.9. It allows
5967 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5968 enable BN_FLG_CONSTTIME.
5970 [Matthew D Wood (Intel Corp)]
5972 *) In the SSL/TLS server implementation, be strict about session ID
5973 context matching (which matters if an application uses a single
5974 external cache for different purposes). Previously,
5975 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5976 set. This did ensure strict client verification, but meant that,
5977 with applications using a single external cache for quite
5978 different requirements, clients could circumvent ciphersuite
5979 restrictions for a given session ID context by starting a session
5980 in a different context.
5983 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5984 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5985 authentication-only ciphersuites.
5988 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5989 not complete and could lead to a possible single byte overflow
5990 (CVE-2007-5135) [Ben Laurie]
5992 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5994 *) Since AES128 and AES256 (and similarly Camellia128 and
5995 Camellia256) share a single mask bit in the logic of
5996 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5997 kludge to work properly if AES128 is available and AES256 isn't
5998 (or if Camellia128 is available and Camellia256 isn't).
6001 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
6002 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
6003 When a point or a seed is encoded in a BIT STRING, we need to
6004 prevent the removal of trailing zero bits to get the proper DER
6005 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
6006 of a NamedBitList, for which trailing 0 bits need to be removed.)
6009 *) Have SSL/TLS server implementation tolerate "mismatched" record
6010 protocol version while receiving ClientHello even if the
6011 ClientHello is fragmented. (The server can't insist on the
6012 particular protocol version it has chosen before the ServerHello
6013 message has informed the client about his choice.)
6016 *) Add RFC 3779 support.
6017 [Rob Austein for ARIN, Ben Laurie]
6019 *) Load error codes if they are not already present instead of using a
6020 static variable. This allows them to be cleanly unloaded and reloaded.
6021 Improve header file function name parsing.
6024 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
6025 or CAPABILITY handshake as required by RFCs.
6028 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
6030 *) Introduce limits to prevent malicious keys being able to
6031 cause a denial of service. (CVE-2006-2940)
6032 [Steve Henson, Bodo Moeller]
6034 *) Fix ASN.1 parsing of certain invalid structures that can result
6035 in a denial of service. (CVE-2006-2937) [Steve Henson]
6037 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6038 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6040 *) Fix SSL client code which could crash if connecting to a
6041 malicious SSLv2 server. (CVE-2006-4343)
6042 [Tavis Ormandy and Will Drewry, Google Security Team]
6044 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
6045 match only those. Before that, "AES256-SHA" would be interpreted
6046 as a pattern and match "AES128-SHA" too (since AES128-SHA got
6047 the same strength classification in 0.9.7h) as we currently only
6048 have a single AES bit in the ciphersuite description bitmap.
6049 That change, however, also applied to ciphersuite strings such as
6050 "RC4-MD5" that intentionally matched multiple ciphersuites --
6051 namely, SSL 2.0 ciphersuites in addition to the more common ones
6052 from SSL 3.0/TLS 1.0.
6054 So we change the selection algorithm again: Naming an explicit
6055 ciphersuite selects this one ciphersuite, and any other similar
6056 ciphersuite (same bitmap) from *other* protocol versions.
6057 Thus, "RC4-MD5" again will properly select both the SSL 2.0
6058 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
6060 Since SSL 2.0 does not have any ciphersuites for which the
6061 128/256 bit distinction would be relevant, this works for now.
6062 The proper fix will be to use different bits for AES128 and
6063 AES256, which would have avoided the problems from the beginning;
6064 however, bits are scarce, so we can only do this in a new release
6065 (not just a patchlevel) when we can change the SSL_CIPHER
6066 definition to split the single 'unsigned long mask' bitmap into
6067 multiple values to extend the available space.
6071 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
6073 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6074 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6076 *) Add AES IGE and biIGE modes.
6079 *) Change the Unix randomness entropy gathering to use poll() when
6080 possible instead of select(), since the latter has some
6081 undesirable limitations.
6082 [Darryl Miles via Richard Levitte and Bodo Moeller]
6084 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
6085 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
6086 cannot be implicitly activated as part of, e.g., the "AES" alias.
6087 However, please upgrade to OpenSSL 0.9.9[-dev] for
6088 non-experimental use of the ECC ciphersuites to get TLS extension
6089 support, which is required for curve and point format negotiation
6090 to avoid potential handshake problems.
6093 *) Disable rogue ciphersuites:
6095 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6096 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6097 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6099 The latter two were purportedly from
6100 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6103 Also deactivate the remaining ciphersuites from
6104 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6105 unofficial, and the ID has long expired.
6108 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6109 dual-core machines) and other potential thread-safety issues.
6112 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
6113 versions), which is now available for royalty-free use
6114 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
6115 Also, add Camellia TLS ciphersuites from RFC 4132.
6117 To minimize changes between patchlevels in the OpenSSL 0.9.8
6118 series, Camellia remains excluded from compilation unless OpenSSL
6119 is configured with 'enable-camellia'.
6122 *) Disable the padding bug check when compression is in use. The padding
6123 bug check assumes the first packet is of even length, this is not
6124 necessarily true if compression is enabled and can result in false
6125 positives causing handshake failure. The actual bug test is ancient
6126 code so it is hoped that implementations will either have fixed it by
6127 now or any which still have the bug do not support compression.
6130 Changes between 0.9.8a and 0.9.8b [04 May 2006]
6132 *) When applying a cipher rule check to see if string match is an explicit
6133 cipher suite and only match that one cipher suite if it is.
6136 *) Link in manifests for VC++ if needed.
6137 [Austin Ziegler <halostatue@gmail.com>]
6139 *) Update support for ECC-based TLS ciphersuites according to
6140 draft-ietf-tls-ecc-12.txt with proposed changes (but without
6141 TLS extensions, which are supported starting with the 0.9.9
6142 branch, not in the OpenSSL 0.9.8 branch).
6145 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
6146 opaque EVP_CIPHER_CTX handling.
6149 *) Fixes and enhancements to zlib compression code. We now only use
6150 "zlib1.dll" and use the default __cdecl calling convention on Win32
6151 to conform with the standards mentioned here:
6152 http://www.zlib.net/DLL_FAQ.txt
6153 Static zlib linking now works on Windows and the new --with-zlib-include
6154 --with-zlib-lib options to Configure can be used to supply the location
6155 of the headers and library. Gracefully handle case where zlib library
6159 *) Several fixes and enhancements to the OID generation code. The old code
6160 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
6161 handle numbers larger than ULONG_MAX, truncated printing and had a
6162 non standard OBJ_obj2txt() behaviour.
6165 *) Add support for building of engines under engine/ as shared libraries
6166 under VC++ build system.
6169 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
6170 Hopefully, we will not see any false combination of paths any more.
6173 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
6175 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6176 (part of SSL_OP_ALL). This option used to disable the
6177 countermeasure against man-in-the-middle protocol-version
6178 rollback in the SSL 2.0 server implementation, which is a bad
6179 idea. (CVE-2005-2969)
6181 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6182 for Information Security, National Institute of Advanced Industrial
6183 Science and Technology [AIST], Japan)]
6185 *) Add two function to clear and return the verify parameter flags.
6188 *) Keep cipherlists sorted in the source instead of sorting them at
6189 runtime, thus removing the need for a lock.
6192 *) Avoid some small subgroup attacks in Diffie-Hellman.
6193 [Nick Mathewson and Ben Laurie]
6195 *) Add functions for well-known primes.
6198 *) Extended Windows CE support.
6199 [Satoshi Nakamura and Andy Polyakov]
6201 *) Initialize SSL_METHOD structures at compile time instead of during
6202 runtime, thus removing the need for a lock.
6205 *) Make PKCS7_decrypt() work even if no certificate is supplied by
6206 attempting to decrypt each encrypted key in turn. Add support to
6210 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
6212 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
6215 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
6218 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
6219 key into the same file any more.
6222 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
6225 *) Add -utf8 command line and config file option to 'ca'.
6226 [Stefan <stf@udoma.org]
6228 *) Removed the macro des_crypt(), as it seems to conflict with some
6229 libraries. Use DES_crypt().
6232 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
6233 involves renaming the source and generated shared-libs for
6234 both. The engines will accept the corrected or legacy ids
6235 ('ncipher' and '4758_cca' respectively) when binding. NB,
6236 this only applies when building 'shared'.
6237 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
6239 *) Add attribute functions to EVP_PKEY structure. Modify
6240 PKCS12_create() to recognize a CSP name attribute and
6241 use it. Make -CSP option work again in pkcs12 utility.
6244 *) Add new functionality to the bn blinding code:
6245 - automatic re-creation of the BN_BLINDING parameters after
6246 a fixed number of uses (currently 32)
6247 - add new function for parameter creation
6248 - introduce flags to control the update behaviour of the
6249 BN_BLINDING parameters
6250 - hide BN_BLINDING structure
6251 Add a second BN_BLINDING slot to the RSA structure to improve
6252 performance when a single RSA object is shared among several
6256 *) Add support for DTLS.
6257 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
6259 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
6260 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
6263 *) Remove buggy and incomplete DH cert support from
6264 ssl/ssl_rsa.c and ssl/s3_both.c
6267 *) Use SHA-1 instead of MD5 as the default digest algorithm for
6268 the apps/openssl applications.
6271 *) Compile clean with "-Wall -Wmissing-prototypes
6272 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6273 DEBUG_SAFESTACK must also be set.
6276 *) Change ./Configure so that certain algorithms can be disabled by default.
6277 The new counterpiece to "no-xxx" is "enable-xxx".
6279 The patented RC5 and MDC2 algorithms will now be disabled unless
6280 "enable-rc5" and "enable-mdc2", respectively, are specified.
6282 (IDEA remains enabled despite being patented. This is because IDEA
6283 is frequently required for interoperability, and there is no license
6284 fee for non-commercial use. As before, "no-idea" can be used to
6285 avoid this algorithm.)
6289 *) Add processing of proxy certificates (see RFC 3820). This work was
6290 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6291 EGEE (Enabling Grids for E-science in Europe).
6294 *) RC4 performance overhaul on modern architectures/implementations, such
6295 as Intel P4, IA-64 and AMD64.
6298 *) New utility extract-section.pl. This can be used specify an alternative
6299 section number in a pod file instead of having to treat each file as
6300 a separate case in Makefile. This can be done by adding two lines to the
6303 =for comment openssl_section:XXX
6305 The blank line is mandatory.
6309 *) New arguments -certform, -keyform and -pass for s_client and s_server
6310 to allow alternative format key and certificate files and passphrase
6314 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
6315 update associated structures and add various utility functions.
6317 Add new policy related verify parameters, include policy checking in
6318 standard verify code. Enhance 'smime' application with extra parameters
6319 to support policy checking and print out.
6322 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6323 Nehemiah processors. These extensions support AES encryption in hardware
6324 as well as RNG (though RNG support is currently disabled).
6325 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6327 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6330 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6331 [Andy Polyakov and a number of other people]
6333 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6334 implementation contributed by IBM.
6335 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6337 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6338 exponent rather than 'unsigned long'. There is a corresponding change to
6339 the new 'rsa_keygen' element of the RSA_METHOD structure.
6340 [Jelte Jansen, Geoff Thorpe]
6342 *) Functionality for creating the initial serial number file is now
6343 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6345 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6346 number file to 1, which is bound to cause problems. To avoid
6347 the problems while respecting compatibility between different 0.9.7
6348 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6349 CA.pl for serial number initialization. With the new release 0.9.8,
6350 we can fix the problem directly in the 'ca' utility.)
6353 *) Reduced header interdependencies by declaring more opaque objects in
6354 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6355 give fewer recursive includes, which could break lazy source code - so
6356 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6357 developers should define this symbol when building and using openssl to
6358 ensure they track the recommended behaviour, interfaces, [etc], but
6359 backwards-compatible behaviour prevails when this isn't defined.
6362 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6365 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6366 This will generate a random key of the appropriate length based on the
6367 cipher context. The EVP_CIPHER can provide its own random key generation
6368 routine to support keys of a specific form. This is used in the des and
6369 3des routines to generate a key of the correct parity. Update S/MIME
6370 code to use new functions and hence generate correct parity DES keys.
6371 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6372 valid (weak or incorrect parity).
6375 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6376 as looking them up. This is useful when the verified structure may contain
6377 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6378 present unless the new PKCS7_NO_CRL flag is asserted.
6381 *) Extend ASN1 oid configuration module. It now additionally accepts the
6384 shortName = some long name, 1.2.3.4
6387 *) Reimplemented the BN_CTX implementation. There is now no more static
6388 limitation on the number of variables it can handle nor the depth of the
6389 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6390 information can now expand as required, and rather than having a single
6391 static array of bignums, BN_CTX now uses a linked-list of such arrays
6392 allowing it to expand on demand whilst maintaining the usefulness of
6393 BN_CTX's "bundling".
6396 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6397 to allow all RSA operations to function using a single BN_CTX.
6400 *) Preliminary support for certificate policy evaluation and checking. This
6401 is initially intended to pass the tests outlined in "Conformance Testing
6402 of Relying Party Client Certificate Path Processing Logic" v1.07.
6405 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6406 remained unused and not that useful. A variety of other little bignum
6407 tweaks and fixes have also been made continuing on from the audit (see
6411 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6412 associated ASN1, EVP and SSL functions and old ASN1 macros.
6415 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6416 and this should never fail. So the return value from the use of
6417 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6418 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6421 *) BN_CTX_get() should return zero-valued bignums, providing the same
6422 initialised value as BN_new().
6423 [Geoff Thorpe, suggested by Ulf Möller]
6425 *) Support for inhibitAnyPolicy certificate extension.
6428 *) An audit of the BIGNUM code is underway, for which debugging code is
6429 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6430 is considered valid when processing BIGNUMs, and causes execution to
6431 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6432 further steps are taken to deliberately pollute unused data in BIGNUM
6433 structures to try and expose faulty code further on. For now, openssl will
6434 (in its default mode of operation) continue to tolerate the inconsistent
6435 forms that it has tolerated in the past, but authors and packagers should
6436 consider trying openssl and their own applications when compiled with
6437 these debugging symbols defined. It will help highlight potential bugs in
6438 their own code, and will improve the test coverage for OpenSSL itself. At
6439 some point, these tighter rules will become openssl's default to improve
6440 maintainability, though the assert()s and other overheads will remain only
6441 in debugging configurations. See bn.h for more details.
6442 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6444 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6445 that can only be obtained through BN_CTX_new() (which implicitly
6446 initialises it). The presence of this function only made it possible
6447 to overwrite an existing structure (and cause memory leaks).
6450 *) Because of the callback-based approach for implementing LHASH as a
6451 template type, lh_insert() adds opaque objects to hash-tables and
6452 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6453 to clean up those corresponding objects before destroying the hash table
6454 (and losing the object pointers). So some over-zealous constifications in
6455 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6456 objects as "const" and the lh_doall[_arg] callback wrappers are not
6457 prototyped to have "const" restrictions on the object pointers they are
6458 given (and so aren't required to cast them away any more).
6461 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6462 (speed) prefers to use its own implementation. The two implementations
6463 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6464 its object type properly exposed (MS_TM) instead of casting to/from "char
6465 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6466 aren't necessarily the greatest nomenclatures - but this is what was used
6467 internally to the implementation so I've used that for now.
6470 *) Ensure that deprecated functions do not get compiled when
6471 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6472 the self-tests were still using deprecated key-generation functions so
6473 these have been updated also.
6476 *) Reorganise PKCS#7 code to separate the digest location functionality
6477 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6478 New function PKCS7_set_digest() to set the digest type for PKCS#7
6479 digestedData type. Add additional code to correctly generate the
6480 digestedData type and add support for this type in PKCS7 initialization
6484 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6485 structure of type "other".
6488 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6489 sure the loop does correctly stop and breaking ("division by zero")
6490 modulus operations are not performed. The (pre-generated) prime
6491 table crypto/bn/bn_prime.h was already correct, but it could not be
6492 re-generated on some platforms because of the "division by zero"
6493 situation in the script.
6494 [Ralf S. Engelschall]
6496 *) Update support for ECC-based TLS ciphersuites according to
6497 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6498 SHA-1 now is only used for "small" curves (where the
6499 representation of a field element takes up to 24 bytes); for
6500 larger curves, the field element resulting from ECDH is directly
6501 used as premaster secret.
6502 [Douglas Stebila (Sun Microsystems Laboratories)]
6504 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6505 curve secp160r1 to the tests.
6506 [Douglas Stebila (Sun Microsystems Laboratories)]
6508 *) Add the possibility to load symbols globally with DSO.
6509 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6511 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6512 control of the error stack.
6515 *) Add support for STORE in ENGINE.
6518 *) Add the STORE type. The intention is to provide a common interface
6519 to certificate and key stores, be they simple file-based stores, or
6520 HSM-type store, or LDAP stores, or...
6521 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6524 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6525 pass a list of arguments to any function as well as provide a way
6526 for a function to pass data back to the caller.
6529 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6530 works like BUF_strdup() but can be used to duplicate a portion of
6531 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6535 *) Add the function sk_find_ex() which works like sk_find(), but will
6536 return an index to an element even if an exact match couldn't be
6537 found. The index is guaranteed to point at the element where the
6538 searched-for key would be inserted to preserve sorting order.
6541 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6542 takes an extra flags argument for optional functionality. Currently,
6543 the following flags are defined:
6545 OBJ_BSEARCH_VALUE_ON_NOMATCH
6546 This one gets OBJ_bsearch_ex() to return a pointer to the first
6547 element where the comparing function returns a negative or zero
6550 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6551 This one gets OBJ_bsearch_ex() to return a pointer to the first
6552 element where the comparing function returns zero. This is useful
6553 if there are more than one element where the comparing function
6557 *) Make it possible to create self-signed certificates with 'openssl ca'
6558 in such a way that the self-signed certificate becomes part of the
6559 CA database and uses the same mechanisms for serial number generation
6560 as all other certificate signing. The new flag '-selfsign' enables
6561 this functionality. Adapt CA.sh and CA.pl.in.
6564 *) Add functionality to check the public key of a certificate request
6565 against a given private. This is useful to check that a certificate
6566 request can be signed by that key (self-signing).
6569 *) Make it possible to have multiple active certificates with the same
6570 subject in the CA index file. This is done only if the keyword
6571 'unique_subject' is set to 'no' in the main CA section (default
6572 if 'CA_default') of the configuration file. The value is saved
6573 with the database itself in a separate index attribute file,
6574 named like the index file with '.attr' appended to the name.
6577 *) Generate multi-valued AVAs using '+' notation in config files for
6581 *) Support for nameConstraints certificate extension.
6584 *) Support for policyConstraints certificate extension.
6587 *) Support for policyMappings certificate extension.
6590 *) Make sure the default DSA_METHOD implementation only uses its
6591 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6592 and change its own handlers to be NULL so as to remove unnecessary
6593 indirection. This lets alternative implementations fallback to the
6594 default implementation more easily.
6597 *) Support for directoryName in GeneralName related extensions
6601 *) Make it possible to link applications using Makefile.shared.
6602 Make that possible even when linking against static libraries!
6605 *) Support for single pass processing for S/MIME signing. This now
6606 means that S/MIME signing can be done from a pipe, in addition
6607 cleartext signing (multipart/signed type) is effectively streaming
6608 and the signed data does not need to be all held in memory.
6610 This is done with a new flag PKCS7_STREAM. When this flag is set
6611 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6612 is done after the data is output (and digests calculated) in
6613 SMIME_write_PKCS7().
6616 *) Add full support for -rpath/-R, both in shared libraries and
6617 applications, at least on the platforms where it's known how
6621 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6622 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6623 will now compute a table of multiples of the generator that
6624 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6625 faster (notably in the case of a single point multiplication,
6626 scalar * generator).
6627 [Nils Larsch, Bodo Moeller]
6629 *) IPv6 support for certificate extensions. The various extensions
6630 which use the IP:a.b.c.d can now take IPv6 addresses using the
6631 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6635 *) Added an ENGINE that implements RSA by performing private key
6636 exponentiations with the GMP library. The conversions to and from
6637 GMP's mpz_t format aren't optimised nor are any montgomery forms
6638 cached, and on x86 it appears OpenSSL's own performance has caught up.
6639 However there are likely to be other architectures where GMP could
6640 provide a boost. This ENGINE is not built in by default, but it can be
6641 specified at Configure time and should be accompanied by the necessary
6642 linker additions, eg;
6643 ./config -DOPENSSL_USE_GMP -lgmp
6646 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6647 testing availability of engines with "-t" - the old behaviour is
6648 produced by increasing the feature's verbosity with "-tt".
6651 *) ECDSA routines: under certain error conditions uninitialized BN objects
6652 could be freed. Solution: make sure initialization is performed early
6653 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6657 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6658 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6659 software implementations. For DSA and DH, parameter generation can
6660 also be overridden by providing the appropriate method callbacks.
6663 *) Change the "progress" mechanism used in key-generation and
6664 primality testing to functions that take a new BN_GENCB pointer in
6665 place of callback/argument pairs. The new API functions have "_ex"
6666 postfixes and the older functions are reimplemented as wrappers for
6667 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6668 declarations of the old functions to help (graceful) attempts to
6669 migrate to the new functions. Also, the new key-generation API
6670 functions operate on a caller-supplied key-structure and return
6671 success/failure rather than returning a key or NULL - this is to
6672 help make "keygen" another member function of RSA_METHOD etc.
6674 Example for using the new callback interface:
6676 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6680 BN_GENCB_set(&my_cb, my_callback, my_arg);
6682 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6683 /* For the meaning of a, b in calls to my_callback(), see the
6684 * documentation of the function that calls the callback.
6685 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6686 * my_callback should return 1 if it wants BN_is_prime_ex()
6687 * to continue, or 0 to stop.
6692 *) Change the ZLIB compression method to be stateful, and make it
6693 available to TLS with the number defined in
6694 draft-ietf-tls-compression-04.txt.
6697 *) Add the ASN.1 structures and functions for CertificatePair, which
6698 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6700 CertificatePair ::= SEQUENCE {
6701 forward [0] Certificate OPTIONAL,
6702 reverse [1] Certificate OPTIONAL,
6703 -- at least one of the pair shall be present -- }
6705 Also implement the PEM functions to read and write certificate
6706 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6708 This needed to be defined, mostly for the sake of the LDAP
6709 attribute crossCertificatePair, but may prove useful elsewhere as
6713 *) Make it possible to inhibit symlinking of shared libraries in
6714 Makefile.shared, for Cygwin's sake.
6717 *) Extend the BIGNUM API by creating a function
6718 void BN_set_negative(BIGNUM *a, int neg);
6719 and a macro that behave like
6720 int BN_is_negative(const BIGNUM *a);
6722 to avoid the need to access 'a->neg' directly in applications.
6725 *) Implement fast modular reduction for pseudo-Mersenne primes
6726 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6727 EC_GROUP_new_curve_GFp() will now automatically use this
6729 [Nils Larsch <nla@trustcenter.de>]
6731 *) Add new lock type (CRYPTO_LOCK_BN).
6734 *) Change the ENGINE framework to automatically load engines
6735 dynamically from specific directories unless they could be
6736 found to already be built in or loaded. Move all the
6737 current engines except for the cryptodev one to a new
6739 The engines in engines/ are built as shared libraries if
6740 the "shared" options was given to ./Configure or ./config.
6741 Otherwise, they are inserted in libcrypto.a.
6742 /usr/local/ssl/engines is the default directory for dynamic
6743 engines, but that can be overridden at configure time through
6744 the usual use of --prefix and/or --openssldir, and at run
6745 time with the environment variable OPENSSL_ENGINES.
6746 [Geoff Thorpe and Richard Levitte]
6748 *) Add Makefile.shared, a helper makefile to build shared
6749 libraries. Adapt Makefile.org.
6752 *) Add version info to Win32 DLLs.
6753 [Peter 'Luna' Runestig" <peter@runestig.com>]
6755 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6756 can be added using this API to created arbitrary PKCS#12
6757 files while avoiding the low level API.
6759 New options to PKCS12_create(), key or cert can be NULL and
6760 will then be omitted from the output file. The encryption
6761 algorithm NIDs can be set to -1 for no encryption, the mac
6762 iteration count can be set to 0 to omit the mac.
6764 Enhance pkcs12 utility by making the -nokeys and -nocerts
6765 options work when creating a PKCS#12 file. New option -nomac
6766 to omit the mac, NONE can be set for an encryption algorithm.
6767 New code is modified to use the enhanced PKCS12_create()
6768 instead of the low level API.
6771 *) Extend ASN1 encoder to support indefinite length constructed
6772 encoding. This can output sequences tags and octet strings in
6773 this form. Modify pk7_asn1.c to support indefinite length
6774 encoding. This is experimental and needs additional code to
6775 be useful, such as an ASN1 bio and some enhanced streaming
6778 Extend template encode functionality so that tagging is passed
6779 down to the template encoder.
6782 *) Let 'openssl req' fail if an argument to '-newkey' is not
6783 recognized instead of using RSA as a default.
6786 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6787 As these are not official, they are not included in "ALL";
6788 the "ECCdraft" ciphersuite group alias can be used to select them.
6789 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6791 *) Add ECDH engine support.
6792 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6794 *) Add ECDH in new directory crypto/ecdh/.
6795 [Douglas Stebila (Sun Microsystems Laboratories)]
6797 *) Let BN_rand_range() abort with an error after 100 iterations
6798 without success (which indicates a broken PRNG).
6801 *) Change BN_mod_sqrt() so that it verifies that the input value
6802 is really the square of the return value. (Previously,
6803 BN_mod_sqrt would show GIGO behaviour.)
6806 *) Add named elliptic curves over binary fields from X9.62, SECG,
6807 and WAP/WTLS; add OIDs that were still missing.
6809 [Sheueling Chang Shantz and Douglas Stebila
6810 (Sun Microsystems Laboratories)]
6812 *) Extend the EC library for elliptic curves over binary fields
6813 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6816 EC_GF2m_simple_method
6820 EC_GROUP_new_curve_GF2m
6821 EC_GROUP_set_curve_GF2m
6822 EC_GROUP_get_curve_GF2m
6823 EC_POINT_set_affine_coordinates_GF2m
6824 EC_POINT_get_affine_coordinates_GF2m
6825 EC_POINT_set_compressed_coordinates_GF2m
6827 Point compression for binary fields is disabled by default for
6828 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6831 As binary polynomials are represented as BIGNUMs, various members
6832 of the EC_GROUP and EC_POINT data structures can be shared
6833 between the implementations for prime fields and binary fields;
6834 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6835 are essentially identical to their ..._GFp counterparts.
6836 (For simplicity, the '..._GFp' prefix has been dropped from
6837 various internal method names.)
6839 An internal 'field_div' method (similar to 'field_mul' and
6840 'field_sqr') has been added; this is used only for binary fields.
6842 [Sheueling Chang Shantz and Douglas Stebila
6843 (Sun Microsystems Laboratories)]
6845 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6846 through methods ('mul', 'precompute_mult').
6848 The generic implementations (now internally called 'ec_wNAF_mul'
6849 and 'ec_wNAF_precomputed_mult') remain the default if these
6850 methods are undefined.
6852 [Sheueling Chang Shantz and Douglas Stebila
6853 (Sun Microsystems Laboratories)]
6855 *) New function EC_GROUP_get_degree, which is defined through
6856 EC_METHOD. For curves over prime fields, this returns the bit
6857 length of the modulus.
6859 [Sheueling Chang Shantz and Douglas Stebila
6860 (Sun Microsystems Laboratories)]
6862 *) New functions EC_GROUP_dup, EC_POINT_dup.
6863 (These simply call ..._new and ..._copy).
6865 [Sheueling Chang Shantz and Douglas Stebila
6866 (Sun Microsystems Laboratories)]
6868 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6869 Polynomials are represented as BIGNUMs (where the sign bit is not
6870 used) in the following functions [macros]:
6873 BN_GF2m_sub [= BN_GF2m_add]
6874 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6875 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6876 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6878 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6879 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6880 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6881 BN_GF2m_cmp [= BN_ucmp]
6883 (Note that only the 'mod' functions are actually for fields GF(2^m).
6884 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6886 For some functions, an the irreducible polynomial defining a
6887 field can be given as an 'unsigned int[]' with strictly
6888 decreasing elements giving the indices of those bits that are set;
6889 i.e., p[] represents the polynomial
6890 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6892 p[0] > p[1] > ... > p[k] = 0.
6893 This applies to the following functions:
6898 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6899 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6901 BN_GF2m_mod_sqrt_arr
6902 BN_GF2m_mod_solve_quad_arr
6906 Conversion can be performed by the following functions:
6911 bntest.c has additional tests for binary polynomial arithmetic.
6913 Two implementations for BN_GF2m_mod_div() are available.
6914 The default algorithm simply uses BN_GF2m_mod_inv() and
6915 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6916 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6917 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6919 [Sheueling Chang Shantz and Douglas Stebila
6920 (Sun Microsystems Laboratories)]
6922 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6923 functionality is disabled at compile-time.
6924 [Douglas Stebila <douglas.stebila@sun.com>]
6926 *) Change default behaviour of 'openssl asn1parse' so that more
6927 information is visible when viewing, e.g., a certificate:
6929 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6930 mode the content of non-printable OCTET STRINGs is output in a
6931 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6932 avoid the appearance of a printable string.
6933 [Nils Larsch <nla@trustcenter.de>]
6935 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6937 EC_GROUP_set_asn1_flag()
6938 EC_GROUP_get_asn1_flag()
6939 EC_GROUP_set_point_conversion_form()
6940 EC_GROUP_get_point_conversion_form()
6941 These control ASN1 encoding details:
6942 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6943 has been set to OPENSSL_EC_NAMED_CURVE.
6944 - Points are encoded in uncompressed form by default; options for
6945 asn1_for are as for point2oct, namely
6946 POINT_CONVERSION_COMPRESSED
6947 POINT_CONVERSION_UNCOMPRESSED
6948 POINT_CONVERSION_HYBRID
6950 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6953 EC_GROUP_get0_seed()
6954 EC_GROUP_get_seed_len()
6955 This is used only for ASN1 purposes (so far).
6956 [Nils Larsch <nla@trustcenter.de>]
6958 *) Add 'field_type' member to EC_METHOD, which holds the NID
6959 of the appropriate field type OID. The new function
6960 EC_METHOD_get_field_type() returns this value.
6961 [Nils Larsch <nla@trustcenter.de>]
6966 EC_POINT_point2hex()
6967 EC_POINT_hex2point()
6968 providing useful interfaces to EC_POINT_point2oct() and
6969 EC_POINT_oct2point().
6970 [Nils Larsch <nla@trustcenter.de>]
6972 *) Change internals of the EC library so that the functions
6973 EC_GROUP_set_generator()
6974 EC_GROUP_get_generator()
6975 EC_GROUP_get_order()
6976 EC_GROUP_get_cofactor()
6977 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6978 to methods, which would lead to unnecessary code duplication when
6979 adding different types of curves.
6980 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6982 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6983 arithmetic, and such that modified wNAFs are generated
6984 (which avoid length expansion in many cases).
6987 *) Add a function EC_GROUP_check_discriminant() (defined via
6988 EC_METHOD) that verifies that the curve discriminant is non-zero.
6990 Add a function EC_GROUP_check() that makes some sanity tests
6991 on a EC_GROUP, its generator and order. This includes
6992 EC_GROUP_check_discriminant().
6993 [Nils Larsch <nla@trustcenter.de>]
6995 *) Add ECDSA in new directory crypto/ecdsa/.
6997 Add applications 'openssl ecparam' and 'openssl ecdsa'
6998 (these are based on 'openssl dsaparam' and 'openssl dsa').
7000 ECDSA support is also included in various other files across the
7001 library. Most notably,
7002 - 'openssl req' now has a '-newkey ecdsa:file' option;
7003 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
7004 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
7005 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
7006 them suitable for ECDSA where domain parameters must be
7007 extracted before the specific public key;
7008 - ECDSA engine support has been added.
7009 [Nils Larsch <nla@trustcenter.de>]
7011 *) Include some named elliptic curves, and add OIDs from X9.62,
7012 SECG, and WAP/WTLS. Each curve can be obtained from the new
7014 EC_GROUP_new_by_curve_name(),
7015 and the list of available named curves can be obtained with
7016 EC_get_builtin_curves().
7017 Also add a 'curve_name' member to EC_GROUP objects, which can be
7019 EC_GROUP_set_curve_name()
7020 EC_GROUP_get_curve_name()
7021 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
7023 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7024 was actually never needed) and in BN_mul(). The removal in BN_mul()
7025 required a small change in bn_mul_part_recursive() and the addition
7026 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7027 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7028 bn_sub_words() and bn_add_words() except they take arrays with
7032 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
7034 *) Cleanse PEM buffers before freeing them since they may contain
7036 [Benjamin Bennett <ben@psc.edu>]
7038 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
7039 a ciphersuite string such as "DEFAULT:RSA" cannot enable
7040 authentication-only ciphersuites.
7043 *) Since AES128 and AES256 share a single mask bit in the logic of
7044 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
7045 kludge to work properly if AES128 is available and AES256 isn't.
7048 *) Expand security boundary to match 1.1.1 module.
7051 *) Remove redundant features: hash file source, editing of test vectors
7052 modify fipsld to use external fips_premain.c signature.
7055 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
7056 run algorithm test programs.
7059 *) Make algorithm test programs more tolerant of whitespace.
7062 *) Have SSL/TLS server implementation tolerate "mismatched" record
7063 protocol version while receiving ClientHello even if the
7064 ClientHello is fragmented. (The server can't insist on the
7065 particular protocol version it has chosen before the ServerHello
7066 message has informed the client about his choice.)
7069 *) Load error codes if they are not already present instead of using a
7070 static variable. This allows them to be cleanly unloaded and reloaded.
7073 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
7075 *) Introduce limits to prevent malicious keys being able to
7076 cause a denial of service. (CVE-2006-2940)
7077 [Steve Henson, Bodo Moeller]
7079 *) Fix ASN.1 parsing of certain invalid structures that can result
7080 in a denial of service. (CVE-2006-2937) [Steve Henson]
7082 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
7083 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
7085 *) Fix SSL client code which could crash if connecting to a
7086 malicious SSLv2 server. (CVE-2006-4343)
7087 [Tavis Ormandy and Will Drewry, Google Security Team]
7089 *) Change ciphersuite string processing so that an explicit
7090 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
7091 will no longer include "AES128-SHA"), and any other similar
7092 ciphersuite (same bitmap) from *other* protocol versions (so that
7093 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
7094 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
7095 changes from 0.9.8b and 0.9.8d.
7098 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
7100 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
7101 (CVE-2006-4339) [Ben Laurie and Google Security Team]
7103 *) Change the Unix randomness entropy gathering to use poll() when
7104 possible instead of select(), since the latter has some
7105 undesirable limitations.
7106 [Darryl Miles via Richard Levitte and Bodo Moeller]
7108 *) Disable rogue ciphersuites:
7110 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
7111 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
7112 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
7114 The latter two were purportedly from
7115 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
7118 Also deactivate the remaining ciphersuites from
7119 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
7120 unofficial, and the ID has long expired.
7123 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
7124 dual-core machines) and other potential thread-safety issues.
7127 Changes between 0.9.7i and 0.9.7j [04 May 2006]
7129 *) Adapt fipsld and the build system to link against the validated FIPS
7130 module in FIPS mode.
7133 *) Fixes for VC++ 2005 build under Windows.
7136 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
7137 from a Windows bash shell such as MSYS. It is autodetected from the
7138 "config" script when run from a VC++ environment. Modify standard VC++
7139 build to use fipscanister.o from the GNU make build.
7142 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
7144 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
7145 The value now differs depending on if you build for FIPS or not.
7146 BEWARE! A program linked with a shared FIPSed libcrypto can't be
7147 safely run with a non-FIPSed libcrypto, as it may crash because of
7148 the difference induced by this change.
7151 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
7153 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
7154 (part of SSL_OP_ALL). This option used to disable the
7155 countermeasure against man-in-the-middle protocol-version
7156 rollback in the SSL 2.0 server implementation, which is a bad
7157 idea. (CVE-2005-2969)
7159 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
7160 for Information Security, National Institute of Advanced Industrial
7161 Science and Technology [AIST], Japan)]
7163 *) Minimal support for X9.31 signatures and PSS padding modes. This is
7164 mainly for FIPS compliance and not fully integrated at this stage.
7167 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
7168 the exponentiation using a fixed-length exponent. (Otherwise,
7169 the information leaked through timing could expose the secret key
7170 after many signatures; cf. Bleichenbacher's attack on DSA with
7174 *) Make a new fixed-window mod_exp implementation the default for
7175 RSA, DSA, and DH private-key operations so that the sequence of
7176 squares and multiplies and the memory access pattern are
7177 independent of the particular secret key. This will mitigate
7178 cache-timing and potential related attacks.
7180 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
7181 and this is automatically used by BN_mod_exp_mont() if the new flag
7182 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
7183 will use this BN flag for private exponents unless the flag
7184 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
7185 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
7187 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
7189 *) Change the client implementation for SSLv23_method() and
7190 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
7191 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
7192 (Previously, the SSL 2.0 backwards compatible Client Hello
7193 message format would be used even with SSL_OP_NO_SSLv2.)
7196 *) Add support for smime-type MIME parameter in S/MIME messages which some
7200 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
7201 a threadsafe manner. Modify rsa code to use new function and add calls
7202 to dsa and dh code (which had race conditions before).
7205 *) Include the fixed error library code in the C error file definitions
7206 instead of fixing them up at runtime. This keeps the error code
7207 structures constant.
7210 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
7212 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
7215 *) Fixes for newer kerberos headers. NB: the casts are needed because
7216 the 'length' field is signed on one version and unsigned on another
7217 with no (?) obvious way to tell the difference, without these VC++
7218 complains. Also the "definition" of FAR (blank) is no longer included
7219 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
7220 some needed definitions.
7223 *) Undo Cygwin change.
7226 *) Added support for proxy certificates according to RFC 3820.
7227 Because they may be a security thread to unaware applications,
7228 they must be explicitly allowed in run-time. See
7229 docs/HOWTO/proxy_certificates.txt for further information.
7232 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
7234 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
7235 server and client random values. Previously
7236 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
7237 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
7239 This change has negligible security impact because:
7241 1. Server and client random values still have 24 bytes of pseudo random
7244 2. Server and client random values are sent in the clear in the initial
7247 3. The master secret is derived using the premaster secret (48 bytes in
7248 size for static RSA ciphersuites) as well as client server and random
7251 The OpenSSL team would like to thank the UK NISCC for bringing this issue
7254 [Stephen Henson, reported by UK NISCC]
7256 *) Use Windows randomness collection on Cygwin.
7259 *) Fix hang in EGD/PRNGD query when communication socket is closed
7260 prematurely by EGD/PRNGD.
7261 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
7263 *) Prompt for pass phrases when appropriate for PKCS12 input format.
7266 *) Back-port of selected performance improvements from development
7267 branch, as well as improved support for PowerPC platforms.
7270 *) Add lots of checks for memory allocation failure, error codes to indicate
7271 failure and freeing up memory if a failure occurs.
7272 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
7274 *) Add new -passin argument to dgst.
7277 *) Perform some character comparisons of different types in X509_NAME_cmp:
7278 this is needed for some certificates that re-encode DNs into UTF8Strings
7279 (in violation of RFC3280) and can't or won't issue name rollover
7283 *) Make an explicit check during certificate validation to see that
7284 the CA setting in each certificate on the chain is correct. As a
7285 side effect always do the following basic checks on extensions,
7286 not just when there's an associated purpose to the check:
7288 - if there is an unhandled critical extension (unless the user
7289 has chosen to ignore this fault)
7290 - if the path length has been exceeded (if one is set at all)
7291 - that certain extensions fit the associated purpose (if one has
7295 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
7297 *) Avoid a race condition when CRLs are checked in a multi threaded
7298 environment. This would happen due to the reordering of the revoked
7299 entries during signature checking and serial number lookup. Now the
7300 encoding is cached and the serial number sort performed under a lock.
7301 Add new STACK function sk_is_sorted().
7304 *) Add Delta CRL to the extension code.
7307 *) Various fixes to s3_pkt.c so alerts are sent properly.
7308 [David Holmes <d.holmes@f5.com>]
7310 *) Reduce the chances of duplicate issuer name and serial numbers (in
7311 violation of RFC3280) using the OpenSSL certificate creation utilities.
7312 This is done by creating a random 64 bit value for the initial serial
7313 number when a serial number file is created or when a self signed
7314 certificate is created using 'openssl req -x509'. The initial serial
7315 number file is created using 'openssl x509 -next_serial' in CA.pl
7316 rather than being initialized to 1.
7319 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7321 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7322 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7323 [Joe Orton, Steve Henson]
7325 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7327 [Joe Orton, Steve Henson]
7329 *) Make it possible to have multiple active certificates with the same
7330 subject in the CA index file. This is done only if the keyword
7331 'unique_subject' is set to 'no' in the main CA section (default
7332 if 'CA_default') of the configuration file. The value is saved
7333 with the database itself in a separate index attribute file,
7334 named like the index file with '.attr' appended to the name.
7337 *) X509 verify fixes. Disable broken certificate workarounds when
7338 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7339 keyUsage extension present. Don't accept CRLs with unhandled critical
7340 extensions: since verify currently doesn't process CRL extensions this
7341 rejects a CRL with *any* critical extensions. Add new verify error codes
7345 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7346 A clarification of RFC2560 will require the use of OCTET STRINGs and
7347 some implementations cannot handle the current raw format. Since OpenSSL
7348 copies and compares OCSP nonces as opaque blobs without any attempt at
7349 parsing them this should not create any compatibility issues.
7352 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7353 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7354 this HMAC (and other) operations are several times slower than OpenSSL
7358 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7359 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7361 *) Use the correct content when signing type "other".
7364 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7366 *) Fix various bugs revealed by running the NISCC test suite:
7368 Stop out of bounds reads in the ASN1 code when presented with
7369 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7371 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7373 If verify callback ignores invalid public key errors don't try to check
7374 certificate signature with the NULL public key.
7378 *) New -ignore_err option in ocsp application to stop the server
7379 exiting on the first error in a request.
7382 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7383 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7387 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7388 extra data after the compression methods not only for TLS 1.0
7389 but also for SSL 3.0 (as required by the specification).
7390 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7392 *) Change X509_certificate_type() to mark the key as exported/exportable
7393 when it's 512 *bits* long, not 512 bytes.
7396 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7397 blocks during encryption.
7400 *) Various fixes to base64 BIO and non blocking I/O. On write
7401 flushes were not handled properly if the BIO retried. On read
7402 data was not being buffered properly and had various logic bugs.
7403 This also affects blocking I/O when the data being decoded is a
7407 *) Various S/MIME bugfixes and compatibility changes:
7408 output correct application/pkcs7 MIME type if
7409 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7410 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7411 of files as .eml work). Correctly handle very long lines in MIME
7415 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7417 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7418 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7419 a protocol version number mismatch like a decryption error
7420 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7423 *) Turn on RSA blinding by default in the default implementation
7424 to avoid a timing attack. Applications that don't want it can call
7425 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7426 They would be ill-advised to do so in most cases.
7427 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7429 *) Change RSA blinding code so that it works when the PRNG is not
7430 seeded (in this case, the secret RSA exponent is abused as
7431 an unpredictable seed -- if it is not unpredictable, there
7432 is no point in blinding anyway). Make RSA blinding thread-safe
7433 by remembering the creator's thread ID in rsa->blinding and
7434 having all other threads use local one-time blinding factors
7435 (this requires more computation than sharing rsa->blinding, but
7436 avoids excessive locking; and if an RSA object is not shared
7437 between threads, blinding will still be very fast).
7440 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7441 ENGINE as defaults for all supported algorithms irrespective of
7442 the 'flags' parameter. 'flags' is now honoured, so applications
7443 should make sure they are passing it correctly.
7446 *) Target "mingw" now allows native Windows code to be generated in
7447 the Cygwin environment as well as with the MinGW compiler.
7450 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7452 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7453 via timing by performing a MAC computation even if incorrect
7454 block cipher padding has been found. This is a countermeasure
7455 against active attacks where the attacker has to distinguish
7456 between bad padding and a MAC verification error. (CVE-2003-0078)
7458 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7459 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7460 Martin Vuagnoux (EPFL, Ilion)]
7462 *) Make the no-err option work as intended. The intention with no-err
7463 is not to have the whole error stack handling routines removed from
7464 libcrypto, it's only intended to remove all the function name and
7465 reason texts, thereby removing some of the footprint that may not
7466 be interesting if those errors aren't displayed anyway.
7468 NOTE: it's still possible for any application or module to have its
7469 own set of error texts inserted. The routines are there, just not
7470 used by default when no-err is given.
7473 *) Add support for FreeBSD on IA64.
7474 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7476 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7477 Kerberos function mit_des_cbc_cksum(). Before this change,
7478 the value returned by DES_cbc_cksum() was like the one from
7479 mit_des_cbc_cksum(), except the bytes were swapped.
7480 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7482 *) Allow an application to disable the automatic SSL chain building.
7483 Before this a rather primitive chain build was always performed in
7484 ssl3_output_cert_chain(): an application had no way to send the
7485 correct chain if the automatic operation produced an incorrect result.
7487 Now the chain builder is disabled if either:
7489 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7491 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7493 The reasoning behind this is that an application would not want the
7494 auto chain building to take place if extra chain certificates are
7495 present and it might also want a means of sending no additional
7496 certificates (for example the chain has two certificates and the
7500 *) Add the possibility to build without the ENGINE framework.
7501 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7503 *) Under Win32 gmtime() can return NULL: check return value in
7504 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7507 *) DSA routines: under certain error conditions uninitialized BN objects
7508 could be freed. Solution: make sure initialization is performed early
7509 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7510 Nils Larsch <nla@trustcenter.de> via PR#459)
7513 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7514 checked on reconnect on the client side, therefore session resumption
7515 could still fail with a "ssl session id is different" error. This
7516 behaviour is masked when SSL_OP_ALL is used due to
7517 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7518 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7519 followup to PR #377.
7522 *) IA-32 assembler support enhancements: unified ELF targets, support
7523 for SCO/Caldera platforms, fix for Cygwin shared build.
7526 *) Add support for FreeBSD on sparc64. As a consequence, support for
7527 FreeBSD on non-x86 processors is separate from x86 processors on
7528 the config script, much like the NetBSD support.
7529 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7531 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7533 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7536 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7537 code (06) was taken as the first octet of the session ID and the last
7538 octet was ignored consequently. As a result SSLv2 client side session
7539 caching could not have worked due to the session ID mismatch between
7541 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7545 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7546 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7550 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7551 seems that in spite of existing for more than a year, many application
7552 author have done nothing to provide the necessary callbacks, which
7553 means that this particular engine will not work properly anywhere.
7554 This is a very unfortunate situation which forces us, in the name
7555 of usability, to give the hw_ncipher.c a static lock, which is part
7557 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7558 appear in 0.9.8 or later. We EXPECT application authors to have
7559 dealt properly with this when 0.9.8 is released (unless we actually
7560 make such changes in the libcrypto locking code that changes will
7561 have to be made anyway).
7564 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7565 octets have been read, EOF or an error occurs. Without this change
7566 some truncated ASN1 structures will not produce an error.
7569 *) Disable Heimdal support, since it hasn't been fully implemented.
7570 Still give the possibility to force the use of Heimdal, but with
7571 warnings and a request that patches get sent to openssl-dev.
7574 *) Add the VC-CE target, introduce the WINCE sysname, and add
7575 INSTALL.WCE and appropriate conditionals to make it build.
7576 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7578 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7579 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7580 edit numbers of the version.
7581 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7583 *) Introduce safe string copy and catenation functions
7584 (BUF_strlcpy() and BUF_strlcat()).
7585 [Ben Laurie (CHATS) and Richard Levitte]
7587 *) Avoid using fixed-size buffers for one-line DNs.
7588 [Ben Laurie (CHATS)]
7590 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7591 resizing buffers containing secrets, and use where appropriate.
7592 [Ben Laurie (CHATS)]
7594 *) Avoid using fixed size buffers for configuration file location.
7595 [Ben Laurie (CHATS)]
7597 *) Avoid filename truncation for various CA files.
7598 [Ben Laurie (CHATS)]
7600 *) Use sizeof in preference to magic numbers.
7601 [Ben Laurie (CHATS)]
7603 *) Avoid filename truncation in cert requests.
7604 [Ben Laurie (CHATS)]
7606 *) Add assertions to check for (supposedly impossible) buffer
7608 [Ben Laurie (CHATS)]
7610 *) Don't cache truncated DNS entries in the local cache (this could
7611 potentially lead to a spoofing attack).
7612 [Ben Laurie (CHATS)]
7614 *) Fix various buffers to be large enough for hex/decimal
7615 representations in a platform independent manner.
7616 [Ben Laurie (CHATS)]
7618 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7619 resizing buffers containing secrets, and use where appropriate.
7620 [Ben Laurie (CHATS)]
7622 *) Add BIO_indent() to avoid much slightly worrying code to do
7624 [Ben Laurie (CHATS)]
7626 *) Convert sprintf()/BIO_puts() to BIO_printf().
7627 [Ben Laurie (CHATS)]
7629 *) buffer_gets() could terminate with the buffer only half
7631 [Ben Laurie (CHATS)]
7633 *) Add assertions to prevent user-supplied crypto functions from
7634 overflowing internal buffers by having large block sizes, etc.
7635 [Ben Laurie (CHATS)]
7637 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7639 [Ben Laurie (CHATS)]
7641 *) Eliminate unused copy of key in RC4.
7642 [Ben Laurie (CHATS)]
7644 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7645 [Ben Laurie (CHATS)]
7647 *) Fix off-by-one error in EGD path.
7648 [Ben Laurie (CHATS)]
7650 *) If RANDFILE path is too long, ignore instead of truncating.
7651 [Ben Laurie (CHATS)]
7653 *) Eliminate unused and incorrectly sized X.509 structure
7655 [Ben Laurie (CHATS)]
7657 *) Eliminate unused and dangerous function knumber().
7658 [Ben Laurie (CHATS)]
7660 *) Eliminate unused and dangerous structure, KSSL_ERR.
7661 [Ben Laurie (CHATS)]
7663 *) Protect against overlong session ID context length in an encoded
7664 session object. Since these are local, this does not appear to be
7666 [Ben Laurie (CHATS)]
7668 *) Change from security patch (see 0.9.6e below) that did not affect
7669 the 0.9.6 release series:
7671 Remote buffer overflow in SSL3 protocol - an attacker could
7672 supply an oversized master key in Kerberos-enabled versions.
7674 [Ben Laurie (CHATS)]
7676 *) Change the SSL kerb5 codes to match RFC 2712.
7679 *) Make -nameopt work fully for req and add -reqopt switch.
7680 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7682 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7683 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7685 *) Make sure tests can be performed even if the corresponding algorithms
7686 have been removed entirely. This was also the last step to make
7687 OpenSSL compilable with DJGPP under all reasonable conditions.
7688 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7690 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7691 to allow version independent disabling of normally unselected ciphers,
7692 which may be activated as a side-effect of selecting a single cipher.
7694 (E.g., cipher list string "RSA" enables ciphersuites that are left
7695 out of "ALL" because they do not provide symmetric encryption.
7696 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7697 [Lutz Jaenicke, Bodo Moeller]
7699 *) Add appropriate support for separate platform-dependent build
7700 directories. The recommended way to make a platform-dependent
7701 build directory is the following (tested on Linux), maybe with
7704 # Place yourself outside of the OpenSSL source tree. In
7705 # this example, the environment variable OPENSSL_SOURCE
7706 # is assumed to contain the absolute OpenSSL source directory.
7707 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7708 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7709 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7710 mkdir -p `dirname $F`
7711 ln -s $OPENSSL_SOURCE/$F $F
7714 To be absolutely sure not to disturb the source tree, a "make clean"
7715 is a good thing. If it isn't successful, don't worry about it,
7716 it probably means the source directory is very clean.
7719 *) Make sure any ENGINE control commands make local copies of string
7720 pointers passed to them whenever necessary. Otherwise it is possible
7721 the caller may have overwritten (or deallocated) the original string
7722 data when a later ENGINE operation tries to use the stored values.
7723 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7725 *) Improve diagnostics in file reading and command-line digests.
7726 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7728 *) Add AES modes CFB and OFB to the object database. Correct an
7729 error in AES-CFB decryption.
7732 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7733 allows existing EVP_CIPHER_CTX structures to be reused after
7734 calling EVP_*Final(). This behaviour is used by encryption
7735 BIOs and some applications. This has the side effect that
7736 applications must explicitly clean up cipher contexts with
7737 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7740 *) Check the values of dna and dnb in bn_mul_recursive before calling
7741 bn_mul_comba (a non zero value means the a or b arrays do not contain
7742 n2 elements) and fallback to bn_mul_normal if either is not zero.
7745 *) Fix escaping of non-ASCII characters when using the -subj option
7746 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7749 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7750 form for "surname", serialNumber has no short form.
7751 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7752 therefore remove "mail" short name for "internet 7".
7753 The OID for unique identifiers in X509 certificates is
7754 x500UniqueIdentifier, not uniqueIdentifier.
7755 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7758 *) Add an "init" command to the ENGINE config module and auto initialize
7759 ENGINEs. Without any "init" command the ENGINE will be initialized
7760 after all ctrl commands have been executed on it. If init=1 the
7761 ENGINE is initialized at that point (ctrls before that point are run
7762 on the uninitialized ENGINE and after on the initialized one). If
7763 init=0 then the ENGINE will not be initialized at all.
7766 *) Fix the 'app_verify_callback' interface so that the user-defined
7767 argument is actually passed to the callback: In the
7768 SSL_CTX_set_cert_verify_callback() prototype, the callback
7769 declaration has been changed from
7772 int (*cb)(X509_STORE_CTX *,void *);
7773 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7774 i=s->ctx->app_verify_callback(&ctx)
7775 has been changed into
7776 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7778 To update applications using SSL_CTX_set_cert_verify_callback(),
7779 a dummy argument can be added to their callback functions.
7780 [D. K. Smetters <smetters@parc.xerox.com>]
7782 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7783 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7785 *) Add and OPENSSL_LOAD_CONF define which will cause
7786 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7787 This allows older applications to transparently support certain
7788 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7789 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7790 load the config file and OPENSSL_add_all_algorithms_conf() which will
7791 always load it have also been added.
7794 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7795 Adjust NIDs and EVP layer.
7796 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7798 *) Config modules support in openssl utility.
7800 Most commands now load modules from the config file,
7801 though in a few (such as version) this isn't done
7802 because it couldn't be used for anything.
7804 In the case of ca and req the config file used is
7805 the same as the utility itself: that is the -config
7806 command line option can be used to specify an
7810 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7811 use "openssl_conf" if filename is NULL use default openssl config file.
7814 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7815 config section name. Add a new flag to tolerate a missing config file
7816 and move code to CONF_modules_load_file().
7819 *) Support for crypto accelerator cards from Accelerated Encryption
7820 Processing, www.aep.ie. (Use engine 'aep')
7821 The support was copied from 0.9.6c [engine] and adapted/corrected
7822 to work with the new engine framework.
7823 [AEP Inc. and Richard Levitte]
7825 *) Support for SureWare crypto accelerator cards from Baltimore
7826 Technologies. (Use engine 'sureware')
7827 The support was copied from 0.9.6c [engine] and adapted
7828 to work with the new engine framework.
7831 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7832 make the newer ENGINE framework commands for the CHIL engine work.
7833 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7835 *) Make it possible to produce shared libraries on ReliantUNIX.
7836 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7838 *) Add the configuration target debug-linux-ppro.
7839 Make 'openssl rsa' use the general key loading routines
7840 implemented in apps.c, and make those routines able to
7841 handle the key format FORMAT_NETSCAPE and the variant
7843 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7845 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7846 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7848 *) Add -keyform to rsautl, and document -engine.
7849 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7851 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7852 BIO_R_NO_SUCH_FILE error code rather than the generic
7853 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7856 *) Add new functions
7858 ERR_peek_last_error_line
7859 ERR_peek_last_error_line_data.
7860 These are similar to
7863 ERR_peek_error_line_data,
7864 but report on the latest error recorded rather than the first one
7865 still in the error queue.
7866 [Ben Laurie, Bodo Moeller]
7868 *) default_algorithms option in ENGINE config module. This allows things
7870 default_algorithms = ALL
7871 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7874 *) Preliminary ENGINE config module.
7877 *) New experimental application configuration code.
7880 *) Change the AES code to follow the same name structure as all other
7881 symmetric ciphers, and behave the same way. Move everything to
7882 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7883 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7885 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7886 [Ben Laurie and Theo de Raadt]
7888 *) Add option to output public keys in req command.
7889 [Massimiliano Pala madwolf@openca.org]
7891 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7892 (up to about 10% better than before for P-192 and P-224).
7895 *) New functions/macros
7897 SSL_CTX_set_msg_callback(ctx, cb)
7898 SSL_CTX_set_msg_callback_arg(ctx, arg)
7899 SSL_set_msg_callback(ssl, cb)
7900 SSL_set_msg_callback_arg(ssl, arg)
7902 to request calling a callback function
7904 void cb(int write_p, int version, int content_type,
7905 const void *buf, size_t len, SSL *ssl, void *arg)
7907 whenever a protocol message has been completely received
7908 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7909 protocol version according to which the SSL library interprets
7910 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7911 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7912 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7913 specification (change_cipher_spec(20), alert(21), handshake(22)).
7914 'buf' and 'len' point to the actual message, 'ssl' to the
7915 SSL object, and 'arg' is the application-defined value set by
7916 SSL[_CTX]_set_msg_callback_arg().
7918 'openssl s_client' and 'openssl s_server' have new '-msg' options
7919 to enable a callback that displays all protocol messages.
7922 *) Change the shared library support so shared libraries are built as
7923 soon as the corresponding static library is finished, and thereby get
7924 openssl and the test programs linked against the shared library.
7925 This still only happens when the keyword "shard" has been given to
7926 the configuration scripts.
7928 NOTE: shared library support is still an experimental thing, and
7929 backward binary compatibility is still not guaranteed.
7930 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7932 *) Add support for Subject Information Access extension.
7933 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7935 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7936 additional bytes when new memory had to be allocated, not just
7937 when reusing an existing buffer.
7940 *) New command line and configuration option 'utf8' for the req command.
7941 This allows field values to be specified as UTF8 strings.
7944 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7945 runs for the former and machine-readable output for the latter.
7948 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7949 of the e-mail address in the DN (i.e., it will go into a certificate
7950 extension only). The new configuration file option 'email_in_dn = no'
7951 has the same effect.
7952 [Massimiliano Pala madwolf@openca.org]
7954 *) Change all functions with names starting with des_ to be starting
7955 with DES_ instead. Add wrappers that are compatible with libdes,
7956 but are named _ossl_old_des_*. Finally, add macros that map the
7957 des_* symbols to the corresponding _ossl_old_des_* if libdes
7958 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7959 desired, the des_* symbols will be mapped to DES_*, with one
7962 Since we provide two compatibility mappings, the user needs to
7963 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7964 compatibility is desired. The default (i.e., when that macro
7965 isn't defined) is OpenSSL 0.9.6c compatibility.
7967 There are also macros that enable and disable the support of old
7968 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7969 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7970 are defined, the default will apply: to support the old des routines.
7972 In either case, one must include openssl/des.h to get the correct
7973 definitions. Do not try to just include openssl/des_old.h, that
7976 NOTE: This is a major break of an old API into a new one. Software
7977 authors are encouraged to switch to the DES_ style functions. Some
7978 time in the future, des_old.h and the libdes compatibility functions
7979 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7980 default), and then completely removed.
7983 *) Test for certificates which contain unsupported critical extensions.
7984 If such a certificate is found during a verify operation it is
7985 rejected by default: this behaviour can be overridden by either
7986 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7987 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7988 X509_supported_extension() has also been added which returns 1 if a
7989 particular extension is supported.
7992 *) Modify the behaviour of EVP cipher functions in similar way to digests
7993 to retain compatibility with existing code.
7996 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7997 compatibility with existing code. In particular the 'ctx' parameter does
7998 not have to be to be initialized before the call to EVP_DigestInit() and
7999 it is tidied up after a call to EVP_DigestFinal(). New function
8000 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
8001 EVP_MD_CTX_copy() changed to not require the destination to be
8002 initialized valid and new function EVP_MD_CTX_copy_ex() added which
8003 requires the destination to be valid.
8005 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
8006 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
8009 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
8010 so that complete 'Handshake' protocol structures are kept in memory
8011 instead of overwriting 'msg_type' and 'length' with 'body' data.
8014 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
8015 [Massimo Santin via Richard Levitte]
8017 *) Major restructuring to the underlying ENGINE code. This includes
8018 reduction of linker bloat, separation of pure "ENGINE" manipulation
8019 (initialisation, etc) from functionality dealing with implementations
8020 of specific crypto interfaces. This change also introduces integrated
8021 support for symmetric ciphers and digest implementations - so ENGINEs
8022 can now accelerate these by providing EVP_CIPHER and EVP_MD
8023 implementations of their own. This is detailed in crypto/engine/README
8024 as it couldn't be adequately described here. However, there are a few
8025 API changes worth noting - some RSA, DSA, DH, and RAND functions that
8026 were changed in the original introduction of ENGINE code have now
8027 reverted back - the hooking from this code to ENGINE is now a good
8028 deal more passive and at run-time, operations deal directly with
8029 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
8030 dereferencing through an ENGINE pointer any more. Also, the ENGINE
8031 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
8032 they were not being used by the framework as there is no concept of a
8033 BIGNUM_METHOD and they could not be generalised to the new
8034 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
8035 ENGINE_cpy() has been removed as it cannot be consistently defined in
8039 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
8042 *) Change mkdef.pl to sort symbols that get the same entry number,
8043 and make sure the automatically generated functions ERR_load_*
8044 become part of libeay.num as well.
8047 *) New function SSL_renegotiate_pending(). This returns true once
8048 renegotiation has been requested (either SSL_renegotiate() call
8049 or HelloRequest/ClientHello received from the peer) and becomes
8050 false once a handshake has been completed.
8051 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
8052 sends a HelloRequest, but does not ensure that a handshake takes
8053 place. SSL_renegotiate_pending() is useful for checking if the
8054 client has followed the request.)
8057 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
8058 By default, clients may request session resumption even during
8059 renegotiation (if session ID contexts permit); with this option,
8060 session resumption is possible only in the first handshake.
8062 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
8063 more bits available for options that should not be part of
8064 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
8067 *) Add some demos for certificate and certificate request creation.
8070 *) Make maximum certificate chain size accepted from the peer application
8071 settable (SSL*_get/set_max_cert_list()), as proposed by
8072 "Douglas E. Engert" <deengert@anl.gov>.
8075 *) Add support for shared libraries for Unixware-7
8076 (Boyd Lynn Gerber <gerberb@zenez.com>).
8079 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
8080 be done prior to destruction. Use this to unload error strings from
8081 ENGINEs that load their own error strings. NB: This adds two new API
8082 functions to "get" and "set" this destroy handler in an ENGINE.
8085 *) Alter all existing ENGINE implementations (except "openssl" and
8086 "openbsd") to dynamically instantiate their own error strings. This
8087 makes them more flexible to be built both as statically-linked ENGINEs
8088 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
8089 Also, add stub code to each that makes building them as self-contained
8090 shared-libraries easier (see README.ENGINE).
8093 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
8094 implementations into applications that are completely implemented in
8095 self-contained shared-libraries. The "dynamic" ENGINE exposes control
8096 commands that can be used to configure what shared-library to load and
8097 to control aspects of the way it is handled. Also, made an update to
8098 the README.ENGINE file that brings its information up-to-date and
8099 provides some information and instructions on the "dynamic" ENGINE
8100 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
8103 *) Make it possible to unload ranges of ERR strings with a new
8104 "ERR_unload_strings" function.
8107 *) Add a copy() function to EVP_MD.
8110 *) Make EVP_MD routines take a context pointer instead of just the
8111 md_data void pointer.
8114 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
8115 that the digest can only process a single chunk of data
8116 (typically because it is provided by a piece of
8117 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
8118 is only going to provide a single chunk of data, and hence the
8119 framework needn't accumulate the data for oneshot drivers.
8122 *) As with "ERR", make it possible to replace the underlying "ex_data"
8123 functions. This change also alters the storage and management of global
8124 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
8125 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
8126 index counters. The API functions that use this state have been changed
8127 to take a "class_index" rather than pointers to the class's local STACK
8128 and counter, and there is now an API function to dynamically create new
8129 classes. This centralisation allows us to (a) plug a lot of the
8130 thread-safety problems that existed, and (b) makes it possible to clean
8131 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
8132 such data would previously have always leaked in application code and
8133 workarounds were in place to make the memory debugging turn a blind eye
8134 to it. Application code that doesn't use this new function will still
8135 leak as before, but their memory debugging output will announce it now
8136 rather than letting it slide.
8138 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
8139 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
8140 has a return value to indicate success or failure.
8143 *) Make it possible to replace the underlying "ERR" functions such that the
8144 global state (2 LHASH tables and 2 locks) is only used by the "default"
8145 implementation. This change also adds two functions to "get" and "set"
8146 the implementation prior to it being automatically set the first time
8147 any other ERR function takes place. Ie. an application can call "get",
8148 pass the return value to a module it has just loaded, and that module
8149 can call its own "set" function using that value. This means the
8150 module's "ERR" operations will use (and modify) the error state in the
8151 application and not in its own statically linked copy of OpenSSL code.
8154 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
8155 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
8156 the operation, and provides a more encapsulated way for external code
8157 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
8158 to use these functions rather than manually incrementing the counts.
8160 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
8163 *) Add EVP test program.
8166 *) Add symmetric cipher support to ENGINE. Expect the API to change!
8169 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
8170 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
8171 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
8172 These allow a CRL to be built without having to access X509_CRL fields
8173 directly. Modify 'ca' application to use new functions.
8176 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
8177 bug workarounds. Rollback attack detection is a security feature.
8178 The problem will only arise on OpenSSL servers when TLSv1 is not
8179 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
8180 Software authors not wanting to support TLSv1 will have special reasons
8181 for their choice and can explicitly enable this option.
8182 [Bodo Moeller, Lutz Jaenicke]
8184 *) Rationalise EVP so it can be extended: don't include a union of
8185 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
8186 (similar to those existing for EVP_CIPHER_CTX).
8191 EVP_MD_CTX_init(&md); /* new function call */
8192 EVP_DigestInit(&md, EVP_sha1());
8193 EVP_DigestUpdate(&md, in, len);
8194 EVP_DigestFinal(&md, out, NULL);
8195 EVP_MD_CTX_cleanup(&md); /* new function call */
8199 *) Make DES key schedule conform to the usual scheme, as well as
8200 correcting its structure. This means that calls to DES functions
8201 now have to pass a pointer to a des_key_schedule instead of a
8202 plain des_key_schedule (which was actually always a pointer
8205 des_key_schedule ks;
8207 des_set_key_checked(..., &ks);
8208 des_ncbc_encrypt(..., &ks, ...);
8210 (Note that a later change renames 'des_...' into 'DES_...'.)
8213 *) Initial reduction of linker bloat: the use of some functions, such as
8214 PEM causes large amounts of unused functions to be linked in due to
8215 poor organisation. For example pem_all.c contains every PEM function
8216 which has a knock on effect of linking in large amounts of (unused)
8217 ASN1 code. Grouping together similar functions and splitting unrelated
8218 functions prevents this.
8221 *) Cleanup of EVP macros.
8224 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
8225 correct _ecb suffix.
8228 *) Add initial OCSP responder support to ocsp application. The
8229 revocation information is handled using the text based index
8230 use by the ca application. The responder can either handle
8231 requests generated internally, supplied in files (for example
8232 via a CGI script) or using an internal minimal server.
8235 *) Add configuration choices to get zlib compression for TLS.
8238 *) Changes to Kerberos SSL for RFC 2712 compliance:
8239 1. Implemented real KerberosWrapper, instead of just using
8240 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
8241 2. Implemented optional authenticator field of KerberosWrapper.
8243 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
8244 and authenticator structs; see crypto/krb5/.
8246 Generalized Kerberos calls to support multiple Kerberos libraries.
8247 [Vern Staats <staatsvr@asc.hpc.mil>,
8248 Jeffrey Altman <jaltman@columbia.edu>
8249 via Richard Levitte]
8251 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
8252 already does with RSA. testdsa.h now has 'priv_key/pub_key'
8253 values for each of the key sizes rather than having just
8254 parameters (and 'speed' generating keys each time).
8257 *) Speed up EVP routines.
8260 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
8261 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
8262 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
8263 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
8265 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
8266 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
8267 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
8270 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
8272 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
8275 *) Added the OS2-EMX target.
8276 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
8278 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
8279 to support NCONF routines in extension code. New function CONF_set_nconf()
8280 to allow functions which take an NCONF to also handle the old LHASH
8281 structure: this means that the old CONF compatible routines can be
8282 retained (in particular wrt extensions) without having to duplicate the
8283 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
8286 *) Enhance the general user interface with mechanisms for inner control
8287 and with possibilities to have yes/no kind of prompts.
8290 *) Change all calls to low level digest routines in the library and
8291 applications to use EVP. Add missing calls to HMAC_cleanup() and
8292 don't assume HMAC_CTX can be copied using memcpy().
8293 [Verdon Walker <VWalker@novell.com>, Steve Henson]
8295 *) Add the possibility to control engines through control names but with
8296 arbitrary arguments instead of just a string.
8297 Change the key loaders to take a UI_METHOD instead of a callback
8298 function pointer. NOTE: this breaks binary compatibility with earlier
8299 versions of OpenSSL [engine].
8300 Adapt the nCipher code for these new conditions and add a card insertion
8304 *) Enhance the general user interface with mechanisms to better support
8305 dialog box interfaces, application-defined prompts, the possibility
8306 to use defaults (for example default passwords from somewhere else)
8307 and interrupts/cancellations.
8310 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
8311 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
8314 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
8315 tidy up some unnecessarily weird code in 'sk_new()').
8316 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8318 *) Change the key loading routines for ENGINEs to use the same kind
8319 callback (pem_password_cb) as all other routines that need this
8323 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8324 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8325 than this minimum value is recommended.
8328 *) New random seeder for OpenVMS, using the system process statistics
8329 that are easily reachable.
8332 *) Windows apparently can't transparently handle global
8333 variables defined in DLLs. Initialisations such as:
8335 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8337 won't compile. This is used by the any applications that need to
8338 declare their own ASN1 modules. This was fixed by adding the option
8339 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8340 needed for static libraries under Win32.
8343 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8344 setting of purpose and trust fields. New X509_STORE trust and
8345 purpose functions and tidy up setting in other SSL functions.
8348 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8349 structure. These are inherited by X509_STORE_CTX when it is
8350 initialised. This allows various defaults to be set in the
8351 X509_STORE structure (such as flags for CRL checking and custom
8352 purpose or trust settings) for functions which only use X509_STORE_CTX
8353 internally such as S/MIME.
8355 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8356 trust settings if they are not set in X509_STORE. This allows X509_STORE
8357 purposes and trust (in S/MIME for example) to override any set by default.
8359 Add command line options for CRL checking to smime, s_client and s_server
8363 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8364 are set then the CRL is looked up in the X509_STORE structure and
8365 its validity and signature checked, then if the certificate is found
8366 in the CRL the verify fails with a revoked error.
8368 Various new CRL related callbacks added to X509_STORE_CTX structure.
8370 Command line options added to 'verify' application to support this.
8372 This needs some additional work, such as being able to handle multiple
8373 CRLs with different times, extension based lookup (rather than just
8374 by subject name) and ultimately more complete V2 CRL extension
8378 *) Add a general user interface API (crypto/ui/). This is designed
8379 to replace things like des_read_password and friends (backward
8380 compatibility functions using this new API are provided).
8381 The purpose is to remove prompting functions from the DES code
8382 section as well as provide for prompting through dialog boxes in
8383 a window system and the like.
8386 *) Add "ex_data" support to ENGINE so implementations can add state at a
8387 per-structure level rather than having to store it globally.
8390 *) Make it possible for ENGINE structures to be copied when retrieved by
8391 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8392 This causes the "original" ENGINE structure to act like a template,
8393 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8394 operational state can be localised to each ENGINE structure, despite the
8395 fact they all share the same "methods". New ENGINE structures returned in
8396 this case have no functional references and the return value is the single
8397 structural reference. This matches the single structural reference returned
8398 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8402 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8403 needs to match any other type at all we need to manually clear the
8407 *) Changes to the "openssl engine" utility to include;
8408 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8409 about an ENGINE's available control commands.
8410 - executing control commands from command line arguments using the
8411 '-pre' and '-post' switches. '-post' is only used if '-t' is
8412 specified and the ENGINE is successfully initialised. The syntax for
8413 the individual commands are colon-separated, for example;
8414 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8417 *) New dynamic control command support for ENGINEs. ENGINEs can now
8418 declare their own commands (numbers), names (strings), descriptions,
8419 and input types for run-time discovery by calling applications. A
8420 subset of these commands are implicitly classed as "executable"
8421 depending on their input type, and only these can be invoked through
8422 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8423 can be based on user input, config files, etc). The distinction is
8424 that "executable" commands cannot return anything other than a boolean
8425 result and can only support numeric or string input, whereas some
8426 discoverable commands may only be for direct use through
8427 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8428 pointers, or other custom uses. The "executable" commands are to
8429 support parameterisations of ENGINE behaviour that can be
8430 unambiguously defined by ENGINEs and used consistently across any
8431 OpenSSL-based application. Commands have been added to all the
8432 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8433 control over shared-library paths without source code alterations.
8436 *) Changed all ENGINE implementations to dynamically allocate their
8437 ENGINEs rather than declaring them statically. Apart from this being
8438 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8439 this also allows the implementations to compile without using the
8440 internal engine_int.h header.
8443 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8444 'const' value. Any code that should be able to modify a RAND_METHOD
8445 should already have non-const pointers to it (ie. they should only
8446 modify their own ones).
8449 *) Made a variety of little tweaks to the ENGINE code.
8450 - "atalla" and "ubsec" string definitions were moved from header files
8451 to C code. "nuron" string definitions were placed in variables
8452 rather than hard-coded - allowing parameterisation of these values
8453 later on via ctrl() commands.
8454 - Removed unused "#if 0"'d code.
8455 - Fixed engine list iteration code so it uses ENGINE_free() to release
8456 structural references.
8457 - Constified the RAND_METHOD element of ENGINE structures.
8458 - Constified various get/set functions as appropriate and added
8459 missing functions (including a catch-all ENGINE_cpy that duplicates
8460 all ENGINE values onto a new ENGINE except reference counts/state).
8461 - Removed NULL parameter checks in get/set functions. Setting a method
8462 or function to NULL is a way of cancelling out a previously set
8463 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8464 and doesn't justify the extra error symbols and code.
8465 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8466 flags from engine_int.h to engine.h.
8467 - Changed prototypes for ENGINE handler functions (init(), finish(),
8468 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8471 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8472 to the algorithm using long division. The binary algorithm can be
8473 used only if the modulus is odd. On 32-bit systems, it is faster
8474 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8475 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8476 up to 450 bits. In 64-bit environments, the binary algorithm
8477 appears to be advantageous for much longer moduli; here we use it
8478 for moduli up to 2048 bits.
8481 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8482 could not support the combine flag in choice fields.
8485 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8486 extensions from a certificate request to the certificate.
8489 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8490 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8491 file: this allows the display of the certificate about to be
8492 signed to be customised, to allow certain fields to be included
8493 or excluded and extension details. The old system didn't display
8494 multicharacter strings properly, omitted fields not in the policy
8495 and couldn't display additional details such as extensions.
8498 *) Function EC_POINTs_mul for multiple scalar multiplication
8499 of an arbitrary number of elliptic curve points
8500 \sum scalars[i]*points[i],
8501 optionally including the generator defined for the EC_GROUP:
8502 scalar*generator + \sum scalars[i]*points[i].
8504 EC_POINT_mul is a simple wrapper function for the typical case
8505 that the point list has just one item (besides the optional
8509 *) First EC_METHODs for curves over GF(p):
8511 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8512 operations and provides various method functions that can also
8513 operate with faster implementations of modular arithmetic.
8515 EC_GFp_mont_method() reuses most functions that are part of
8516 EC_GFp_simple_method, but uses Montgomery arithmetic.
8518 [Bodo Moeller; point addition and point doubling
8519 implementation directly derived from source code provided by
8520 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8522 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8523 crypto/ec/ec_lib.c):
8525 Curves are EC_GROUP objects (with an optional group generator)
8526 based on EC_METHODs that are built into the library.
8528 Points are EC_POINT objects based on EC_GROUP objects.
8530 Most of the framework would be able to handle curves over arbitrary
8531 finite fields, but as there are no obvious types for fields other
8532 than GF(p), some functions are limited to that for now.
8535 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8536 that the file contains a complete HTTP response.
8539 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8540 change the def and num file printf format specifier from "%-40sXXX"
8541 to "%-39s XXX". The latter will always guarantee a space after the
8542 field while the former will cause them to run together if the field
8543 is 40 of more characters long.
8546 *) Constify the cipher and digest 'method' functions and structures
8547 and modify related functions to take constant EVP_MD and EVP_CIPHER
8551 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8552 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8555 *) Modify EVP_Digest*() routines so they now return values. Although the
8556 internal software routines can never fail additional hardware versions
8560 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8562 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8563 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8566 ERR_R_NESTED_ASN1_ERROR
8568 ERR_R_MISSING_ASN1_EOS
8569 were 4 .. 9, conflicting with
8570 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8572 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8573 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8575 Add new error code 'ERR_R_INTERNAL_ERROR'.
8578 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8582 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8583 sets the subject name for a new request or supersedes the
8584 subject name in a given request. Formats that can be parsed are
8585 'CN=Some Name, OU=myOU, C=IT'
8587 'CN=Some Name/OU=myOU/C=IT'.
8589 Add options '-batch' and '-verbose' to 'openssl req'.
8590 [Massimiliano Pala <madwolf@hackmasters.net>]
8592 *) Introduce the possibility to access global variables through
8593 functions on platform were that's the best way to handle exporting
8594 global variables in shared libraries. To enable this functionality,
8595 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8596 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8597 is normally done by Configure or something similar).
8599 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8600 in the source file (foo.c) like this:
8602 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8603 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8605 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8606 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8608 OPENSSL_DECLARE_GLOBAL(int,foo);
8609 #define foo OPENSSL_GLOBAL_REF(foo)
8610 OPENSSL_DECLARE_GLOBAL(double,bar);
8611 #define bar OPENSSL_GLOBAL_REF(bar)
8613 The #defines are very important, and therefore so is including the
8614 header file everywhere where the defined globals are used.
8616 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8617 of ASN.1 items, but that structure is a bit different.
8619 The largest change is in util/mkdef.pl which has been enhanced with
8620 better and easier to understand logic to choose which symbols should
8621 go into the Windows .def files as well as a number of fixes and code
8622 cleanup (among others, algorithm keywords are now sorted
8623 lexicographically to avoid constant rewrites).
8626 *) In BN_div() keep a copy of the sign of 'num' before writing the
8627 result to 'rm' because if rm==num the value will be overwritten
8628 and produce the wrong result if 'num' is negative: this caused
8629 problems with BN_mod() and BN_nnmod().
8632 *) Function OCSP_request_verify(). This checks the signature on an
8633 OCSP request and verifies the signer certificate. The signer
8634 certificate is just checked for a generic purpose and OCSP request
8638 *) Add OCSP_check_validity() function to check the validity of OCSP
8639 responses. OCSP responses are prepared in real time and may only
8640 be a few seconds old. Simply checking that the current time lies
8641 between thisUpdate and nextUpdate max reject otherwise valid responses
8642 caused by either OCSP responder or client clock inaccuracy. Instead
8643 we allow thisUpdate and nextUpdate to fall within a certain period of
8644 the current time. The age of the response can also optionally be
8645 checked. Two new options -validity_period and -status_age added to
8649 *) If signature or public key algorithm is unrecognized print out its
8650 OID rather that just UNKNOWN.
8653 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8654 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8655 ID to be generated from the issuer certificate alone which can then be
8656 passed to OCSP_id_issuer_cmp().
8659 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8660 ASN1 modules to export functions returning ASN1_ITEM pointers
8661 instead of the ASN1_ITEM structures themselves. This adds several
8662 new macros which allow the underlying ASN1 function/structure to
8663 be accessed transparently. As a result code should not use ASN1_ITEM
8664 references directly (such as &X509_it) but instead use the relevant
8665 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8666 use of the new ASN1 code on platforms where exporting structures
8667 is problematical (for example in shared libraries) but exporting
8668 functions returning pointers to structures is not.
8671 *) Add support for overriding the generation of SSL/TLS session IDs.
8672 These callbacks can be registered either in an SSL_CTX or per SSL.
8673 The purpose of this is to allow applications to control, if they wish,
8674 the arbitrary values chosen for use as session IDs, particularly as it
8675 can be useful for session caching in multiple-server environments. A
8676 command-line switch for testing this (and any client code that wishes
8677 to use such a feature) has been added to "s_server".
8678 [Geoff Thorpe, Lutz Jaenicke]
8680 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8681 of the form '#if defined(...) || defined(...) || ...' and
8682 '#if !defined(...) && !defined(...) && ...'. This also avoids
8683 the growing number of special cases it was previously handling.
8686 *) Make all configuration macros available for application by making
8687 sure they are available in opensslconf.h, by giving them names starting
8688 with "OPENSSL_" to avoid conflicts with other packages and by making
8689 sure e_os2.h will cover all platform-specific cases together with
8691 Additionally, it is now possible to define configuration/platform-
8692 specific names (called "system identities"). In the C code, these
8693 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8694 macro with the name beginning with "OPENSSL_SYS_", which is determined
8695 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8699 *) New option -set_serial to 'req' and 'x509' this allows the serial
8700 number to use to be specified on the command line. Previously self
8701 signed certificates were hard coded with serial number 0 and the
8702 CA options of 'x509' had to use a serial number in a file which was
8706 *) New options to 'ca' utility to support V2 CRL entry extensions.
8707 Currently CRL reason, invalidity date and hold instruction are
8708 supported. Add new CRL extensions to V3 code and some new objects.
8711 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8712 disable standard block padding (aka PKCS#5 padding) in the EVP
8713 API, which was previously mandatory. This means that the data is
8714 not padded in any way and so the total length much be a multiple
8715 of the block size, otherwise an error occurs.
8718 *) Initial (incomplete) OCSP SSL support.
8721 *) New function OCSP_parse_url(). This splits up a URL into its host,
8722 port and path components: primarily to parse OCSP URLs. New -url
8723 option to ocsp utility.
8726 *) New nonce behavior. The return value of OCSP_check_nonce() now
8727 reflects the various checks performed. Applications can decide
8728 whether to tolerate certain situations such as an absent nonce
8729 in a response when one was present in a request: the ocsp application
8730 just prints out a warning. New function OCSP_add1_basic_nonce()
8731 this is to allow responders to include a nonce in a response even if
8732 the request is nonce-less.
8735 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8736 skipped when using openssl x509 multiple times on a single input file,
8737 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8740 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8741 set string type: to handle setting ASN1_TIME structures. Fix ca
8742 utility to correctly initialize revocation date of CRLs.
8745 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8746 the clients preferred ciphersuites and rather use its own preferences.
8747 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8748 Internet Explorer by ensuring unchanged hash method during stepup.
8749 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8752 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8753 to aes and add a new 'exist' option to print out symbols that don't
8757 *) Additional options to ocsp utility to allow flags to be set and
8758 additional certificates supplied.
8761 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8762 OCSP client a number of certificate to only verify the response
8766 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8767 handle the new API. Currently only ECB, CBC modes supported. Add new
8770 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8771 Encryption Standard (AES) Ciphersuites for Transport Layer
8772 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8773 not enabled by default and were not part of the "ALL" ciphersuite
8774 alias because they were not yet official; they could be
8775 explicitly requested by specifying the "AESdraft" ciphersuite
8776 group alias. In the final release of OpenSSL 0.9.7, the group
8777 alias is called "AES" and is part of "ALL".)
8778 [Ben Laurie, Steve Henson, Bodo Moeller]
8780 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8781 request to response.
8784 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8785 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8786 extract information from a certificate request. OCSP_response_create()
8787 creates a response and optionally adds a basic response structure.
8788 OCSP_basic_add1_status() adds a complete single response to a basic
8789 response and returns the OCSP_SINGLERESP structure just added (to allow
8790 extensions to be included for example). OCSP_basic_add1_cert() adds a
8791 certificate to a basic response and OCSP_basic_sign() signs a basic
8792 response with various flags. New helper functions ASN1_TIME_check()
8793 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8794 (converts ASN1_TIME to GeneralizedTime).
8797 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8798 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8799 structure from a certificate. X509_pubkey_digest() digests the public_key
8800 contents: this is used in various key identifiers.
8803 *) Make sk_sort() tolerate a NULL argument.
8804 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8806 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8807 passed by the function are trusted implicitly. If any of them signed the
8808 response then it is assumed to be valid and is not verified.
8811 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8812 to data. This was previously part of the PKCS7 ASN1 code. This
8813 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8814 [Steve Henson, reported by Kenneth R. Robinette
8815 <support@securenetterm.com>]
8817 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8818 routines: without these tracing memory leaks is very painful.
8819 Fix leaks in PKCS12 and PKCS7 routines.
8822 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8823 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8824 effectively meant GeneralizedTime would never be used. Now it
8825 is initialised to -1 but X509_time_adj() now has to check the value
8826 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8827 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8828 [Steve Henson, reported by Kenneth R. Robinette
8829 <support@securenetterm.com>]
8831 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8832 result in a zero length in the ASN1_INTEGER structure which was
8833 not consistent with the structure when d2i_ASN1_INTEGER() was used
8834 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8835 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8836 where it did not print out a minus for negative ASN1_INTEGER.
8839 *) Add summary printout to ocsp utility. The various functions which
8840 convert status values to strings have been renamed to:
8841 OCSP_response_status_str(), OCSP_cert_status_str() and
8842 OCSP_crl_reason_str() and are no longer static. New options
8843 to verify nonce values and to disable verification. OCSP response
8844 printout format cleaned up.
8847 *) Add additional OCSP certificate checks. These are those specified
8848 in RFC2560. This consists of two separate checks: the CA of the
8849 certificate being checked must either be the OCSP signer certificate
8850 or the issuer of the OCSP signer certificate. In the latter case the
8851 OCSP signer certificate must contain the OCSP signing extended key
8852 usage. This check is performed by attempting to match the OCSP
8853 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8854 in the OCSP_CERTID structures of the response.
8857 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8858 and related routines. This uses the standard OpenSSL certificate
8859 verify routines to perform initial checks (just CA validity) and
8860 to obtain the certificate chain. Then additional checks will be
8861 performed on the chain. Currently the root CA is checked to see
8862 if it is explicitly trusted for OCSP signing. This is used to set
8863 a root CA as a global signing root: that is any certificate that
8864 chains to that CA is an acceptable OCSP signing certificate.
8867 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8868 extensions from a separate configuration file.
8869 As when reading extensions from the main configuration file,
8870 the '-extensions ...' option may be used for specifying the
8872 [Massimiliano Pala <madwolf@comune.modena.it>]
8874 *) New OCSP utility. Allows OCSP requests to be generated or
8875 read. The request can be sent to a responder and the output
8876 parsed, outputted or printed in text form. Not complete yet:
8877 still needs to check the OCSP response validity.
8880 *) New subcommands for 'openssl ca':
8881 'openssl ca -status <serial>' prints the status of the cert with
8882 the given serial number (according to the index file).
8883 'openssl ca -updatedb' updates the expiry status of certificates
8885 [Massimiliano Pala <madwolf@comune.modena.it>]
8887 *) New '-newreq-nodes' command option to CA.pl. This is like
8888 '-newreq', but calls 'openssl req' with the '-nodes' option
8889 so that the resulting key is not encrypted.
8890 [Damien Miller <djm@mindrot.org>]
8892 *) New configuration for the GNU Hurd.
8893 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8895 *) Initial code to implement OCSP basic response verify. This
8896 is currently incomplete. Currently just finds the signer's
8897 certificate and verifies the signature on the response.
8900 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8901 value of OPENSSLDIR. This is available via the new '-d' option
8902 to 'openssl version', and is also included in 'openssl version -a'.
8905 *) Allowing defining memory allocation callbacks that will be given
8906 file name and line number information in additional arguments
8907 (a const char* and an int). The basic functionality remains, as
8908 well as the original possibility to just replace malloc(),
8909 realloc() and free() by functions that do not know about these
8910 additional arguments. To register and find out the current
8911 settings for extended allocation functions, the following
8912 functions are provided:
8914 CRYPTO_set_mem_ex_functions
8915 CRYPTO_set_locked_mem_ex_functions
8916 CRYPTO_get_mem_ex_functions
8917 CRYPTO_get_locked_mem_ex_functions
8919 These work the same way as CRYPTO_set_mem_functions and friends.
8920 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8921 extended allocation function is enabled.
8922 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8923 a conventional allocation function is enabled.
8924 [Richard Levitte, Bodo Moeller]
8926 *) Finish off removing the remaining LHASH function pointer casts.
8927 There should no longer be any prototype-casting required when using
8928 the LHASH abstraction, and any casts that remain are "bugs". See
8929 the callback types and macros at the head of lhash.h for details
8930 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8933 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8934 If /dev/[u]random devices are not available or do not return enough
8935 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8937 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8938 /etc/entropy will be queried once each in this sequence, querying stops
8939 when enough entropy was collected without querying more sockets.
8942 *) Change the Unix RAND_poll() variant to be able to poll several
8943 random devices, as specified by DEVRANDOM, until a sufficient amount
8944 of data has been collected. We spend at most 10 ms on each file
8945 (select timeout) and read in non-blocking mode. DEVRANDOM now
8946 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8947 (previously it was just the string "/dev/urandom"), so on typical
8948 platforms the 10 ms delay will never occur.
8949 Also separate out the Unix variant to its own file, rand_unix.c.
8950 For VMS, there's a currently-empty rand_vms.c.
8953 *) Move OCSP client related routines to ocsp_cl.c. These
8954 provide utility functions which an application needing
8955 to issue a request to an OCSP responder and analyse the
8956 response will typically need: as opposed to those which an
8957 OCSP responder itself would need which will be added later.
8959 OCSP_request_sign() signs an OCSP request with an API similar
8960 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8961 response. OCSP_response_get1_basic() extracts basic response
8962 from response. OCSP_resp_find_status(): finds and extracts status
8963 information from an OCSP_CERTID structure (which will be created
8964 when the request structure is built). These are built from lower
8965 level functions which work on OCSP_SINGLERESP structures but
8966 won't normally be used unless the application wishes to examine
8967 extensions in the OCSP response for example.
8969 Replace nonce routines with a pair of functions.
8970 OCSP_request_add1_nonce() adds a nonce value and optionally
8971 generates a random value. OCSP_check_nonce() checks the
8972 validity of the nonce in an OCSP response.
8975 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8976 This doesn't copy the supplied OCSP_CERTID and avoids the
8977 need to free up the newly created id. Change return type
8978 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8979 This can then be used to add extensions to the request.
8980 Deleted OCSP_request_new(), since most of its functionality
8981 is now in OCSP_REQUEST_new() (and the case insensitive name
8982 clash) apart from the ability to set the request name which
8983 will be added elsewhere.
8986 *) Update OCSP API. Remove obsolete extensions argument from
8987 various functions. Extensions are now handled using the new
8988 OCSP extension code. New simple OCSP HTTP function which
8989 can be used to send requests and parse the response.
8992 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8993 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8994 uses the special reorder version of SET OF to sort the attributes
8995 and reorder them to match the encoded order. This resolves a long
8996 standing problem: a verify on a PKCS7 structure just after signing
8997 it used to fail because the attribute order did not match the
8998 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8999 it uses the received order. This is necessary to tolerate some broken
9000 software that does not order SET OF. This is handled by encoding
9001 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
9002 to produce the required SET OF.
9005 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
9006 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
9007 files to get correct declarations of the ASN.1 item variables.
9010 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
9011 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
9012 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
9013 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
9014 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
9015 ASN1_ITEM and no wrapper functions.
9018 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
9019 replace the old function pointer based I/O routines. Change most of
9020 the *_d2i_bio() and *_d2i_fp() functions to use these.
9023 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
9024 lines, recognize more "algorithms" that can be deselected, and make
9025 it complain about algorithm deselection that isn't recognised.
9028 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
9029 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
9030 to use new functions. Add NO_ASN1_OLD which can be set to remove
9031 some old style ASN1 functions: this can be used to determine if old
9032 code will still work when these eventually go away.
9035 *) New extension functions for OCSP structures, these follow the
9036 same conventions as certificates and CRLs.
9039 *) New function X509V3_add1_i2d(). This automatically encodes and
9040 adds an extension. Its behaviour can be customised with various
9041 flags to append, replace or delete. Various wrappers added for
9042 certificates and CRLs.
9045 *) Fix to avoid calling the underlying ASN1 print routine when
9046 an extension cannot be parsed. Correct a typo in the
9047 OCSP_SERVICELOC extension. Tidy up print OCSP format.
9050 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
9051 entries for variables.
9054 *) Add functionality to apps/openssl.c for detecting locking
9055 problems: As the program is single-threaded, all we have
9056 to do is register a locking callback using an array for
9057 storing which locks are currently held by the program.
9060 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
9061 SSL_get_ex_data_X509_STORE_idx(), which is used in
9062 ssl_verify_cert_chain() and thus can be called at any time
9063 during TLS/SSL handshakes so that thread-safety is essential.
9064 Unfortunately, the ex_data design is not at all suited
9065 for multi-threaded use, so it probably should be abolished.
9068 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
9069 [Broadcom, tweaked and integrated by Geoff Thorpe]
9071 *) Move common extension printing code to new function
9072 X509V3_print_extensions(). Reorganise OCSP print routines and
9073 implement some needed OCSP ASN1 functions. Add OCSP extensions.
9076 *) New function X509_signature_print() to remove duplication in some
9080 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
9081 set (this was treated exactly the same as SET OF previously). This
9082 is used to reorder the STACK representing the structure to match the
9083 encoding. This will be used to get round a problem where a PKCS7
9084 structure which was signed could not be verified because the STACK
9085 order did not reflect the encoded order.
9088 *) Reimplement the OCSP ASN1 module using the new code.
9091 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
9092 for its ASN1 operations. The old style function pointers still exist
9093 for now but they will eventually go away.
9096 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
9097 completely replaces the old ASN1 functionality with a table driven
9098 encoder and decoder which interprets an ASN1_ITEM structure describing
9099 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
9100 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
9101 has also been converted to the new form.
9104 *) Change BN_mod_exp_recp so that negative moduli are tolerated
9105 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
9106 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
9107 for negative moduli.
9110 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
9111 of not touching the result's sign bit.
9114 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
9118 *) Changed the LHASH code to use prototypes for callbacks, and created
9119 macros to declare and implement thin (optionally static) functions
9120 that provide type-safety and avoid function pointer casting for the
9121 type-specific callbacks.
9124 *) Added Kerberos Cipher Suites to be used with TLS, as written in
9126 [Veers Staats <staatsvr@asc.hpc.mil>,
9127 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
9129 *) Reformat the FAQ so the different questions and answers can be divided
9130 in sections depending on the subject.
9133 *) Have the zlib compression code load ZLIB.DLL dynamically under
9137 *) New function BN_mod_sqrt for computing square roots modulo a prime
9138 (using the probabilistic Tonelli-Shanks algorithm unless
9139 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
9140 be handled deterministically).
9141 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9143 *) Make BN_mod_inverse faster by explicitly handling small quotients
9144 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
9145 512 bits], about 30% for larger ones [1024 or 2048 bits].)
9148 *) New function BN_kronecker.
9151 *) Fix BN_gcd so that it works on negative inputs; the result is
9152 positive unless both parameters are zero.
9153 Previously something reasonably close to an infinite loop was
9154 possible because numbers could be growing instead of shrinking
9155 in the implementation of Euclid's algorithm.
9158 *) Fix BN_is_word() and BN_is_one() macros to take into account the
9159 sign of the number in question.
9161 Fix BN_is_word(a,w) to work correctly for w == 0.
9163 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
9164 because its test if the absolute value of 'a' equals 'w'.
9165 Note that BN_abs_is_word does *not* handle w == 0 reliably;
9166 it exists mostly for use in the implementations of BN_is_zero(),
9167 BN_is_one(), and BN_is_word().
9170 *) New function BN_swap.
9173 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
9174 the exponentiation functions are more likely to produce reasonable
9175 results on negative inputs.
9178 *) Change BN_mod_mul so that the result is always non-negative.
9179 Previously, it could be negative if one of the factors was negative;
9180 I don't think anyone really wanted that behaviour.
9183 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
9184 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
9185 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
9186 and add new functions:
9195 BN_mod_lshift1_quick
9199 These functions always generate non-negative results.
9201 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
9202 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
9204 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
9205 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
9206 be reduced modulo m.
9207 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
9210 The following entry accidentally appeared in the CHANGES file
9211 distributed with OpenSSL 0.9.7. The modifications described in
9212 it do *not* apply to OpenSSL 0.9.7.
9214 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
9215 was actually never needed) and in BN_mul(). The removal in BN_mul()
9216 required a small change in bn_mul_part_recursive() and the addition
9217 of the functions bn_cmp_part_words(), bn_sub_part_words() and
9218 bn_add_part_words(), which do the same thing as bn_cmp_words(),
9219 bn_sub_words() and bn_add_words() except they take arrays with
9224 *) In 'openssl passwd', verify passwords read from the terminal
9225 unless the '-salt' option is used (which usually means that
9226 verification would just waste user's time since the resulting
9227 hash is going to be compared with some given password hash)
9228 or the new '-noverify' option is used.
9230 This is an incompatible change, but it does not affect
9231 non-interactive use of 'openssl passwd' (passwords on the command
9232 line, '-stdin' option, '-in ...' option) and thus should not
9236 *) Remove all references to RSAref, since there's no more need for it.
9239 *) Make DSO load along a path given through an environment variable
9240 (SHLIB_PATH) with shl_load().
9243 *) Constify the ENGINE code as a result of BIGNUM constification.
9244 Also constify the RSA code and most things related to it. In a
9245 few places, most notable in the depth of the ASN.1 code, ugly
9246 casts back to non-const were required (to be solved at a later
9250 *) Make it so the openssl application has all engines loaded by default.
9253 *) Constify the BIGNUM routines a little more.
9256 *) Add the following functions:
9258 ENGINE_load_cswift()
9260 ENGINE_load_atalla()
9262 ENGINE_load_builtin_engines()
9264 That way, an application can itself choose if external engines that
9265 are built-in in OpenSSL shall ever be used or not. The benefit is
9266 that applications won't have to be linked with libdl or other dso
9267 libraries unless it's really needed.
9269 Changed 'openssl engine' to load all engines on demand.
9270 Changed the engine header files to avoid the duplication of some
9271 declarations (they differed!).
9274 *) 'openssl engine' can now list capabilities.
9277 *) Better error reporting in 'openssl engine'.
9280 *) Never call load_dh_param(NULL) in s_server.
9283 *) Add engine application. It can currently list engines by name and
9284 identity, and test if they are actually available.
9287 *) Improve RPM specification file by forcing symbolic linking and making
9288 sure the installed documentation is also owned by root.root.
9289 [Damien Miller <djm@mindrot.org>]
9291 *) Give the OpenSSL applications more possibilities to make use of
9292 keys (public as well as private) handled by engines.
9295 *) Add OCSP code that comes from CertCo.
9298 *) Add VMS support for the Rijndael code.
9301 *) Added untested support for Nuron crypto accelerator.
9304 *) Add support for external cryptographic devices. This code was
9305 previously distributed separately as the "engine" branch.
9306 [Geoff Thorpe, Richard Levitte]
9308 *) Rework the filename-translation in the DSO code. It is now possible to
9309 have far greater control over how a "name" is turned into a filename
9310 depending on the operating environment and any oddities about the
9311 different shared library filenames on each system.
9314 *) Support threads on FreeBSD-elf in Configure.
9317 *) Fix for SHA1 assembly problem with MASM: it produces
9318 warnings about corrupt line number information when assembling
9319 with debugging information. This is caused by the overlapping
9321 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9324 NCONF_get_number() has no error checking at all. As a replacement,
9325 NCONF_get_number_e() is defined (_e for "error checking") and is
9326 promoted strongly. The old NCONF_get_number is kept around for
9327 binary backward compatibility.
9328 Make it possible for methods to load from something other than a BIO,
9329 by providing a function pointer that is given a name instead of a BIO.
9330 For example, this could be used to load configuration data from an
9334 *) Fix for non blocking accept BIOs. Added new I/O special reason
9335 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9336 with non blocking I/O was not possible because no retry code was
9337 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9341 *) Added the beginnings of Rijndael support.
9344 *) Fix for bug in DirectoryString mask setting. Add support for
9345 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9346 to allow certificate printing to more controllable, additional
9347 'certopt' option to 'x509' to allow new printing options to be
9351 *) Clean old EAY MD5 hack from e_os.h.
9354 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9356 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9357 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9358 [Joe Orton, Steve Henson]
9360 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9362 *) Fix additional bug revealed by the NISCC test suite:
9364 Stop bug triggering large recursion when presented with
9365 certain ASN.1 tags (CVE-2003-0851)
9368 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9370 *) Fix various bugs revealed by running the NISCC test suite:
9372 Stop out of bounds reads in the ASN1 code when presented with
9373 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9375 If verify callback ignores invalid public key errors don't try to check
9376 certificate signature with the NULL public key.
9380 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9381 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9385 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9386 extra data after the compression methods not only for TLS 1.0
9387 but also for SSL 3.0 (as required by the specification).
9388 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9390 *) Change X509_certificate_type() to mark the key as exported/exportable
9391 when it's 512 *bits* long, not 512 bytes.
9394 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9396 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9397 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9398 a protocol version number mismatch like a decryption error
9399 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9402 *) Turn on RSA blinding by default in the default implementation
9403 to avoid a timing attack. Applications that don't want it can call
9404 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9405 They would be ill-advised to do so in most cases.
9406 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9408 *) Change RSA blinding code so that it works when the PRNG is not
9409 seeded (in this case, the secret RSA exponent is abused as
9410 an unpredictable seed -- if it is not unpredictable, there
9411 is no point in blinding anyway). Make RSA blinding thread-safe
9412 by remembering the creator's thread ID in rsa->blinding and
9413 having all other threads use local one-time blinding factors
9414 (this requires more computation than sharing rsa->blinding, but
9415 avoids excessive locking; and if an RSA object is not shared
9416 between threads, blinding will still be very fast).
9419 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9421 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9422 via timing by performing a MAC computation even if incorrect
9423 block cipher padding has been found. This is a countermeasure
9424 against active attacks where the attacker has to distinguish
9425 between bad padding and a MAC verification error. (CVE-2003-0078)
9427 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9428 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9429 Martin Vuagnoux (EPFL, Ilion)]
9431 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9433 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9434 memory from its contents. This is done with a counter that will
9435 place alternating values in each byte. This can be used to solve
9436 two issues: 1) the removal of calls to memset() by highly optimizing
9437 compilers, and 2) cleansing with other values than 0, since those can
9438 be read through on certain media, for example a swap space on disk.
9441 *) Bugfix: client side session caching did not work with external caching,
9442 because the session->cipher setting was not restored when reloading
9443 from the external cache. This problem was masked, when
9444 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9445 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9448 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9449 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9450 [Zeev Lieber <zeev-l@yahoo.com>]
9452 *) Undo an undocumented change introduced in 0.9.6e which caused
9453 repeated calls to OpenSSL_add_all_ciphers() and
9454 OpenSSL_add_all_digests() to be ignored, even after calling
9458 *) Change the default configuration reader to deal with last line not
9459 being properly terminated.
9462 *) Change X509_NAME_cmp() so it applies the special rules on handling
9463 DN values that are of type PrintableString, as well as RDNs of type
9464 emailAddress where the value has the type ia5String.
9465 [stefank@valicert.com via Richard Levitte]
9467 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9468 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9469 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9470 the bitwise-OR of the two for use by the majority of applications
9471 wanting this behaviour, and update the docs. The documented
9472 behaviour and actual behaviour were inconsistent and had been
9473 changing anyway, so this is more a bug-fix than a behavioural
9475 [Geoff Thorpe, diagnosed by Nadav Har'El]
9477 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9478 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9481 *) Fix initialization code race conditions in
9482 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9483 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9484 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9485 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9486 ssl2_get_cipher_by_char(),
9487 ssl3_get_cipher_by_char().
9488 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9490 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9491 the cached sessions are flushed, as the remove_cb() might use ex_data
9492 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9493 (see [openssl.org #212]).
9494 [Geoff Thorpe, Lutz Jaenicke]
9496 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9497 length, instead of the encoding length to d2i_ASN1_OBJECT.
9500 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9502 *) [In 0.9.6g-engine release:]
9503 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9504 [Lynn Gazis <lgazis@rainbow.com>]
9506 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9508 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9509 and get fix the header length calculation.
9510 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9511 Alon Kantor <alonk@checkpoint.com> (and others),
9514 *) Use proper error handling instead of 'assertions' in buffer
9515 overflow checks added in 0.9.6e. This prevents DoS (the
9516 assertions could call abort()).
9517 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9519 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9521 *) Add various sanity checks to asn1_get_length() to reject
9522 the ASN1 length bytes if they exceed sizeof(long), will appear
9523 negative or the content length exceeds the length of the
9525 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9527 *) Fix cipher selection routines: ciphers without encryption had no flags
9528 for the cipher strength set and where therefore not handled correctly
9529 by the selection routines (PR #130).
9532 *) Fix EVP_dsa_sha macro.
9536 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9537 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9538 that was added in OpenSSL 0.9.6d.
9540 As the countermeasure turned out to be incompatible with some
9541 broken SSL implementations, the new option is part of SSL_OP_ALL.
9542 SSL_OP_ALL is usually employed when compatibility with weird SSL
9543 implementations is desired (e.g. '-bugs' option to 's_client' and
9544 's_server'), so the new option is automatically set in many
9548 *) Changes in security patch:
9550 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9551 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9552 Air Force Materiel Command, USAF, under agreement number
9555 *) Add various sanity checks to asn1_get_length() to reject
9556 the ASN1 length bytes if they exceed sizeof(long), will appear
9557 negative or the content length exceeds the length of the
9558 supplied buffer. (CVE-2002-0659)
9559 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9561 *) Assertions for various potential buffer overflows, not known to
9563 [Ben Laurie (CHATS)]
9565 *) Various temporary buffers to hold ASCII versions of integers were
9566 too small for 64 bit platforms. (CVE-2002-0655)
9567 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9569 *) Remote buffer overflow in SSL3 protocol - an attacker could
9570 supply an oversized session ID to a client. (CVE-2002-0656)
9571 [Ben Laurie (CHATS)]
9573 *) Remote buffer overflow in SSL2 protocol - an attacker could
9574 supply an oversized client master key. (CVE-2002-0656)
9575 [Ben Laurie (CHATS)]
9577 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9579 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9580 encoded as NULL) with id-dsa-with-sha1.
9581 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9583 *) Check various X509_...() return values in apps/req.c.
9584 [Nils Larsch <nla@trustcenter.de>]
9586 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9587 an end-of-file condition would erroneously be flagged, when the CRLF
9588 was just at the end of a processed block. The bug was discovered when
9589 processing data through a buffering memory BIO handing the data to a
9590 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9591 <ptsekov@syntrex.com> and Nedelcho Stanev.
9594 *) Implement a countermeasure against a vulnerability recently found
9595 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9596 before application data chunks to avoid the use of known IVs
9597 with data potentially chosen by the attacker.
9600 *) Fix length checks in ssl3_get_client_hello().
9603 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9604 to prevent ssl3_read_internal() from incorrectly assuming that
9605 ssl3_read_bytes() found application data while handshake
9606 processing was enabled when in fact s->s3->in_read_app_data was
9607 merely automatically cleared during the initial handshake.
9608 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9610 *) Fix object definitions for Private and Enterprise: they were not
9611 recognized in their shortname (=lowercase) representation. Extend
9612 obj_dat.pl to issue an error when using undefined keywords instead
9613 of silently ignoring the problem (Svenning Sorensen
9614 <sss@sss.dnsalias.net>).
9617 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9618 generators, i.e. generators other than 2 and 5. (Previously, the
9619 code did not properly initialise the 'add' and 'rem' values to
9620 BN_generate_prime().)
9622 In the new general case, we do not insist that 'generator' is
9623 actually a primitive root: This requirement is rather pointless;
9624 a generator of the order-q subgroup is just as good, if not
9628 *) Map new X509 verification errors to alerts. Discovered and submitted by
9629 Tom Wu <tom@arcot.com>.
9632 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9633 returning non-zero before the data has been completely received
9634 when using non-blocking I/O.
9635 [Bodo Moeller; problem pointed out by John Hughes]
9637 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9638 [Ben Laurie, Lutz Jaenicke]
9640 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9641 Yoram Zahavi <YoramZ@gilian.com>).
9644 *) Add information about CygWin 1.3 and on, and preserve proper
9645 configuration for the versions before that.
9646 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9648 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9649 check whether we deal with a copy of a session and do not delete from
9650 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9651 <izhar@checkpoint.com>.
9654 *) Do not store session data into the internal session cache, if it
9655 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9656 flag is set). Proposed by Aslam <aslam@funk.com>.
9659 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9663 *) [In 0.9.6d-engine release:]
9664 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9665 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9667 *) Add the configuration target linux-s390x.
9668 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9670 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9671 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9672 variable as an indication that a ClientHello message has been
9673 received. As the flag value will be lost between multiple
9674 invocations of ssl3_accept when using non-blocking I/O, the
9675 function may not be aware that a handshake has actually taken
9676 place, thus preventing a new session from being added to the
9679 To avoid this problem, we now set s->new_session to 2 instead of
9680 using a local variable.
9681 [Lutz Jaenicke, Bodo Moeller]
9683 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9684 if the SSL_R_LENGTH_MISMATCH error is detected.
9685 [Geoff Thorpe, Bodo Moeller]
9687 *) New 'shared_ldflag' column in Configure platform table.
9690 *) Fix EVP_CIPHER_mode macro.
9691 ["Dan S. Camper" <dan@bti.net>]
9693 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9694 type, we must throw them away by setting rr->length to 0.
9695 [D P Chang <dpc@qualys.com>]
9697 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9699 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9700 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9701 worked incorrectly for those cases where range = 10..._2 and
9702 3*range is two bits longer than range.)
9705 *) Only add signing time to PKCS7 structures if it is not already
9709 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9710 OBJ_ld_ce should be OBJ_id_ce.
9711 Also some ip-pda OIDs in crypto/objects/objects.txt were
9712 incorrect (cf. RFC 3039).
9713 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9715 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9716 returns early because it has nothing to do.
9717 [Andy Schneider <andy.schneider@bjss.co.uk>]
9719 *) [In 0.9.6c-engine release:]
9720 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9721 [Andy Schneider <andy.schneider@bjss.co.uk>]
9723 *) [In 0.9.6c-engine release:]
9724 Add support for Cryptographic Appliance's keyserver technology.
9725 (Use engine 'keyclient')
9726 [Cryptographic Appliances and Geoff Thorpe]
9728 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9729 is called via tools/c89.sh because arguments have to be
9730 rearranged (all '-L' options must appear before the first object
9732 [Richard Shapiro <rshapiro@abinitio.com>]
9734 *) [In 0.9.6c-engine release:]
9735 Add support for Broadcom crypto accelerator cards, backported
9737 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9739 *) [In 0.9.6c-engine release:]
9740 Add support for SureWare crypto accelerator cards from
9741 Baltimore Technologies. (Use engine 'sureware')
9742 [Baltimore Technologies and Mark Cox]
9744 *) [In 0.9.6c-engine release:]
9745 Add support for crypto accelerator cards from Accelerated
9746 Encryption Processing, www.aep.ie. (Use engine 'aep')
9747 [AEP Inc. and Mark Cox]
9749 *) Add a configuration entry for gcc on UnixWare.
9750 [Gary Benson <gbenson@redhat.com>]
9752 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9753 messages are stored in a single piece (fixed-length part and
9754 variable-length part combined) and fix various bugs found on the way.
9757 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9758 instead. BIO_gethostbyname() does not know what timeouts are
9759 appropriate, so entries would stay in cache even when they have
9761 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9763 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9764 faced with a pathologically small ClientHello fragment that does
9765 not contain client_version: Instead of aborting with an error,
9766 simply choose the highest available protocol version (i.e.,
9767 TLS 1.0 unless it is disabled). In practice, ClientHello
9768 messages are never sent like this, but this change gives us
9769 strictly correct behaviour at least for TLS.
9772 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9773 never resets s->method to s->ctx->method when called from within
9774 one of the SSL handshake functions.
9775 [Bodo Moeller; problem pointed out by Niko Baric]
9777 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9778 (sent using the client's version number) if client_version is
9779 smaller than the protocol version in use. Also change
9780 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9781 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9782 the client will at least see that alert.
9785 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9789 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9790 client receives HelloRequest while in a handshake.
9791 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9793 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9794 should end in 'break', not 'goto end' which circumvents various
9795 cleanups done in state SSL_ST_OK. But session related stuff
9796 must be disabled for SSL_ST_OK in the case that we just sent a
9799 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9800 before just sending a HelloRequest.
9801 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9803 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9804 reveal whether illegal block cipher padding was found or a MAC
9805 verification error occurred. (Neither SSLerr() codes nor alerts
9806 are directly visible to potential attackers, but the information
9807 may leak via logfiles.)
9809 Similar changes are not required for the SSL 2.0 implementation
9810 because the number of padding bytes is sent in clear for SSL 2.0,
9811 and the extra bytes are just ignored. However ssl/s2_pkt.c
9812 failed to verify that the purported number of padding bytes is in
9816 *) Add OpenUNIX-8 support including shared libraries
9817 (Boyd Lynn Gerber <gerberb@zenez.com>).
9820 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9821 'wristwatch attack' using huge encoding parameters (cf.
9822 James H. Manger's CRYPTO 2001 paper). Note that the
9823 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9824 encoding parameters and hence was not vulnerable.
9827 *) BN_sqr() bug fix.
9828 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9830 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9831 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9832 followed by modular reduction.
9833 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9835 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9836 equivalent based on BN_pseudo_rand() instead of BN_rand().
9839 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9840 This function was broken, as the check for a new client hello message
9841 to handle SGC did not allow these large messages.
9842 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9845 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9848 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9849 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9852 *) Rework the configuration and shared library support for Tru64 Unix.
9853 The configuration part makes use of modern compiler features and
9854 still retains old compiler behavior for those that run older versions
9855 of the OS. The shared library support part includes a variant that
9856 uses the RPATH feature, and is available through the special
9857 configuration target "alpha-cc-rpath", which will never be selected
9859 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9861 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9862 with the same message size as in ssl3_get_certificate_request().
9863 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9864 messages might inadvertently be reject as too long.
9865 [Petr Lampa <lampa@fee.vutbr.cz>]
9867 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9870 *) Modified SSL library such that the verify_callback that has been set
9871 specifically for an SSL object with SSL_set_verify() is actually being
9872 used. Before the change, a verify_callback set with this function was
9873 ignored and the verify_callback() set in the SSL_CTX at the time of
9874 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9875 to allow the necessary settings.
9878 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9879 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9880 done automatically (in contradiction to the requirements of the C
9881 standard). This made problems when used from OpenSSH.
9884 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9885 dh->length and always used
9887 BN_rand_range(priv_key, dh->p).
9889 BN_rand_range() is not necessary for Diffie-Hellman, and this
9890 specific range makes Diffie-Hellman unnecessarily inefficient if
9891 dh->length (recommended exponent length) is much smaller than the
9892 length of dh->p. We could use BN_rand_range() if the order of
9893 the subgroup was stored in the DH structure, but we only have
9898 BN_rand(priv_key, l, ...)
9900 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9906 RSA_eay_public_encrypt
9907 RSA_eay_private_decrypt
9908 RSA_eay_private_encrypt (signing)
9909 RSA_eay_public_decrypt (signature verification)
9911 (default implementations for RSA_public_encrypt,
9912 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9913 always reject numbers >= n.
9916 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9917 to synchronize access to 'locking_thread'. This is necessary on
9918 systems where access to 'locking_thread' (an 'unsigned long'
9919 variable) is not atomic.
9922 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9923 *before* setting the 'crypto_lock_rand' flag. The previous code had
9924 a race condition if 0 is a valid thread ID.
9925 [Travis Vitek <vitek@roguewave.com>]
9927 *) Add support for shared libraries under Irix.
9928 [Albert Chin-A-Young <china@thewrittenword.com>]
9930 *) Add configuration option to build on Linux on both big-endian and
9932 [Ralf Baechle <ralf@uni-koblenz.de>]
9934 *) Add the possibility to create shared libraries on HP-UX.
9937 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9939 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9940 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9941 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9942 PRNG state recovery was possible based on the output of
9943 one PRNG request appropriately sized to gain knowledge on
9944 'md' followed by enough consecutive 1-byte PRNG requests
9945 to traverse all of 'state'.
9947 1. When updating 'md_local' (the current thread's copy of 'md')
9948 during PRNG output generation, hash all of the previous
9949 'md_local' value, not just the half used for PRNG output.
9951 2. Make the number of bytes from 'state' included into the hash
9952 independent from the number of PRNG bytes requested.
9954 The first measure alone would be sufficient to avoid
9955 Markku-Juhani's attack. (Actually it had never occurred
9956 to me that the half of 'md_local' used for chaining was the
9957 half from which PRNG output bytes were taken -- I had always
9958 assumed that the secret half would be used.) The second
9959 measure makes sure that additional data from 'state' is never
9960 mixed into 'md_local' in small portions; this heuristically
9961 further strengthens the PRNG.
9964 *) Fix crypto/bn/asm/mips3.s.
9967 *) When only the key is given to "enc", the IV is undefined. Print out
9968 an error message in this case.
9971 *) Handle special case when X509_NAME is empty in X509 printing routines.
9974 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9975 positive and less than q.
9978 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9979 used: it isn't thread safe and the add_lock_callback should handle
9981 [Paul Rose <Paul.Rose@bridge.com>]
9983 *) Verify that incoming data obeys the block size in
9984 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9988 [Ulf Möller, Bodo Möller]
9990 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9991 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9992 when fixing the server behaviour for backwards-compatible 'client
9993 hello' messages. (Note that the attack is impractical against
9994 SSL 3.0 and TLS 1.0 anyway because length and version checking
9995 means that the probability of guessing a valid ciphertext is
9996 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9999 Before 0.9.5, the countermeasure (hide the error by generating a
10000 random 'decryption result') did not work properly because
10001 ERR_clear_error() was missing, meaning that SSL_get_error() would
10002 detect the supposedly ignored error.
10004 Both problems are now fixed.
10007 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
10008 (previously it was 1024).
10011 *) Fix for compatibility mode trust settings: ignore trust settings
10012 unless some valid trust or reject settings are present.
10015 *) Fix for blowfish EVP: its a variable length cipher.
10018 *) Fix various bugs related to DSA S/MIME verification. Handle missing
10019 parameters in DSA public key structures and return an error in the
10020 DSA routines if parameters are absent.
10023 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
10024 in the current directory if neither $RANDFILE nor $HOME was set.
10025 RAND_file_name() in 0.9.6a returned NULL in this case. This has
10026 caused some confusion to Windows users who haven't defined $HOME.
10027 Thus RAND_file_name() is changed again: e_os.h can define a
10028 DEFAULT_HOME, which will be used if $HOME is not set.
10029 For Windows, we use "C:"; on other platforms, we still require
10030 environment variables.
10032 *) Move 'if (!initialized) RAND_poll()' into regions protected by
10033 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
10034 having multiple threads call RAND_poll() concurrently.
10037 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
10038 combination of a flag and a thread ID variable.
10039 Otherwise while one thread is in ssleay_rand_bytes (which sets the
10040 flag), *other* threads can enter ssleay_add_bytes without obeying
10041 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
10042 that they do not hold after the first thread unsets add_do_not_lock).
10045 *) Change bctest again: '-x' expressions are not available in all
10046 versions of 'test'.
10049 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
10051 *) Fix a couple of memory leaks in PKCS7_dataDecode()
10052 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
10054 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
10055 the default extension for executables, if any. Also, make the perl
10056 scripts that use symlink() to test if it really exists and use "cp"
10057 if it doesn't. All this made OpenSSL compilable and installable in
10061 *) Fix for asn1_GetSequence() for indefinite length constructed data.
10062 If SEQUENCE is length is indefinite just set c->slen to the total
10063 amount of data available.
10064 [Steve Henson, reported by shige@FreeBSD.org]
10065 [This change does not apply to 0.9.7.]
10067 *) Change bctest to avoid here-documents inside command substitution
10068 (workaround for FreeBSD /bin/sh bug).
10069 For compatibility with Ultrix, avoid shell functions (introduced
10070 in the bctest version that searches along $PATH).
10073 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
10074 with des_encrypt() defined on some operating systems, like Solaris
10078 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
10079 On the Importance of Eliminating Errors in Cryptographic
10080 Computations, J. Cryptology 14 (2001) 2, 101-119,
10081 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
10084 *) MIPS assembler BIGNUM division bug fix.
10087 *) Disabled incorrect Alpha assembler code.
10090 *) Fix PKCS#7 decode routines so they correctly update the length
10091 after reading an EOC for the EXPLICIT tag.
10093 [This change does not apply to 0.9.7.]
10095 *) Fix bug in PKCS#12 key generation routines. This was triggered
10096 if a 3DES key was generated with a 0 initial byte. Include
10097 PKCS12_BROKEN_KEYGEN compilation option to retain the old
10098 (but broken) behaviour.
10101 *) Enhance bctest to search for a working bc along $PATH and print
10103 [Tim Rice <tim@multitalents.net> via Richard Levitte]
10105 *) Fix memory leaks in err.c: free err_data string if necessary;
10106 don't write to the wrong index in ERR_set_error_data.
10109 *) Implement ssl23_peek (analogous to ssl23_read), which previously
10113 *) Replace rdtsc with _emit statements for VC++ version 5.
10114 [Jeremy Cooper <jeremy@baymoo.org>]
10116 *) Make it possible to reuse SSLv2 sessions.
10119 *) In copy_email() check for >= 0 as a return value for
10120 X509_NAME_get_index_by_NID() since 0 is a valid index.
10121 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
10123 *) Avoid coredump with unsupported or invalid public keys by checking if
10124 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
10125 PKCS7_verify() fails with non detached data.
10128 *) Don't use getenv in library functions when run as setuid/setgid.
10129 New function OPENSSL_issetugid().
10132 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
10133 due to incorrect handling of multi-threading:
10135 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
10137 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
10139 3. Count how many times MemCheck_off() has been called so that
10140 nested use can be treated correctly. This also avoids
10141 inband-signalling in the previous code (which relied on the
10142 assumption that thread ID 0 is impossible).
10145 *) Add "-rand" option also to s_client and s_server.
10148 *) Fix CPU detection on Irix 6.x.
10149 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
10150 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10152 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
10155 [This change does not apply to 0.9.7.]
10157 *) Use the cached encoding of an X509_NAME structure rather than
10158 copying it. This is apparently the reason for the libsafe "errors"
10159 but the code is actually correct.
10162 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
10163 Bleichenbacher's DSA attack.
10164 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
10165 to be set and top=0 forces the highest bit to be set; top=-1 is new
10166 and leaves the highest bit random.
10167 [Ulf Moeller, Bodo Moeller]
10169 *) In the NCONF_...-based implementations for CONF_... queries
10170 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
10171 a temporary CONF structure with the data component set to NULL
10172 (which gives segmentation faults in lh_retrieve).
10173 Instead, use NULL for the CONF pointer in CONF_get_string and
10174 CONF_get_number (which may use environment variables) and directly
10175 return NULL from CONF_get_section.
10178 *) Fix potential buffer overrun for EBCDIC.
10181 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
10182 keyUsage if basicConstraints absent for a CA.
10185 *) Make SMIME_write_PKCS7() write mail header values with a format that
10186 is more generally accepted (no spaces before the semicolon), since
10187 some programs can't parse those values properly otherwise. Also make
10188 sure BIO's that break lines after each write do not create invalid
10192 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
10193 macros previously used would not encode an empty SEQUENCE OF
10194 and break the signature.
10196 [This change does not apply to 0.9.7.]
10198 *) Zero the premaster secret after deriving the master secret in
10202 *) Add some EVP_add_digest_alias registrations (as found in
10203 OpenSSL_add_all_digests()) to SSL_library_init()
10204 aka OpenSSL_add_ssl_algorithms(). This provides improved
10205 compatibility with peers using X.509 certificates
10206 with unconventional AlgorithmIdentifier OIDs.
10209 *) Fix for Irix with NO_ASM.
10210 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
10212 *) ./config script fixes.
10213 [Ulf Moeller, Richard Levitte]
10215 *) Fix 'openssl passwd -1'.
10218 *) Change PKCS12_key_gen_asc() so it can cope with non null
10219 terminated strings whose length is passed in the passlen
10220 parameter, for example from PEM callbacks. This was done
10221 by adding an extra length parameter to asc2uni().
10222 [Steve Henson, reported by <oddissey@samsung.co.kr>]
10224 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
10225 call failed, free the DSA structure.
10228 *) Fix to uni2asc() to cope with zero length Unicode strings.
10229 These are present in some PKCS#12 files.
10232 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
10233 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
10234 when writing a 32767 byte record.
10235 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
10237 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
10238 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
10240 (RSA objects have a reference count access to which is protected
10241 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
10242 so they are meant to be shared between threads.)
10243 [Bodo Moeller, Geoff Thorpe; original patch submitted by
10244 "Reddie, Steven" <Steven.Reddie@ca.com>]
10246 *) Fix a deadlock in CRYPTO_mem_leaks().
10249 *) Use better test patterns in bntest.
10252 *) rand_win.c fix for Borland C.
10255 *) BN_rshift bugfix for n == 0.
10258 *) Add a 'bctest' script that checks for some known 'bc' bugs
10259 so that 'make test' does not abort just because 'bc' is broken.
10262 *) Store verify_result within SSL_SESSION also for client side to
10263 avoid potential security hole. (Re-used sessions on the client side
10264 always resulted in verify_result==X509_V_OK, not using the original
10265 result of the server certificate verification.)
10268 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
10269 SSL3_RT_APPLICATION_DATA, return 0.
10270 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
10274 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
10275 releases, have been re-implemented by renaming the previous
10276 implementations of ssl2_read and ssl3_read to ssl2_read_internal
10277 and ssl3_read_internal, respectively, and adding 'peek' parameters
10278 to them. The new ssl[23]_{read,peek} functions are calls to
10279 ssl[23]_read_internal with the 'peek' flag set appropriately.
10280 A 'peek' parameter has also been added to ssl3_read_bytes, which
10281 does the actual work for ssl3_read_internal.
10284 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
10285 the method-specific "init()" handler. Also clean up ex_data after
10286 calling the method-specific "finish()" handler. Previously, this was
10287 happening the other way round.
10290 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
10291 The previous value, 12, was not always sufficient for BN_mod_exp().
10294 *) Make sure that shared libraries get the internal name engine with
10295 the full version number and not just 0. This should mark the
10296 shared libraries as not backward compatible. Of course, this should
10297 be changed again when we can guarantee backward binary compatibility.
10300 *) Fix typo in get_cert_by_subject() in by_dir.c
10301 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
10303 *) Rework the system to generate shared libraries:
10305 - Make note of the expected extension for the shared libraries and
10306 if there is a need for symbolic links from for example libcrypto.so.0
10307 to libcrypto.so.0.9.7. There is extended info in Configure for
10310 - Make as few rebuilds of the shared libraries as possible.
10312 - Still avoid linking the OpenSSL programs with the shared libraries.
10314 - When installing, install the shared libraries separately from the
10318 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10320 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10321 and not in SSL_clear because the latter is also used by the
10322 accept/connect functions; previously, the settings made by
10323 SSL_set_read_ahead would be lost during the handshake.
10324 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10326 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10327 Previously, it would create entries for disabled algorithms no
10331 *) Added several new manual pages for SSL_* function.
10334 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10336 *) In ssl23_get_client_hello, generate an error message when faced
10337 with an initial SSL 3.0/TLS record that is too small to contain the
10338 first two bytes of the ClientHello message, i.e. client_version.
10339 (Note that this is a pathologic case that probably has never happened
10340 in real life.) The previous approach was to use the version number
10341 from the record header as a substitute; but our protocol choice
10342 should not depend on that one because it is not authenticated
10343 by the Finished messages.
10346 *) More robust randomness gathering functions for Windows.
10347 [Jeffrey Altman <jaltman@columbia.edu>]
10349 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10350 not set then we don't setup the error code for issuer check errors
10351 to avoid possibly overwriting other errors which the callback does
10352 handle. If an application does set the flag then we assume it knows
10353 what it is doing and can handle the new informational codes
10357 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10358 a general "ANY" type, as such it should be able to decode anything
10359 including tagged types. However it didn't check the class so it would
10360 wrongly interpret tagged types in the same way as their universal
10361 counterpart and unknown types were just rejected. Changed so that the
10362 tagged and unknown types are handled in the same way as a SEQUENCE:
10363 that is the encoding is stored intact. There is also a new type
10364 "V_ASN1_OTHER" which is used when the class is not universal, in this
10365 case we have no idea what the actual type is so we just lump them all
10369 *) On VMS, stdout may very well lead to a file that is written to
10370 in a record-oriented fashion. That means that every write() will
10371 write a separate record, which will be read separately by the
10372 programs trying to read from it. This can be very confusing.
10374 The solution is to put a BIO filter in the way that will buffer
10375 text until a linefeed is reached, and then write everything a
10376 line at a time, so every record written will be an actual line,
10377 not chunks of lines and not (usually doesn't happen, but I've
10378 seen it once) several lines in one record. BIO_f_linebuffer() is
10381 Currently, it's a VMS-only method, because that's where it has
10382 been tested well enough.
10385 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10386 it can return incorrect results.
10387 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10388 but it was in 0.9.6-beta[12].)
10391 *) Disable the check for content being present when verifying detached
10392 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10393 include zero length content when signing messages.
10396 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10397 BIO_ctrl (for BIO pairs).
10400 *) Add DSO method for VMS.
10403 *) Bug fix: Montgomery multiplication could produce results with the
10407 *) Add RPM specification openssl.spec and modify it to build three
10408 packages. The default package contains applications, application
10409 documentation and run-time libraries. The devel package contains
10410 include files, static libraries and function documentation. The
10411 doc package contains the contents of the doc directory. The original
10412 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10415 *) Add a large number of documentation files for many SSL routines.
10416 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10418 *) Add a configuration entry for Sony News 4.
10419 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10421 *) Don't set the two most significant bits to one when generating a
10422 random number < q in the DSA library.
10425 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10426 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10427 the underlying transport is blocking) if a handshake took place.
10428 (The default behaviour is needed by applications such as s_client
10429 and s_server that use select() to determine when to use SSL_read;
10430 but for applications that know in advance when to expect data, it
10431 just makes things more complicated.)
10434 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10438 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10439 work better on such systems.
10440 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10442 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10443 Update PKCS12_parse() so it copies the friendlyName and the
10444 keyid to the certificates aux info.
10447 *) Fix bug in PKCS7_verify() which caused an infinite loop
10448 if there was more than one signature.
10449 [Sven Uszpelkat <su@celocom.de>]
10451 *) Major change in util/mkdef.pl to include extra information
10452 about each symbol, as well as presenting variables as well
10453 as functions. This change means that there's n more need
10454 to rebuild the .num files when some algorithms are excluded.
10457 *) Allow the verify time to be set by an application,
10458 rather than always using the current time.
10461 *) Phase 2 verify code reorganisation. The certificate
10462 verify code now looks up an issuer certificate by a
10463 number of criteria: subject name, authority key id
10464 and key usage. It also verifies self signed certificates
10465 by the same criteria. The main comparison function is
10466 X509_check_issued() which performs these checks.
10468 Lot of changes were necessary in order to support this
10469 without completely rewriting the lookup code.
10471 Authority and subject key identifier are now cached.
10473 The LHASH 'certs' is X509_STORE has now been replaced
10474 by a STACK_OF(X509_OBJECT). This is mainly because an
10475 LHASH can't store or retrieve multiple objects with
10476 the same hash value.
10478 As a result various functions (which were all internal
10479 use only) have changed to handle the new X509_STORE
10480 structure. This will break anything that messed round
10481 with X509_STORE internally.
10483 The functions X509_STORE_add_cert() now checks for an
10484 exact match, rather than just subject name.
10486 The X509_STORE API doesn't directly support the retrieval
10487 of multiple certificates matching a given criteria, however
10488 this can be worked round by performing a lookup first
10489 (which will fill the cache with candidate certificates)
10490 and then examining the cache for matches. This is probably
10491 the best we can do without throwing out X509_LOOKUP
10492 entirely (maybe later...).
10494 The X509_VERIFY_CTX structure has been enhanced considerably.
10496 All certificate lookup operations now go via a get_issuer()
10497 callback. Although this currently uses an X509_STORE it
10498 can be replaced by custom lookups. This is a simple way
10499 to bypass the X509_STORE hackery necessary to make this
10500 work and makes it possible to use more efficient techniques
10501 in future. A very simple version which uses a simple
10502 STACK for its trusted certificate store is also provided
10503 using X509_STORE_CTX_trusted_stack().
10505 The verify_cb() and verify() callbacks now have equivalents
10506 in the X509_STORE_CTX structure.
10508 X509_STORE_CTX also has a 'flags' field which can be used
10509 to customise the verify behaviour.
10512 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10513 excludes S/MIME capabilities.
10516 *) When a certificate request is read in keep a copy of the
10517 original encoding of the signed data and use it when outputting
10518 again. Signatures then use the original encoding rather than
10519 a decoded, encoded version which may cause problems if the
10520 request is improperly encoded.
10523 *) For consistency with other BIO_puts implementations, call
10524 buffer_write(b, ...) directly in buffer_puts instead of calling
10527 In BIO_puts, increment b->num_write as in BIO_write.
10528 [Peter.Sylvester@EdelWeb.fr]
10530 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10531 BN_zero, we may not return a BIGNUM with an array consisting of
10532 words set to zero.)
10535 *) Avoid calling abort() from within the library when problems are
10536 detected, except if preprocessor symbols have been defined
10537 (such as REF_CHECK, BN_DEBUG etc.).
10540 *) New openssl application 'rsautl'. This utility can be
10541 used for low level RSA operations. DER public key
10542 BIO/fp routines also added.
10545 *) New Configure entry and patches for compiling on QNX 4.
10546 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10548 *) A demo state-machine implementation was sponsored by
10549 Nuron (http://www.nuron.com/) and is now available in
10550 demos/state_machine.
10553 *) New options added to the 'dgst' utility for signature
10554 generation and verification.
10557 *) Unrecognized PKCS#7 content types are now handled via a
10558 catch all ASN1_TYPE structure. This allows unsupported
10559 types to be stored as a "blob" and an application can
10560 encode and decode it manually.
10563 *) Fix various signed/unsigned issues to make a_strex.c
10564 compile under VC++.
10565 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10567 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10568 length if passed a buffer. ASN1_INTEGER_to_BN failed
10569 if passed a NULL BN and its argument was negative.
10570 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10572 *) Modification to PKCS#7 encoding routines to output definite
10573 length encoding. Since currently the whole structures are in
10574 memory there's not real point in using indefinite length
10575 constructed encoding. However if OpenSSL is compiled with
10576 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10579 *) Added BIO_vprintf() and BIO_vsnprintf().
10582 *) Added more prefixes to parse for in the strings written
10583 through a logging bio, to cover all the levels that are available
10584 through syslog. The prefixes are now:
10586 PANIC, EMERG, EMR => LOG_EMERG
10587 ALERT, ALR => LOG_ALERT
10588 CRIT, CRI => LOG_CRIT
10589 ERROR, ERR => LOG_ERR
10590 WARNING, WARN, WAR => LOG_WARNING
10591 NOTICE, NOTE, NOT => LOG_NOTICE
10592 INFO, INF => LOG_INFO
10593 DEBUG, DBG => LOG_DEBUG
10595 and as before, if none of those prefixes are present at the
10596 beginning of the string, LOG_ERR is chosen.
10598 On Win32, the LOG_* levels are mapped according to this:
10600 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10601 LOG_WARNING => EVENTLOG_WARNING_TYPE
10602 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10606 *) Made it possible to reconfigure with just the configuration
10607 argument "reconf" or "reconfigure". The command line arguments
10608 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10609 and are retrieved from there when reconfiguring.
10612 *) MD4 implemented.
10613 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10615 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10618 *) The obj_dat.pl script was messing up the sorting of object
10619 names. The reason was that it compared the quoted version
10620 of strings as a result "OCSP" > "OCSP Signing" because
10621 " > SPACE. Changed script to store unquoted versions of
10622 names and add quotes on output. It was also omitting some
10623 names from the lookup table if they were given a default
10624 value (that is if SN is missing it is given the same
10625 value as LN and vice versa), these are now added on the
10626 grounds that if an object has a name we should be able to
10627 look it up. Finally added warning output when duplicate
10628 short or long names are found.
10631 *) Changes needed for Tandem NSK.
10632 [Scott Uroff <scott@xypro.com>]
10634 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10635 RSA_padding_check_SSLv23(), special padding was never detected
10636 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10637 version rollback attacks was not effective.
10639 In s23_clnt.c, don't use special rollback-attack detection padding
10640 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10641 client; similarly, in s23_srvr.c, don't do the rollback check if
10642 SSL 2.0 is the only protocol enabled in the server.
10645 *) Make it possible to get hexdumps of unprintable data with 'openssl
10646 asn1parse'. By implication, the functions ASN1_parse_dump() and
10647 BIO_dump_indent() are added.
10650 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10651 these print out strings and name structures based on various
10652 flags including RFC2253 support and proper handling of
10653 multibyte characters. Added options to the 'x509' utility
10654 to allow the various flags to be set.
10657 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10658 Also change the functions X509_cmp_current_time() and
10659 X509_gmtime_adj() work with an ASN1_TIME structure,
10660 this will enable certificates using GeneralizedTime in validity
10661 dates to be checked.
10664 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10665 negative public key encodings) on by default,
10666 NO_NEG_PUBKEY_BUG can be set to disable it.
10669 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10670 content octets. An i2c_ASN1_OBJECT is unnecessary because
10671 the encoding can be trivially obtained from the structure.
10674 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10675 not read locks (CRYPTO_r_[un]lock).
10678 *) A first attempt at creating official support for shared
10679 libraries through configuration. I've kept it so the
10680 default is static libraries only, and the OpenSSL programs
10681 are always statically linked for now, but there are
10682 preparations for dynamic linking in place.
10683 This has been tested on Linux and Tru64.
10686 *) Randomness polling function for Win9x, as described in:
10687 Peter Gutmann, Software Generation of Practically Strong
10691 *) Fix so PRNG is seeded in req if using an already existing
10695 *) New options to smime application. -inform and -outform
10696 allow alternative formats for the S/MIME message including
10697 PEM and DER. The -content option allows the content to be
10698 specified separately. This should allow things like Netscape
10699 form signing output easier to verify.
10702 *) Fix the ASN1 encoding of tags using the 'long form'.
10705 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10706 STRING types. These convert content octets to and from the
10707 underlying type. The actual tag and length octets are
10708 already assumed to have been read in and checked. These
10709 are needed because all other string types have virtually
10710 identical handling apart from the tag. By having versions
10711 of the ASN1 functions that just operate on content octets
10712 IMPLICIT tagging can be handled properly. It also allows
10713 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10714 and ASN1_INTEGER are identical apart from the tag.
10717 *) Change the handling of OID objects as follows:
10719 - New object identifiers are inserted in objects.txt, following
10720 the syntax given in objects.README.
10721 - objects.pl is used to process obj_mac.num and create a new
10723 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10726 This is currently kind of a hack, and the perl code in objects.pl
10727 isn't very elegant, but it works as I intended. The simplest way
10728 to check that it worked correctly is to look in obj_dat.h and
10729 check the array nid_objs and make sure the objects haven't moved
10730 around (this is important!). Additions are OK, as well as
10731 consistent name changes.
10734 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10737 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10738 The given file adds to whatever has already been seeded into the
10739 random pool through the RANDFILE configuration file option or
10740 environment variable, or the default random state file.
10743 *) mkstack.pl now sorts each macro group into lexical order.
10744 Previously the output order depended on the order the files
10745 appeared in the directory, resulting in needless rewriting
10749 *) Patches to make OpenSSL compile under Win32 again. Mostly
10750 work arounds for the VC++ problem that it treats func() as
10751 func(void). Also stripped out the parts of mkdef.pl that
10752 added extra typesafe functions: these no longer exist.
10755 *) Reorganisation of the stack code. The macros are now all
10756 collected in safestack.h . Each macro is defined in terms of
10757 a "stack macro" of the form SKM_<name>(type, a, b). The
10758 DEBUG_SAFESTACK is now handled in terms of function casts,
10759 this has the advantage of retaining type safety without the
10760 use of additional functions. If DEBUG_SAFESTACK is not defined
10761 then the non typesafe macros are used instead. Also modified the
10762 mkstack.pl script to handle the new form. Needs testing to see
10763 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10764 the default if no major problems. Similar behaviour for ASN1_SET_OF
10765 and PKCS12_STACK_OF.
10768 *) When some versions of IIS use the 'NET' form of private key the
10769 key derivation algorithm is different. Normally MD5(password) is
10770 used as a 128 bit RC4 key. In the modified case
10771 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10772 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10773 as the old Netscape_RSA functions except they have an additional
10774 'sgckey' parameter which uses the modified algorithm. Also added
10775 an -sgckey command line option to the rsa utility. Thanks to
10776 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10777 algorithm to openssl-dev.
10780 *) The evp_local.h macros were using 'c.##kname' which resulted in
10781 invalid expansion on some systems (SCO 5.0.5 for example).
10782 Corrected to 'c.kname'.
10783 [Phillip Porch <root@theporch.com>]
10785 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10786 a STACK of email addresses from a certificate or request, these look
10787 in the subject name and the subject alternative name extensions and
10788 omit any duplicate addresses.
10791 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10792 This makes DSA verification about 2 % faster.
10795 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10796 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10797 plus overhead for 1024 bit moduli).
10798 This makes exponentiations about 0.5 % faster for 1024 bit
10799 exponents (as measured by "openssl speed rsa2048").
10802 *) Rename memory handling macros to avoid conflicts with other
10804 Malloc => OPENSSL_malloc
10805 Malloc_locked => OPENSSL_malloc_locked
10806 Realloc => OPENSSL_realloc
10807 Free => OPENSSL_free
10810 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10811 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10814 *) CygWin32 support.
10815 [John Jarvie <jjarvie@newsguy.com>]
10817 *) The type-safe stack code has been rejigged. It is now only compiled
10818 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10819 by default all type-specific stack functions are "#define"d back to
10820 standard stack functions. This results in more streamlined output
10821 but retains the type-safety checking possibilities of the original
10825 *) The STACK code has been cleaned up, and certain type declarations
10826 that didn't make a lot of sense have been brought in line. This has
10827 also involved a cleanup of sorts in safestack.h to more correctly
10828 map type-safe stack functions onto their plain stack counterparts.
10829 This work has also resulted in a variety of "const"ifications of
10830 lots of the code, especially "_cmp" operations which should normally
10831 be prototyped with "const" parameters anyway.
10834 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10835 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10836 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10837 where all of 'md' is used each time the PRNG is used, but 'state'
10838 is used only indexed by a cyclic counter. As entropy may not be
10839 well distributed from the beginning, 'md' is important as a
10840 chaining variable. However, the output function chains only half
10841 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10842 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10843 in all of 'state' being rewritten, with the new values depending
10844 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10847 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10848 the handshake is continued after ssl_verify_cert_chain();
10849 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10850 can lead to 'unexplainable' connection aborts later.
10851 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10853 *) Major EVP API cipher revision.
10854 Add hooks for extra EVP features. This allows various cipher
10855 parameters to be set in the EVP interface. Support added for variable
10856 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10857 setting of RC2 and RC5 parameters.
10859 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10862 Remove lots of duplicated code from the EVP library. For example *every*
10863 cipher init() function handles the 'iv' in the same way according to the
10864 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10865 for CFB and OFB modes they zero ctx->num.
10867 New functionality allows removal of S/MIME code RC2 hack.
10869 Most of the routines have the same form and so can be declared in terms
10872 By shifting this to the top level EVP_CipherInit() it can be removed from
10873 all individual ciphers. If the cipher wants to handle IVs or keys
10874 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10877 Change lots of functions like EVP_EncryptUpdate() to now return a
10878 value: although software versions of the algorithms cannot fail
10879 any installed hardware versions can.
10882 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10883 this option is set, tolerate broken clients that send the negotiated
10884 protocol version number instead of the requested protocol version
10888 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10889 i.e. non-zero for export ciphersuites, zero otherwise.
10890 Previous versions had this flag inverted, inconsistent with
10891 rsa_tmp_cb (..._TMP_RSA_CB).
10892 [Bodo Moeller; problem reported by Amit Chopra]
10894 *) Add missing DSA library text string. Work around for some IIS
10895 key files with invalid SEQUENCE encoding.
10898 *) Add a document (doc/standards.txt) that list all kinds of standards
10899 and so on that are implemented in OpenSSL.
10902 *) Enhance c_rehash script. Old version would mishandle certificates
10903 with the same subject name hash and wouldn't handle CRLs at all.
10904 Added -fingerprint option to crl utility, to support new c_rehash
10908 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10911 *) Fix for SSL server purpose checking. Server checking was
10912 rejecting certificates which had extended key usage present
10913 but no ssl client purpose.
10914 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10916 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10917 is a little unclear about how a blank password is handled.
10918 Since the password in encoded as a BMPString with terminating
10919 double NULL a zero length password would end up as just the
10920 double NULL. However no password at all is different and is
10921 handled differently in the PKCS#12 key generation code. NS
10922 treats a blank password as zero length. MSIE treats it as no
10923 password on export: but it will try both on import. We now do
10924 the same: PKCS12_parse() tries zero length and no password if
10925 the password is set to "" or NULL (NULL is now a valid password:
10926 it wasn't before) as does the pkcs12 application.
10929 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10930 perror when PEM_read_bio_X509_REQ fails, the error message must
10931 be obtained from the error queue.
10934 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10935 it in ERR_remove_state if appropriate, and change ERR_get_state
10936 accordingly to avoid race conditions (this is necessary because
10937 thread_hash is no longer constant once set).
10940 *) Bugfix for linux-elf makefile.one.
10943 *) RSA_get_default_method() will now cause a default
10944 RSA_METHOD to be chosen if one doesn't exist already.
10945 Previously this was only set during a call to RSA_new()
10946 or RSA_new_method(NULL) meaning it was possible for
10947 RSA_get_default_method() to return NULL.
10950 *) Added native name translation to the existing DSO code
10951 that will convert (if the flag to do so is set) filenames
10952 that are sufficiently small and have no path information
10953 into a canonical native form. Eg. "blah" converted to
10954 "libblah.so" or "blah.dll" etc.
10957 *) New function ERR_error_string_n(e, buf, len) which is like
10958 ERR_error_string(e, buf), but writes at most 'len' bytes
10959 including the 0 terminator. For ERR_error_string_n, 'buf'
10961 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10963 *) CONF library reworked to become more general. A new CONF
10964 configuration file reader "class" is implemented as well as a
10965 new functions (NCONF_*, for "New CONF") to handle it. The now
10966 old CONF_* functions are still there, but are reimplemented to
10967 work in terms of the new functions. Also, a set of functions
10968 to handle the internal storage of the configuration data is
10969 provided to make it easier to write new configuration file
10970 reader "classes" (I can definitely see something reading a
10971 configuration file in XML format, for example), called _CONF_*,
10972 or "the configuration storage API"...
10974 The new configuration file reading functions are:
10976 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10977 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10979 NCONF_default, NCONF_WIN32
10981 NCONF_dump_fp, NCONF_dump_bio
10983 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10984 NCONF_new creates a new CONF object. This works in the same way
10985 as other interfaces in OpenSSL, like the BIO interface.
10986 NCONF_dump_* dump the internal storage of the configuration file,
10987 which is useful for debugging. All other functions take the same
10988 arguments as the old CONF_* functions with the exception of the
10989 first that must be a `CONF *' instead of a `LHASH *'.
10991 To make it easier to use the new classes with the old CONF_* functions,
10992 the function CONF_set_default_method is provided.
10995 *) Add '-tls1' option to 'openssl ciphers', which was already
10996 mentioned in the documentation but had not been implemented.
10997 (This option is not yet really useful because even the additional
10998 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
11001 *) Initial DSO code added into libcrypto for letting OpenSSL (and
11002 OpenSSL-based applications) load shared libraries and bind to
11003 them in a portable way.
11004 [Geoff Thorpe, with contributions from Richard Levitte]
11006 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
11008 *) Make sure _lrotl and _lrotr are only used with MSVC.
11010 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
11011 (the default implementation of RAND_status).
11013 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
11014 to '-clrext' (= clear extensions), as intended and documented.
11015 [Bodo Moeller; inconsistency pointed out by Michael Attili
11016 <attili@amaxo.com>]
11018 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
11019 was larger than the MD block size.
11020 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
11022 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
11023 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
11024 using the passed key: if the passed key was a private key the result
11025 of X509_print(), for example, would be to print out all the private key
11029 *) des_quad_cksum() byte order bug fix.
11030 [Ulf Möller, using the problem description in krb4-0.9.7, where
11031 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
11033 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
11035 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
11037 *) For easily testing in shell scripts whether some command
11038 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
11039 returns with exit code 0 iff no command of the given name is available.
11040 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
11041 the output goes to stdout and nothing is printed to stderr.
11042 Additional arguments are always ignored.
11044 Since for each cipher there is a command of the same name,
11045 the 'no-cipher' compilation switches can be tested this way.
11047 ('openssl no-XXX' is not able to detect pseudo-commands such
11048 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
11051 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
11054 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
11055 is set; it will be thrown away anyway because each handshake creates
11057 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
11058 to parameters -- in previous versions (since OpenSSL 0.9.3) the
11059 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
11060 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
11063 *) New s_client option -ign_eof: EOF at stdin is ignored, and
11064 'Q' and 'R' lose their special meanings (quit/renegotiate).
11065 This is part of what -quiet does; unlike -quiet, -ign_eof
11066 does not suppress any output.
11069 *) Add compatibility options to the purpose and trust code. The
11070 purpose X509_PURPOSE_ANY is "any purpose" which automatically
11071 accepts a certificate or CA, this was the previous behaviour,
11072 with all the associated security issues.
11074 X509_TRUST_COMPAT is the old trust behaviour: only and
11075 automatically trust self signed roots in certificate store. A
11076 new trust setting X509_TRUST_DEFAULT is used to specify that
11077 a purpose has no associated trust setting and it should instead
11078 use the value in the default purpose.
11081 *) Fix the PKCS#8 DSA private key code so it decodes keys again
11082 and fix a memory leak.
11085 *) In util/mkerr.pl (which implements 'make errors'), preserve
11086 reason strings from the previous version of the .c file, as
11087 the default to have only downcase letters (and digits) in
11088 automatically generated reasons codes is not always appropriate.
11091 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
11092 using strerror. Previously, ERR_reason_error_string() returned
11093 library names as reason strings for SYSerr; but SYSerr is a special
11094 case where small numbers are errno values, not library numbers.
11097 *) Add '-dsaparam' option to 'openssl dhparam' application. This
11098 converts DSA parameters into DH parameters. (When creating parameters,
11099 DSA_generate_parameters is used.)
11102 *) Include 'length' (recommended exponent length) in C code generated
11103 by 'openssl dhparam -C'.
11106 *) The second argument to set_label in perlasm was already being used
11107 so couldn't be used as a "file scope" flag. Moved to third argument
11111 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
11112 instead of RAND_bytes for encryption IVs and salts.
11115 *) Include RAND_status() into RAND_METHOD instead of implementing
11116 it only for md_rand.c Otherwise replacing the PRNG by calling
11117 RAND_set_rand_method would be impossible.
11120 *) Don't let DSA_generate_key() enter an infinite loop if the random
11121 number generation fails.
11124 *) New 'rand' application for creating pseudo-random output.
11127 *) Added configuration support for Linux/IA64
11128 [Rolf Haberrecker <rolf@suse.de>]
11130 *) Assembler module support for Mingw32.
11133 *) Shared library support for HPUX (in shlib/).
11134 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
11136 *) Shared library support for Solaris gcc.
11137 [Lutz Behnke <behnke@trustcenter.de>]
11139 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
11141 *) PKCS7_encrypt() was adding text MIME headers twice because they
11142 were added manually and by SMIME_crlf_copy().
11145 *) In bntest.c don't call BN_rand with zero bits argument.
11146 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
11148 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
11149 case was implemented. This caused BN_div_recp() to fail occasionally.
11152 *) Add an optional second argument to the set_label() in the perl
11153 assembly language builder. If this argument exists and is set
11154 to 1 it signals that the assembler should use a symbol whose
11155 scope is the entire file, not just the current function. This
11156 is needed with MASM which uses the format label:: for this scope.
11157 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
11159 *) Change the ASN1 types so they are typedefs by default. Before
11160 almost all types were #define'd to ASN1_STRING which was causing
11161 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
11165 *) Change names of new functions to the new get1/get0 naming
11166 convention: After 'get1', the caller owns a reference count
11167 and has to call ..._free; 'get0' returns a pointer to some
11168 data structure without incrementing reference counters.
11169 (Some of the existing 'get' functions increment a reference
11170 counter, some don't.)
11171 Similarly, 'set1' and 'add1' functions increase reference
11172 counters or duplicate objects.
11175 *) Allow for the possibility of temp RSA key generation failure:
11176 the code used to assume it always worked and crashed on failure.
11179 *) Fix potential buffer overrun problem in BIO_printf().
11180 [Ulf Möller, using public domain code by Patrick Powell; problem
11181 pointed out by David Sacerdote <das33@cornell.edu>]
11183 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
11184 RAND_egd() and RAND_status(). In the command line application,
11185 the EGD socket can be specified like a seed file using RANDFILE
11189 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
11190 Some CAs (e.g. Verisign) distribute certificates in this form.
11193 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
11194 list to exclude them. This means that no special compilation option
11195 is needed to use anonymous DH: it just needs to be included in the
11199 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
11200 EVP_MD_type. The old functionality is available in a new macro called
11201 EVP_MD_md(). Change code that uses it and update docs.
11204 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
11205 where the 'void *' argument is replaced by a function pointer argument.
11206 Previously 'void *' was abused to point to functions, which works on
11207 many platforms, but is not correct. As these functions are usually
11208 called by macros defined in OpenSSL header files, most source code
11209 should work without changes.
11212 *) <openssl/opensslconf.h> (which is created by Configure) now contains
11213 sections with information on -D... compiler switches used for
11214 compiling the library so that applications can see them. To enable
11215 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
11216 must be defined. E.g.,
11217 #define OPENSSL_ALGORITHM_DEFINES
11218 #include <openssl/opensslconf.h>
11219 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
11220 [Richard Levitte, Ulf and Bodo Möller]
11222 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
11226 *) Change the 'other' type in certificate aux info to a STACK_OF
11227 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
11228 the required ASN1 format: arbitrary types determined by an OID.
11231 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
11232 argument to 'req'. This is not because the function is newer or
11233 better than others it just uses the work 'NEW' in the certificate
11234 request header lines. Some software needs this.
11237 *) Reorganise password command line arguments: now passwords can be
11238 obtained from various sources. Delete the PEM_cb function and make
11239 it the default behaviour: i.e. if the callback is NULL and the
11240 usrdata argument is not NULL interpret it as a null terminated pass
11241 phrase. If usrdata and the callback are NULL then the pass phrase
11242 is prompted for as usual.
11245 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
11246 the support is automatically enabled. The resulting binaries will
11247 autodetect the card and use it if present.
11248 [Ben Laurie and Compaq Inc.]
11250 *) Work around for Netscape hang bug. This sends certificate request
11251 and server done in one record. Since this is perfectly legal in the
11252 SSL/TLS protocol it isn't a "bug" option and is on by default. See
11253 the bugs/SSLv3 entry for more info.
11256 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
11259 *) Add -rand argument to smime and pkcs12 applications and read/write
11263 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
11266 *) Add command line password options to the remaining applications.
11269 *) Bug fix for BN_div_recp() for numerators with an even number of
11273 *) More tests in bntest.c, and changed test_bn output.
11276 *) ./config recognizes MacOS X now.
11279 *) Bug fix for BN_div() when the first words of num and divisor are
11280 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
11283 *) Add support for various broken PKCS#8 formats, and command line
11284 options to produce them.
11287 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
11288 get temporary BIGNUMs from a BN_CTX.
11291 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
11295 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
11296 include a #define from the old name to the new. The original intent
11297 was that statically linked binaries could for example just call
11298 SSLeay_add_all_ciphers() to just add ciphers to the table and not
11299 link with digests. This never worked because SSLeay_add_all_digests()
11300 and SSLeay_add_all_ciphers() were in the same source file so calling
11301 one would link with the other. They are now in separate source files.
11304 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
11307 *) Use a less unusual form of the Miller-Rabin primality test (it used
11308 a binary algorithm for exponentiation integrated into the Miller-Rabin
11309 loop, our standard modexp algorithms are faster).
11312 *) Support for the EBCDIC character set completed.
11313 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
11315 *) Source code cleanups: use const where appropriate, eliminate casts,
11316 use void * instead of char * in lhash.
11319 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11320 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11321 this the server could overwrite ephemeral keys that the client
11325 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11326 using 50 iterations of the Rabin-Miller test.
11328 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11329 iterations of the Rabin-Miller test as required by the appendix
11330 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11331 As BN_is_prime_fasttest includes trial division, DSA parameter
11332 generation becomes much faster.
11334 This implies a change for the callback functions in DSA_is_prime
11335 and DSA_generate_parameters: The callback function is called once
11336 for each positive witness in the Rabin-Miller test, not just
11337 occasionally in the inner loop; and the parameters to the
11338 callback function now provide an iteration count for the outer
11339 loop rather than for the current invocation of the inner loop.
11340 DSA_generate_parameters additionally can call the callback
11341 function with an 'iteration count' of -1, meaning that a
11342 candidate has passed the trial division test (when q is generated
11343 from an application-provided seed, trial division is skipped).
11346 *) New function BN_is_prime_fasttest that optionally does trial
11347 division before starting the Rabin-Miller test and has
11348 an additional BN_CTX * argument (whereas BN_is_prime always
11349 has to allocate at least one BN_CTX).
11350 'callback(1, -1, cb_arg)' is called when a number has passed the
11351 trial division stage.
11354 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11358 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11361 *) New function BN_pseudo_rand().
11364 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11365 bignum version of BN_from_montgomery() with the working code from
11366 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11370 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11371 made it impossible to use the same SSL_SESSION data structure in
11372 SSL2 clients in multiple threads.
11375 *) The return value of RAND_load_file() no longer counts bytes obtained
11376 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11377 to seed the PRNG (previously an explicit byte count was required).
11378 [Ulf Möller, Bodo Möller]
11380 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11381 used (char *) instead of (void *) and had casts all over the place.
11384 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11387 *) Retain source code compatibility for BN_prime_checks macro:
11388 BN_is_prime(..., BN_prime_checks, ...) now uses
11389 BN_prime_checks_for_size to determine the appropriate number of
11390 Rabin-Miller iterations.
11393 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11394 DH_CHECK_P_NOT_SAFE_PRIME.
11395 (Check if this is true? OpenPGP calls them "strong".)
11398 *) Merge the functionality of "dh" and "gendh" programs into a new program
11399 "dhparam". The old programs are retained for now but will handle DH keys
11400 (instead of parameters) in future.
11403 *) Make the ciphers, s_server and s_client programs check the return values
11404 when a new cipher list is set.
11407 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11408 ciphers. Before when the 56bit ciphers were enabled the sorting was
11411 The syntax for the cipher sorting has been extended to support sorting by
11412 cipher-strength (using the strength_bits hard coded in the tables).
11413 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11415 Fix a bug in the cipher-command parser: when supplying a cipher command
11416 string with an "undefined" symbol (neither command nor alphanumeric
11417 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11418 an error is flagged.
11420 Due to the strength-sorting extension, the code of the
11421 ssl_create_cipher_list() function was completely rearranged. I hope that
11422 the readability was also increased :-)
11423 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11425 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11426 for the first serial number and places 2 in the serial number file. This
11427 avoids problems when the root CA is created with serial number zero and
11428 the first user certificate has the same issuer name and serial number
11432 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11433 the new code. Add documentation for this stuff.
11436 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11437 X509_*() to X509at_*() on the grounds that they don't handle X509
11438 structures and behave in an analogous way to the X509v3 functions:
11439 they shouldn't be called directly but wrapper functions should be used
11442 So we also now have some wrapper functions that call the X509at functions
11443 when passed certificate requests. (TO DO: similar things can be done with
11444 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11445 things. Some of these need some d2i or i2d and print functionality
11446 because they handle more complex structures.)
11449 *) Add missing #ifndefs that caused missing symbols when building libssl
11450 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11451 NO_RSA in ssl/s2*.c.
11452 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11454 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11455 has a return value which indicates the quality of the random data
11456 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11457 error queue. New function RAND_pseudo_bytes() generates output that is
11458 guaranteed to be unique but not unpredictable. RAND_add is like
11459 RAND_seed, but takes an extra argument for an entropy estimate
11460 (RAND_seed always assumes full entropy).
11463 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11464 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11465 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11466 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11467 false-positive rate of at most 2^-80 for random input.
11470 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11473 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11474 in the 0.9.5 release), this returns the chain
11475 from an X509_CTX structure with a dup of the stack and all
11476 the X509 reference counts upped: so the stack will exist
11477 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11480 Also make SSL_SESSION_print() print out the verify return
11484 *) Add manpage for the pkcs12 command. Also change the default
11485 behaviour so MAC iteration counts are used unless the new
11486 -nomaciter option is used. This improves file security and
11487 only older versions of MSIE (4.0 for example) need it.
11490 *) Honor the no-xxx Configure options when creating .DEF files.
11493 *) Add PKCS#10 attributes to field table: challengePassword,
11494 unstructuredName and unstructuredAddress. These are taken from
11495 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11496 international characters are used.
11498 More changes to X509_ATTRIBUTE code: allow the setting of types
11499 based on strings. Remove the 'loc' parameter when adding
11500 attributes because these will be a SET OF encoding which is sorted
11504 *) Initial changes to the 'req' utility to allow request generation
11505 automation. This will allow an application to just generate a template
11506 file containing all the field values and have req construct the
11509 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11510 used all over the place including certificate requests and PKCS#7
11511 structures. They are currently handled manually where necessary with
11512 some primitive wrappers for PKCS#7. The new functions behave in a
11513 manner analogous to the X509 extension functions: they allow
11514 attributes to be looked up by NID and added.
11516 Later something similar to the X509V3 code would be desirable to
11517 automatically handle the encoding, decoding and printing of the
11518 more complex types. The string types like challengePassword can
11519 be handled by the string table functions.
11521 Also modified the multi byte string table handling. Now there is
11522 a 'global mask' which masks out certain types. The table itself
11523 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11524 is useful when for example there is only one permissible type
11525 (as in countryName) and using the mask might result in no valid
11529 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11530 SSL_get_peer_finished to allow applications to obtain the latest
11531 Finished messages sent to the peer or expected from the peer,
11532 respectively. (SSL_get_peer_finished is usually the Finished message
11533 actually received from the peer, otherwise the protocol will be aborted.)
11535 As the Finished message are message digests of the complete handshake
11536 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11537 be used for external authentication procedures when the authentication
11538 provided by SSL/TLS is not desired or is not enough.
11541 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11542 the host supports BWX extension and if Compaq C is present on the
11543 $PATH. Just exploiting of the BWX extension results in 20-30%
11544 performance kick for some algorithms, e.g. DES and RC4 to mention
11545 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11549 *) Add support for MS "fast SGC". This is arguably a violation of the
11550 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11551 weak crypto and after checking the certificate is SGC a second one
11552 with strong crypto. MS SGC stops the first handshake after receiving
11553 the server certificate message and sends a second client hello. Since
11554 a server will typically do all the time consuming operations before
11555 expecting any further messages from the client (server key exchange
11556 is the most expensive) there is little difference between the two.
11558 To get OpenSSL to support MS SGC we have to permit a second client
11559 hello message after we have sent server done. In addition we have to
11560 reset the MAC if we do get this second client hello.
11563 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11564 if a DER encoded private key is RSA or DSA traditional format. Changed
11565 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11566 format DER encoded private key. Newer code should use PKCS#8 format which
11567 has the key type encoded in the ASN1 structure. Added DER private key
11568 support to pkcs8 application.
11571 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11572 ciphersuites has been selected (as required by the SSL 3/TLS 1
11573 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11574 is set, we interpret this as a request to violate the specification
11575 (the worst that can happen is a handshake failure, and 'correct'
11576 behaviour would result in a handshake failure anyway).
11579 *) In SSL_CTX_add_session, take into account that there might be multiple
11580 SSL_SESSION structures with the same session ID (e.g. when two threads
11581 concurrently obtain them from an external cache).
11582 The internal cache can handle only one SSL_SESSION with a given ID,
11583 so if there's a conflict, we now throw out the old one to achieve
11587 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11588 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11589 some routines that use cipher OIDs: some ciphers do not have OIDs
11590 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11594 *) Simplify the trust setting structure and code. Now we just have
11595 two sequences of OIDs for trusted and rejected settings. These will
11596 typically have values the same as the extended key usage extension
11597 and any application specific purposes.
11599 The trust checking code now has a default behaviour: it will just
11600 check for an object with the same NID as the passed id. Functions can
11601 be provided to override either the default behaviour or the behaviour
11602 for a given id. SSL client, server and email already have functions
11603 in place for compatibility: they check the NID and also return "trusted"
11604 if the certificate is self signed.
11607 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11608 traditional format into an EVP_PKEY structure.
11611 *) Add a password callback function PEM_cb() which either prompts for
11612 a password if usr_data is NULL or otherwise assumes it is a null
11613 terminated password. Allow passwords to be passed on command line
11614 environment or config files in a few more utilities.
11617 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11618 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11619 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11620 Update documentation.
11623 *) Support for ASN1 "NULL" type. This could be handled before by using
11624 ASN1_TYPE but there wasn't any function that would try to read a NULL
11625 and produce an error if it couldn't. For compatibility we also have
11626 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11627 don't allocate anything because they don't need to.
11630 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11632 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11634 *) Rebuild of the memory allocation routines used by OpenSSL code and
11635 possibly others as well. The purpose is to make an interface that
11636 provide hooks so anyone can build a separate set of allocation and
11637 deallocation routines to be used by OpenSSL, for example memory
11638 pool implementations, or something else, which was previously hard
11639 since Malloc(), Realloc() and Free() were defined as macros having
11640 the values malloc, realloc and free, respectively (except for Win32
11641 compilations). The same is provided for memory debugging code.
11642 OpenSSL already comes with functionality to find memory leaks, but
11643 this gives people a chance to debug other memory problems.
11645 With these changes, a new set of functions and macros have appeared:
11647 CRYPTO_set_mem_debug_functions() [F]
11648 CRYPTO_get_mem_debug_functions() [F]
11649 CRYPTO_dbg_set_options() [F]
11650 CRYPTO_dbg_get_options() [F]
11651 CRYPTO_malloc_debug_init() [M]
11653 The memory debug functions are NULL by default, unless the library
11654 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11655 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11656 gives the standard debugging functions that come with OpenSSL) or
11657 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11658 provided by the library user) must be used. When the standard
11659 debugging functions are used, CRYPTO_dbg_set_options can be used to
11660 request additional information:
11661 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11662 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11664 Also, things like CRYPTO_set_mem_functions will always give the
11665 expected result (the new set of functions is used for allocation
11666 and deallocation) at all times, regardless of platform and compiler
11669 To finish it up, some functions that were never use in any other
11670 way than through macros have a new API and new semantic:
11672 CRYPTO_dbg_malloc()
11673 CRYPTO_dbg_realloc()
11676 All macros of value have retained their old syntax.
11677 [Richard Levitte and Bodo Moeller]
11679 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11680 ordering of SMIMECapabilities wasn't in "strength order" and there
11681 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11685 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11686 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11687 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11689 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11690 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11691 functionality to handle multipart/signed properly) and a utility
11692 called 'smime' to call all this stuff. This is based on code I
11693 originally wrote for Celo who have kindly allowed it to be
11694 included in OpenSSL.
11697 *) Add variants des_set_key_checked and des_set_key_unchecked of
11698 des_set_key (aka des_key_sched). Global variable des_check_key
11699 decides which of these is called by des_set_key; this way
11700 des_check_key behaves as it always did, but applications and
11701 the library itself, which was buggy for des_check_key == 1,
11702 have a cleaner way to pick the version they need.
11705 *) New function PKCS12_newpass() which changes the password of a
11709 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11710 dynamic mix. In both cases the ids can be used as an index into the
11711 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11712 functions so they accept a list of the field values and the
11713 application doesn't need to directly manipulate the X509_TRUST
11717 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11721 *) Modify the way the V3 extension code looks up extensions. This now
11722 works in a similar way to the object code: we have some "standard"
11723 extensions in a static table which is searched with OBJ_bsearch()
11724 and the application can add dynamic ones if needed. The file
11725 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11726 updated whenever a new extension is added to the core code and kept
11727 in ext_nid order. There is a simple program 'tabtest.c' which checks
11728 this. New extensions are not added too often so this file can readily
11729 be maintained manually.
11731 There are two big advantages in doing things this way. The extensions
11732 can be looked up immediately and no longer need to be "added" using
11733 X509V3_add_standard_extensions(): this function now does nothing.
11734 [Side note: I get *lots* of email saying the extension code doesn't
11735 work because people forget to call this function]
11736 Also no dynamic allocation is done unless new extensions are added:
11737 so if we don't add custom extensions there is no need to call
11738 X509V3_EXT_cleanup().
11741 *) Modify enc utility's salting as follows: make salting the default. Add a
11742 magic header, so unsalted files fail gracefully instead of just decrypting
11743 to garbage. This is because not salting is a big security hole, so people
11744 should be discouraged from doing it.
11747 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11748 digest to be passed on the command line but it only used this
11749 parameter when signing a certificate. Modified so all relevant
11750 operations are affected by the digest parameter including the
11751 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11752 DSA key was used because it didn't fix the digest.
11755 *) Initial certificate chain verify code. Currently tests the untrusted
11756 certificates for consistency with the verify purpose (which is set
11757 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11759 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11760 this is because it will reject chains with invalid extensions whereas
11761 every previous version of OpenSSL and SSLeay made no checks at all.
11763 Trust code: checks the root CA for the relevant trust settings. Trust
11764 settings have an initial value consistent with the verify purpose: e.g.
11765 if the verify purpose is for SSL client use it expects the CA to be
11766 trusted for SSL client use. However the default value can be changed to
11767 permit custom trust settings: one example of this would be to only trust
11768 certificates from a specific "secure" set of CAs.
11770 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11771 which should be used for version portability: especially since the
11772 verify structure is likely to change more often now.
11774 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11775 to set them. If not set then assume SSL clients will verify SSL servers
11778 Two new options to the verify program: -untrusted allows a set of
11779 untrusted certificates to be passed in and -purpose which sets the
11780 intended purpose of the certificate. If a purpose is set then the
11781 new chain verify code is used to check extension consistency.
11784 *) Support for the authority information access extension.
11787 *) Modify RSA and DSA PEM read routines to transparently handle
11788 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11789 public keys in a format compatible with certificate
11790 SubjectPublicKeyInfo structures. Unfortunately there were already
11791 functions called *_PublicKey_* which used various odd formats so
11792 these are retained for compatibility: however the DSA variants were
11793 never in a public release so they have been deleted. Changed dsa/rsa
11794 utilities to handle the new format: note no releases ever handled public
11795 keys so we should be OK.
11797 The primary motivation for this change is to avoid the same fiasco
11798 that dogs private keys: there are several incompatible private key
11799 formats some of which are standard and some OpenSSL specific and
11800 require various evil hacks to allow partial transparent handling and
11801 even then it doesn't work with DER formats. Given the option anything
11802 other than PKCS#8 should be dumped: but the other formats have to
11803 stay in the name of compatibility.
11805 With public keys and the benefit of hindsight one standard format
11806 is used which works with EVP_PKEY, RSA or DSA structures: though
11807 it clearly returns an error if you try to read the wrong kind of key.
11809 Added a -pubkey option to the 'x509' utility to output the public key.
11810 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11811 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11812 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11813 that do the same as the EVP_PKEY_assign_*() except they up the
11814 reference count of the added key (they don't "swallow" the
11818 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11819 CRLs would fail if the file contained no certificates or no CRLs:
11820 added a new function to read in both types and return the number
11821 read: this means that if none are read it will be an error. The
11822 DER versions of the certificate and CRL reader would always fail
11823 because it isn't possible to mix certificates and CRLs in DER format
11824 without choking one or the other routine. Changed this to just read
11825 a certificate: this is the best we can do. Also modified the code
11826 in apps/verify.c to take notice of return codes: it was previously
11827 attempting to read in certificates from NULL pointers and ignoring
11828 any errors: this is one reason why the cert and CRL reader seemed
11829 to work. It doesn't check return codes from the default certificate
11830 routines: these may well fail if the certificates aren't installed.
11833 *) Code to support otherName option in GeneralName.
11836 *) First update to verify code. Change the verify utility
11837 so it warns if it is passed a self signed certificate:
11838 for consistency with the normal behaviour. X509_verify
11839 has been modified to it will now verify a self signed
11840 certificate if *exactly* the same certificate appears
11841 in the store: it was previously impossible to trust a
11842 single self signed certificate. This means that:
11843 openssl verify ss.pem
11844 now gives a warning about a self signed certificate but
11845 openssl verify -CAfile ss.pem ss.pem
11849 *) For servers, store verify_result in SSL_SESSION data structure
11850 (and add it to external session representation).
11851 This is needed when client certificate verifications fails,
11852 but an application-provided verification callback (set by
11853 SSL_CTX_set_cert_verify_callback) allows accepting the session
11854 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11855 but returns 1): When the session is reused, we have to set
11856 ssl->verify_result to the appropriate error code to avoid
11858 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11860 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11861 case in PKCS7_dataInit() where the signed PKCS7 structure
11862 didn't contain any existing data because it was being created.
11863 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11865 *) Add a salt to the key derivation routines in enc.c. This
11866 forms the first 8 bytes of the encrypted file. Also add a
11867 -S option to allow a salt to be input on the command line.
11870 *) New function X509_cmp(). Oddly enough there wasn't a function
11871 to compare two certificates. We do this by working out the SHA1
11872 hash and comparing that. X509_cmp() will be needed by the trust
11876 *) SSL_get1_session() is like SSL_get_session(), but increments
11877 the reference count in the SSL_SESSION returned.
11878 [Geoff Thorpe <geoff@eu.c2.net>]
11880 *) Fix for 'req': it was adding a null to request attributes.
11881 Also change the X509_LOOKUP and X509_INFO code to handle
11882 certificate auxiliary information.
11885 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11889 *) Add the possibility to add extra information to the memory leak
11890 detecting output, to form tracebacks, showing from where each
11891 allocation was originated: CRYPTO_push_info("constant string") adds
11892 the string plus current file name and line number to a per-thread
11893 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11894 is like calling CYRPTO_pop_info() until the stack is empty.
11895 Also updated memory leak detection code to be multi-thread-safe.
11898 *) Add options -text and -noout to pkcs7 utility and delete the
11899 encryption options which never did anything. Update docs.
11902 *) Add options to some of the utilities to allow the pass phrase
11903 to be included on either the command line (not recommended on
11904 OSes like Unix) or read from the environment. Update the
11905 manpages and fix a few bugs.
11908 *) Add a few manpages for some of the openssl commands.
11911 *) Fix the -revoke option in ca. It was freeing up memory twice,
11912 leaking and not finding already revoked certificates.
11915 *) Extensive changes to support certificate auxiliary information.
11916 This involves the use of X509_CERT_AUX structure and X509_AUX
11917 functions. An X509_AUX function such as PEM_read_X509_AUX()
11918 can still read in a certificate file in the usual way but it
11919 will also read in any additional "auxiliary information". By
11920 doing things this way a fair degree of compatibility can be
11921 retained: existing certificates can have this information added
11922 using the new 'x509' options.
11924 Current auxiliary information includes an "alias" and some trust
11925 settings. The trust settings will ultimately be used in enhanced
11926 certificate chain verification routines: currently a certificate
11927 can only be trusted if it is self signed and then it is trusted
11931 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11932 The problem was that one of the replacement routines had not been working
11933 since SSLeay releases. For now the offending routine has been replaced
11934 with non-optimised assembler. Even so, this now gives around 95%
11935 performance improvement for 1024 bit RSA signs.
11938 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11939 handling. Most clients have the effective key size in bits equal to
11940 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11941 A few however don't do this and instead use the size of the decrypted key
11942 to determine the RC2 key length and the AlgorithmIdentifier to determine
11943 the effective key length. In this case the effective key length can still
11944 be 40 bits but the key length can be 168 bits for example. This is fixed
11945 by manually forcing an RC2 key into the EVP_PKEY structure because the
11946 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11947 the key length and effective key length are equal.
11950 *) Add a bunch of functions that should simplify the creation of
11951 X509_NAME structures. Now you should be able to do:
11952 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11953 and have it automatically work out the correct field type and fill in
11954 the structures. The more adventurous can try:
11955 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11956 and it will (hopefully) work out the correct multibyte encoding.
11959 *) Change the 'req' utility to use the new field handling and multibyte
11960 copy routines. Before the DN field creation was handled in an ad hoc
11961 way in req, ca, and x509 which was rather broken and didn't support
11962 BMPStrings or UTF8Strings. Since some software doesn't implement
11963 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11964 using the dirstring_type option. See the new comment in the default
11965 openssl.cnf for more info.
11968 *) Make crypto/rand/md_rand.c more robust:
11969 - Assure unique random numbers after fork().
11970 - Make sure that concurrent threads access the global counter and
11971 md serializably so that we never lose entropy in them
11972 or use exactly the same state in multiple threads.
11973 Access to the large state is not always serializable because
11974 the additional locking could be a performance killer, and
11975 md should be large enough anyway.
11978 *) New file apps/app_rand.c with commonly needed functionality
11979 for handling the random seed file.
11981 Use the random seed file in some applications that previously did not:
11983 dsaparam -genkey (which also ignored its '-rand' option),
11986 x509 (when signing).
11987 Except on systems with /dev/urandom, it is crucial to have a random
11988 seed file at least for key creation, DSA signing, and for DH exchanges;
11989 for RSA signatures we could do without one.
11991 gendh and gendsa (unlike genrsa) used to read only the first byte
11992 of each file listed in the '-rand' option. The function as previously
11993 found in genrsa is now in app_rand.c and is used by all programs
11994 that support '-rand'.
11997 *) In RAND_write_file, use mode 0600 for creating files;
11998 don't just chmod when it may be too late.
12001 *) Report an error from X509_STORE_load_locations
12002 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
12005 *) New function ASN1_mbstring_copy() this copies a string in either
12006 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
12007 into an ASN1_STRING type. A mask of permissible types is passed
12008 and it chooses the "minimal" type to use or an error if not type
12012 *) Add function equivalents to the various macros in asn1.h. The old
12013 macros are retained with an M_ prefix. Code inside the library can
12014 use the M_ macros. External code (including the openssl utility)
12015 should *NOT* in order to be "shared library friendly".
12018 *) Add various functions that can check a certificate's extensions
12019 to see if it usable for various purposes such as SSL client,
12020 server or S/MIME and CAs of these types. This is currently
12021 VERY EXPERIMENTAL but will ultimately be used for certificate chain
12022 verification. Also added a -purpose flag to x509 utility to
12023 print out all the purposes.
12026 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
12030 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
12031 for, obtain and decode and extension and obtain its critical flag.
12032 This allows all the necessary extension code to be handled in a
12033 single function call.
12036 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
12037 platforms. See crypto/rc4/rc4_enc.c for further details.
12040 *) New -noout option to asn1parse. This causes no output to be produced
12041 its main use is when combined with -strparse and -out to extract data
12042 from a file (which may not be in ASN.1 format).
12045 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
12046 when producing the local key id.
12047 [Richard Levitte <levitte@stacken.kth.se>]
12049 *) New option -dhparam in s_server. This allows a DH parameter file to be
12050 stated explicitly. If it is not stated then it tries the first server
12051 certificate file. The previous behaviour hard coded the filename
12055 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
12056 a public key to be input or output. For example:
12057 openssl rsa -in key.pem -pubout -out pubkey.pem
12058 Also added necessary DSA public key functions to handle this.
12061 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
12062 in the message. This was handled by allowing
12063 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
12064 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
12066 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
12067 to the end of the strings whereas this didn't. This would cause problems
12068 if strings read with d2i_ASN1_bytes() were later modified.
12069 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
12071 *) Fix for base64 decode bug. When a base64 bio reads only one line of
12072 data and it contains EOF it will end up returning an error. This is
12073 caused by input 46 bytes long. The cause is due to the way base64
12074 BIOs find the start of base64 encoded data. They do this by trying a
12075 trial decode on each line until they find one that works. When they
12076 do a flag is set and it starts again knowing it can pass all the
12077 data directly through the decoder. Unfortunately it doesn't reset
12078 the context it uses. This means that if EOF is reached an attempt
12079 is made to pass two EOFs through the context and this causes the
12080 resulting error. This can also cause other problems as well. As is
12081 usual with these problems it takes *ages* to find and the fix is
12082 trivial: move one line.
12083 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
12085 *) Ugly workaround to get s_client and s_server working under Windows. The
12086 old code wouldn't work because it needed to select() on sockets and the
12087 tty (for keypresses and to see if data could be written). Win32 only
12088 supports select() on sockets so we select() with a 1s timeout on the
12089 sockets and then see if any characters are waiting to be read, if none
12090 are present then we retry, we also assume we can always write data to
12091 the tty. This isn't nice because the code then blocks until we've
12092 received a complete line of data and it is effectively polling the
12093 keyboard at 1s intervals: however it's quite a bit better than not
12094 working at all :-) A dedicated Windows application might handle this
12095 with an event loop for example.
12098 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
12099 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
12100 will be called when RSA_sign() and RSA_verify() are used. This is useful
12101 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
12102 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
12103 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
12104 This necessitated the support of an extra signature type NID_md5_sha1
12105 for SSL signatures and modifications to the SSL library to use it instead
12106 of calling RSA_public_decrypt() and RSA_private_encrypt().
12109 *) Add new -verify -CAfile and -CApath options to the crl program, these
12110 will lookup a CRL issuers certificate and verify the signature in a
12111 similar way to the verify program. Tidy up the crl program so it
12112 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
12113 less strict. It will now permit CRL extensions even if it is not
12114 a V2 CRL: this will allow it to tolerate some broken CRLs.
12117 *) Initialize all non-automatic variables each time one of the openssl
12118 sub-programs is started (this is necessary as they may be started
12119 multiple times from the "OpenSSL>" prompt).
12120 [Lennart Bang, Bodo Moeller]
12122 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
12123 removing all other RSA functionality (this is what NO_RSA does). This
12124 is so (for example) those in the US can disable those operations covered
12125 by the RSA patent while allowing storage and parsing of RSA keys and RSA
12129 *) Non-copying interface to BIO pairs.
12130 (still largely untested)
12133 *) New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
12134 ASCII string. This was handled independently in various places before.
12137 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
12138 UTF8 strings a character at a time.
12141 *) Use client_version from client hello to select the protocol
12142 (s23_srvr.c) and for RSA client key exchange verification
12143 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
12146 *) Add various utility functions to handle SPKACs, these were previously
12147 handled by poking round in the structure internals. Added new function
12148 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
12149 print, verify and generate SPKACs. Based on an original idea from
12150 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
12153 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
12156 *) Allow the config file extension section to be overwritten on the
12157 command line. Based on an original idea from Massimiliano Pala
12158 <madwolf@comune.modena.it>. The new option is called -extensions
12159 and can be applied to ca, req and x509. Also -reqexts to override
12160 the request extensions in req and -crlexts to override the crl extensions
12164 *) Add new feature to the SPKAC handling in ca. Now you can include
12165 the same field multiple times by preceding it by "XXXX." for example:
12168 this is the same syntax as used in the req config file.
12171 *) Allow certificate extensions to be added to certificate requests. These
12172 are specified in a 'req_extensions' option of the req section of the
12173 config file. They can be printed out with the -text option to req but
12174 are otherwise ignored at present.
12177 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
12178 data read consists of only the final block it would not decrypted because
12179 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
12180 A misplaced 'break' also meant the decrypted final block might not be
12181 copied until the next read.
12184 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
12185 a few extra parameters to the DH structure: these will be useful if
12186 for example we want the value of 'q' or implement X9.42 DH.
12189 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
12190 provides hooks that allow the default DSA functions or functions on a
12191 "per key" basis to be replaced. This allows hardware acceleration and
12192 hardware key storage to be handled without major modification to the
12193 library. Also added low level modexp hooks and CRYPTO_EX structure and
12194 associated functions.
12197 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
12198 as "read only": it can't be written to and the buffer it points to will
12199 not be freed. Reading from a read only BIO is much more efficient than
12200 a normal memory BIO. This was added because there are several times when
12201 an area of memory needs to be read from a BIO. The previous method was
12202 to create a memory BIO and write the data to it, this results in two
12203 copies of the data and an O(n^2) reading algorithm. There is a new
12204 function BIO_new_mem_buf() which creates a read only memory BIO from
12205 an area of memory. Also modified the PKCS#7 routines to use read only
12209 *) Bugfix: ssl23_get_client_hello did not work properly when called in
12210 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
12211 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
12212 but a retry condition occurred while trying to read the rest.
12215 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
12216 NID_pkcs7_encrypted by default: this was wrong since this should almost
12217 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
12218 the encrypted data type: this is a more sensible place to put it and it
12219 allows the PKCS#12 code to be tidied up that duplicated this
12223 *) Changed obj_dat.pl script so it takes its input and output files on
12224 the command line. This should avoid shell escape redirection problems
12228 *) Initial support for certificate extension requests, these are included
12229 in things like Xenroll certificate requests. Included functions to allow
12230 extensions to be obtained and added.
12233 *) -crlf option to s_client and s_server for sending newlines as
12234 CRLF (as required by many protocols).
12237 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
12239 *) Install libRSAglue.a when OpenSSL is built with RSAref.
12240 [Ralf S. Engelschall]
12242 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
12243 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
12245 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
12249 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
12250 DH parameters/keys (q is lost during that conversion, but the resulting
12251 DH parameters contain its length).
12253 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
12254 much faster than DH_generate_parameters (which creates parameters
12255 where p = 2*q + 1), and also the smaller q makes DH computations
12256 much more efficient (160-bit exponentiation instead of 1024-bit
12257 exponentiation); so this provides a convenient way to support DHE
12258 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
12259 utter importance to use
12260 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12262 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
12263 when such DH parameters are used, because otherwise small subgroup
12264 attacks may become possible!
12267 *) Avoid memory leak in i2d_DHparams.
12270 *) Allow the -k option to be used more than once in the enc program:
12271 this allows the same encrypted message to be read by multiple recipients.
12274 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
12275 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
12276 it will always use the numerical form of the OID, even if it has a short
12280 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
12281 method only got called if p,q,dmp1,dmq1,iqmp components were present,
12282 otherwise bn_mod_exp was called. In the case of hardware keys for example
12283 no private key components need be present and it might store extra data
12284 in the RSA structure, which cannot be accessed from bn_mod_exp.
12285 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
12286 private key operations.
12289 *) Added support for SPARC Linux.
12292 *) pem_password_cb function type incompatibly changed from
12293 typedef int pem_password_cb(char *buf, int size, int rwflag);
12295 ....(char *buf, int size, int rwflag, void *userdata);
12296 so that applications can pass data to their callbacks:
12297 The PEM[_ASN1]_{read,write}... functions and macros now take an
12298 additional void * argument, which is just handed through whenever
12299 the password callback is called.
12300 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
12302 New function SSL_CTX_set_default_passwd_cb_userdata.
12304 Compatibility note: As many C implementations push function arguments
12305 onto the stack in reverse order, the new library version is likely to
12306 interoperate with programs that have been compiled with the old
12307 pem_password_cb definition (PEM_whatever takes some data that
12308 happens to be on the stack as its last argument, and the callback
12309 just ignores this garbage); but there is no guarantee whatsoever that
12312 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
12313 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
12314 problems not only on Windows, but also on some Unix platforms.
12315 To avoid problematic command lines, these definitions are now in an
12316 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
12317 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12320 *) MIPS III/IV assembler module is reimplemented.
12323 *) More DES library cleanups: remove references to srand/rand and
12324 delete an unused file.
12327 *) Add support for the free Netwide assembler (NASM) under Win32,
12328 since not many people have MASM (ml) and it can be hard to obtain.
12329 This is currently experimental but it seems to work OK and pass all
12330 the tests. Check out INSTALL.W32 for info.
12333 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12334 without temporary keys kept an extra copy of the server key,
12335 and connections with temporary keys did not free everything in case
12339 *) New function RSA_check_key and new openssl rsa option -check
12340 for verifying the consistency of RSA keys.
12341 [Ulf Moeller, Bodo Moeller]
12343 *) Various changes to make Win32 compile work:
12344 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12345 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12346 comparison" warnings.
12347 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12350 *) Add a debugging option to PKCS#5 v2 key generation function: when
12351 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12352 derived keys are printed to stderr.
12355 *) Copy the flags in ASN1_STRING_dup().
12356 [Roman E. Pavlov <pre@mo.msk.ru>]
12358 *) The x509 application mishandled signing requests containing DSA
12359 keys when the signing key was also DSA and the parameters didn't match.
12361 It was supposed to omit the parameters when they matched the signing key:
12362 the verifying software was then supposed to automatically use the CA's
12363 parameters if they were absent from the end user certificate.
12365 Omitting parameters is no longer recommended. The test was also
12366 the wrong way round! This was probably due to unusual behaviour in
12367 EVP_cmp_parameters() which returns 1 if the parameters match.
12368 This meant that parameters were omitted when they *didn't* match and
12369 the certificate was useless. Certificates signed with 'ca' didn't have
12371 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12373 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12374 The interface is as follows:
12375 Applications can use
12376 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12377 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12378 "off" is now the default.
12379 The library internally uses
12380 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12381 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12382 to disable memory-checking temporarily.
12384 Some inconsistent states that previously were possible (and were
12385 even the default) are now avoided.
12387 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12388 with each memory chunk allocated; this is occasionally more helpful
12389 than just having a counter.
12391 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12393 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12397 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12398 which largely parallels "options", but is for changing API behaviour,
12399 whereas "options" are about protocol behaviour.
12400 Initial "mode" flags are:
12402 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12403 a single record has been written.
12404 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12405 retries use the same buffer location.
12406 (But all of the contents must be
12410 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12413 *) Fix problems with no-hmac etc.
12414 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12416 *) New functions RSA_get_default_method(), RSA_set_method() and
12417 RSA_get_method(). These allows replacement of RSA_METHODs without having
12418 to mess around with the internals of an RSA structure.
12421 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12422 Also really enable memory leak checks in openssl.c and in some
12424 [Chad C. Mulligan, Bodo Moeller]
12426 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12427 up the length of negative integers. This has now been simplified to just
12428 store the length when it is first determined and use it later, rather
12429 than trying to keep track of where data is copied and updating it to
12431 [Steve Henson, reported by Brien Wheeler
12432 <bwheeler@authentica-security.com>]
12434 *) Add a new function PKCS7_signatureVerify. This allows the verification
12435 of a PKCS#7 signature but with the signing certificate passed to the
12436 function itself. This contrasts with PKCS7_dataVerify which assumes the
12437 certificate is present in the PKCS#7 structure. This isn't always the
12438 case: certificates can be omitted from a PKCS#7 structure and be
12439 distributed by "out of band" means (such as a certificate database).
12442 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12443 function prototypes in pem.h, also change util/mkdef.pl to add the
12444 necessary function names.
12447 *) mk1mf.pl (used by Windows builds) did not properly read the
12448 options set by Configure in the top level Makefile, and Configure
12449 was not even able to write more than one option correctly.
12450 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12453 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12454 file to be loaded from a BIO or FILE pointer. The BIO version will
12455 for example allow memory BIOs to contain config info.
12458 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12459 Whoever hopes to achieve shared-library compatibility across versions
12460 must use this, not the compile-time macro.
12461 (Exercise 0.9.4: Which is the minimum library version required by
12463 Note: All this applies only to multi-threaded programs, others don't
12467 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12468 through a BIO pair triggered the default case, i.e.
12469 SSLerr(...,SSL_R_UNKNOWN_STATE).
12472 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12473 can use the SSL library even if none of the specific BIOs is
12477 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12478 for the encoded length.
12479 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12481 *) Add initial documentation of the X509V3 functions.
12484 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12485 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12486 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12487 secure PKCS#8 private key format with a high iteration count.
12490 *) Fix determination of Perl interpreter: A perl or perl5
12491 _directory_ in $PATH was also accepted as the interpreter.
12492 [Ralf S. Engelschall]
12494 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12495 wrong with it but it was very old and did things like calling
12496 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12497 unusual formatting.
12500 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12501 to use the new extension code.
12504 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12505 with macros. This should make it easier to change their form, add extra
12506 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12510 *) Add to configuration table a new entry that can specify an alternative
12511 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12512 according to Mark Crispin <MRC@Panda.COM>.
12516 *) DES CBC did not update the IV. Weird.
12519 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12520 Changing the behaviour of the former might break existing programs --
12521 where IV updating is needed, des_ncbc_encrypt can be used.
12524 *) When bntest is run from "make test" it drives bc to check its
12525 calculations, as well as internally checking them. If an internal check
12526 fails, it needs to cause bc to give a non-zero result or make test carries
12527 on without noticing the failure. Fixed.
12530 *) DES library cleanups.
12533 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12534 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12535 ciphers. NOTE: although the key derivation function has been verified
12536 against some published test vectors it has not been extensively tested
12537 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12541 *) Instead of "mkdir -p", which is not fully portable, use new
12542 Perl script "util/mkdir-p.pl".
12545 *) Rewrite the way password based encryption (PBE) is handled. It used to
12546 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12547 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12548 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12549 the 'parameter' field of the AlgorithmIdentifier is passed to the
12550 underlying key generation function so it must do its own ASN1 parsing.
12551 This has also changed the EVP_PBE_CipherInit() function which now has a
12552 'parameter' argument instead of literal salt and iteration count values
12553 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12556 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12557 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12558 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12559 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12560 value was just used as a "magic string" and not used directly its
12561 value doesn't matter.
12564 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12568 *) "linux-sparc64" configuration (ultrapenguin).
12569 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12570 "linux-sparc" configuration.
12571 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12573 *) config now generates no-xxx options for missing ciphers.
12576 *) Support the EBCDIC character set (work in progress).
12577 File ebcdic.c not yet included because it has a different license.
12578 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12580 *) Support BS2000/OSD-POSIX.
12581 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12583 *) Make callbacks for key generation use void * instead of char *.
12586 *) Make S/MIME samples compile (not yet tested).
12589 *) Additional typesafe stacks.
12592 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12596 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12598 *) New configuration variant "sco5-gcc".
12600 *) Updated some demos.
12601 [Sean O Riordain, Wade Scholine]
12603 *) Add missing BIO_free at exit of pkcs12 application.
12606 *) Fix memory leak in conf.c.
12609 *) Updates for Win32 to assembler version of MD5.
12612 *) Set #! path to perl in apps/der_chop to where we found it
12613 instead of using a fixed path.
12616 *) SHA library changes for irix64-mips4-cc.
12619 *) Improvements for VMS support.
12623 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12625 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12626 This also avoids the problems with SC4.2 and unpatched SC5.
12627 [Andy Polyakov <appro@fy.chalmers.se>]
12629 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12630 These are required because of the typesafe stack would otherwise break
12631 existing code. If old code used a structure member which used to be STACK
12632 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12633 sk_num or sk_value it would produce an error because the num, data members
12634 are not present in STACK_OF. Now it just produces a warning. sk_set
12635 replaces the old method of assigning a value to sk_value
12636 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12637 that does this will no longer work (and should use sk_set instead) but
12638 this could be regarded as a "questionable" behaviour anyway.
12641 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12642 correctly handle encrypted S/MIME data.
12645 *) Change type of various DES function arguments from des_cblock
12646 (which means, in function argument declarations, pointer to char)
12647 to des_cblock * (meaning pointer to array with 8 char elements),
12648 which allows the compiler to do more typechecking; it was like
12649 that back in SSLeay, but with lots of ugly casts.
12651 Introduce new type const_des_cblock.
12654 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12655 problems: find RecipientInfo structure that matches recipient certificate
12656 and initialise the ASN1 structures properly based on passed cipher.
12659 *) Belatedly make the BN tests actually check the results.
12662 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12663 to and from BNs: it was completely broken. New compilation option
12664 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12665 key elements as negative integers.
12668 *) Reorganize and speed up MD5.
12669 [Andy Polyakov <appro@fy.chalmers.se>]
12672 [Richard Levitte <richard@levitte.org>]
12674 *) New option -out to asn1parse to allow the parsed structure to be
12675 output to a file. This is most useful when combined with the -strparse
12676 option to examine the output of things like OCTET STRINGS.
12679 *) Make SSL library a little more fool-proof by not requiring any longer
12680 that SSL_set_{accept,connect}_state be called before
12681 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12682 in many applications because usually everything *appeared* to work as
12683 intended anyway -- now it really works as intended).
12686 *) Move openssl.cnf out of lib/.
12689 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12690 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12691 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12692 [Ralf S. Engelschall]
12694 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12695 handle PKCS#7 enveloped data properly.
12696 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12698 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12699 copying pointers. The cert_st handling is changed by this in
12700 various ways (and thus what used to be known as ctx->default_cert
12701 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12702 any longer when s->cert does not give us what we need).
12703 ssl_cert_instantiate becomes obsolete by this change.
12704 As soon as we've got the new code right (possibly it already is?),
12705 we have solved a couple of bugs of the earlier code where s->cert
12706 was used as if it could not have been shared with other SSL structures.
12708 Note that using the SSL API in certain dirty ways now will result
12709 in different behaviour than observed with earlier library versions:
12710 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12711 does not influence s as it used to.
12713 In order to clean up things more thoroughly, inside SSL_SESSION
12714 we don't use CERT any longer, but a new structure SESS_CERT
12715 that holds per-session data (if available); currently, this is
12716 the peer's certificate chain and, for clients, the server's certificate
12717 and temporary key. CERT holds only those values that can have
12718 meaningful defaults in an SSL_CTX.
12721 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12722 from the internal representation. Various PKCS#7 fixes: remove some
12723 evil casts and set the enc_dig_alg field properly based on the signing
12727 *) Allow PKCS#12 password to be set from the command line or the
12728 environment. Let 'ca' get its config file name from the environment
12729 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12733 *) Allow certificate policies extension to use an IA5STRING for the
12734 organization field. This is contrary to the PKIX definition but
12735 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12739 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12740 without disallowing inline assembler and the like for non-pedantic builds.
12743 *) Support Borland C++ builder.
12744 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12746 *) Support Mingw32.
12749 *) SHA-1 cleanups and performance enhancements.
12750 [Andy Polyakov <appro@fy.chalmers.se>]
12752 *) Sparc v8plus assembler for the bignum library.
12753 [Andy Polyakov <appro@fy.chalmers.se>]
12755 *) Accept any -xxx and +xxx compiler options in Configure.
12758 *) Update HPUX configuration.
12761 *) Add missing sk_<type>_unshift() function to safestack.h
12762 [Ralf S. Engelschall]
12764 *) New function SSL_CTX_use_certificate_chain_file that sets the
12765 "extra_cert"s in addition to the certificate. (This makes sense
12766 only for "PEM" format files, as chains as a whole are not
12770 *) Support verify_depth from the SSL API.
12771 x509_vfy.c had what can be considered an off-by-one-error:
12772 Its depth (which was not part of the external interface)
12773 was actually counting the number of certificates in a chain;
12774 now it really counts the depth.
12777 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12778 instead of X509err, which often resulted in confusing error
12779 messages since the error codes are not globally unique
12780 (e.g. an alleged error in ssl3_accept when a certificate
12781 didn't match the private key).
12783 *) New function SSL_CTX_set_session_id_context that allows to set a default
12784 value (so that you don't need SSL_set_session_id_context for each
12785 connection using the SSL_CTX).
12788 *) OAEP decoding bug fix.
12791 *) Support INSTALL_PREFIX for package builders, as proposed by
12795 *) New Configure options "threads" and "no-threads". For systems
12796 where the proper compiler options are known (currently Solaris
12797 and Linux), "threads" is the default.
12800 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12803 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12804 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12805 such as /usr/local/bin.
12808 *) "make linux-shared" to build shared libraries.
12809 [Niels Poppe <niels@netbox.org>]
12811 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12814 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12815 extension adding in x509 utility.
12818 *) Remove NOPROTO sections and error code comments.
12821 *) Partial rewrite of the DEF file generator to now parse the ANSI
12825 *) New Configure options --prefix=DIR and --openssldir=DIR.
12828 *) Complete rewrite of the error code script(s). It is all now handled
12829 by one script at the top level which handles error code gathering,
12830 header rewriting and C source file generation. It should be much better
12831 than the old method: it now uses a modified version of Ulf's parser to
12832 read the ANSI prototypes in all header files (thus the old K&R definitions
12833 aren't needed for error creation any more) and do a better job of
12834 translating function codes into names. The old 'ASN1 error code embedded
12835 in a comment' is no longer necessary and it doesn't use .err files which
12836 have now been deleted. Also the error code call doesn't have to appear all
12837 on one line (which resulted in some large lines...).
12840 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12843 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12844 0 (which usually indicates a closed connection), but continue reading.
12847 *) Fix some race conditions.
12850 *) Add support for CRL distribution points extension. Add Certificate
12851 Policies and CRL distribution points documentation.
12854 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12857 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12858 8 of keying material. Merlin has also confirmed interop with this fix
12859 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12860 [Merlin Hughes <merlin@baltimore.ie>]
12862 *) Fix lots of warnings.
12863 [Richard Levitte <levitte@stacken.kth.se>]
12865 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12866 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12867 [Richard Levitte <levitte@stacken.kth.se>]
12869 *) Fix problems with sizeof(long) == 8.
12870 [Andy Polyakov <appro@fy.chalmers.se>]
12872 *) Change functions to ANSI C.
12875 *) Fix typos in error codes.
12876 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12878 *) Remove defunct assembler files from Configure.
12881 *) SPARC v8 assembler BIGNUM implementation.
12882 [Andy Polyakov <appro@fy.chalmers.se>]
12884 *) Support for Certificate Policies extension: both print and set.
12885 Various additions to support the r2i method this uses.
12888 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12889 return a const string when you are expecting an allocated buffer.
12892 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12893 types DirectoryString and DisplayText.
12896 *) Add code to allow r2i extensions to access the configuration database,
12897 add an LHASH database driver and add several ctx helper functions.
12900 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12901 fail when they extended the size of a BIGNUM.
12904 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12905 support typesafe stack.
12908 *) Fix typo in SSL_[gs]et_options().
12909 [Nils Frostberg <nils@medcom.se>]
12911 *) Delete various functions and files that belonged to the (now obsolete)
12912 old X509V3 handling code.
12915 *) New Configure option "rsaref".
12918 *) Don't auto-generate pem.h.
12921 *) Introduce type-safe ASN.1 SETs.
12924 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12925 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12927 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12928 that links with OpenSSL (well at least cause lots of warnings), but fear
12929 not: the conversion is trivial, and it eliminates loads of evil casts. A
12930 few STACKed things have been converted already. Feel free to convert more.
12931 In the fullness of time, I'll do away with the STACK type altogether.
12934 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12935 specified in <certfile> by updating the entry in the index.txt file.
12936 This way one no longer has to edit the index.txt file manually for
12937 revoking a certificate. The -revoke option does the gory details now.
12938 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12940 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12941 `-text' option at all and this way the `-noout -text' combination was
12942 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12943 [Ralf S. Engelschall]
12945 *) Make sure a corresponding plain text error message exists for the
12946 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12947 verify callback function determined that a certificate was revoked.
12948 [Ralf S. Engelschall]
12950 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12951 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12952 all available ciphers including rc5, which was forgotten until now.
12953 In order to let the testing shell script know which algorithms
12954 are available, a new (up to now undocumented) command
12955 "openssl list-cipher-commands" is used.
12958 *) Bugfix: s_client occasionally would sleep in select() when
12959 it should have checked SSL_pending() first.
12962 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12963 the raw DSA values prior to ASN.1 encoding.
12966 *) Tweaks to Configure
12967 [Niels Poppe <niels@netbox.org>]
12969 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12973 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12976 *) New config option to avoid instructions that are illegal on the 80386.
12977 The default code is faster, but requires at least a 486.
12980 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12981 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12982 same as SSL2_VERSION anyway.
12985 *) New "-showcerts" option for s_client.
12988 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12989 application. Various cleanups and fixes.
12992 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12993 modify error routines to work internally. Add error codes and PBE init
12994 to library startup routines.
12997 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12998 packing functions to asn1 and evp. Changed function names and error
12999 codes along the way.
13002 *) PKCS12 integration: and so it begins... First of several patches to
13003 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
13004 objects to objects.h
13007 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
13008 and display support for Thawte strong extranet extension.
13011 *) Add LinuxPPC support.
13012 [Jeff Dubrule <igor@pobox.org>]
13014 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
13015 bn_div_words in alpha.s.
13016 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
13018 *) Make sure the RSA OAEP test is skipped under -DRSAref because
13019 OAEP isn't supported when OpenSSL is built with RSAref.
13020 [Ulf Moeller <ulf@fitug.de>]
13022 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
13023 so they no longer are missing under -DNOPROTO.
13024 [Soren S. Jorvang <soren@t.dk>]
13027 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
13029 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
13030 doesn't work when the session is reused. Coming soon!
13033 *) Fix a security hole, that allows sessions to be reused in the wrong
13034 context thus bypassing client cert protection! All software that uses
13035 client certs and session caches in multiple contexts NEEDS PATCHING to
13036 allow session reuse! A fuller solution is in the works.
13037 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
13039 *) Some more source tree cleanups (removed obsolete files
13040 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
13041 permission on "config" script to be executable) and a fix for the INSTALL
13043 [Ulf Moeller <ulf@fitug.de>]
13045 *) Remove some legacy and erroneous uses of malloc, free instead of
13047 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
13049 *) Make rsa_oaep_test return non-zero on error.
13050 [Ulf Moeller <ulf@fitug.de>]
13052 *) Add support for native Solaris shared libraries. Configure
13053 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
13054 if someone would make that last step automatic.
13055 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
13057 *) ctx_size was not built with the right compiler during "make links". Fixed.
13060 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
13061 except NULL ciphers". This means the default cipher list will no longer
13062 enable NULL ciphers. They need to be specifically enabled e.g. with
13063 the string "DEFAULT:eNULL".
13066 *) Fix to RSA private encryption routines: if p < q then it would
13067 occasionally produce an invalid result. This will only happen with
13068 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
13071 *) Be less restrictive and allow also `perl util/perlpath.pl
13072 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
13073 because this way one can also use an interpreter named `perl5' (which is
13074 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
13075 installed as `perl').
13076 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13078 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
13079 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13081 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
13082 advapi32.lib to Win32 build and change the pem test comparison
13083 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
13084 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
13085 and crypto/des/ede_cbcm_enc.c.
13088 *) DES quad checksum was broken on big-endian architectures. Fixed.
13091 *) Comment out two functions in bio.h that aren't implemented. Fix up the
13092 Win32 test batch file so it (might) work again. The Win32 test batch file
13093 is horrible: I feel ill....
13096 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
13097 in e_os.h. Audit of header files to check ANSI and non ANSI
13098 sections: 10 functions were absent from non ANSI section and not exported
13099 from Windows DLLs. Fixed up libeay.num for new functions.
13102 *) Make `openssl version' output lines consistent.
13103 [Ralf S. Engelschall]
13105 *) Fix Win32 symbol export lists for BIO functions: Added
13106 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
13107 to ms/libeay{16,32}.def.
13108 [Ralf S. Engelschall]
13110 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
13111 fine under Unix and passes some trivial tests I've now added. But the
13112 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
13113 added to make sure no one expects that this stuff really works in the
13114 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
13115 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
13117 [Ralf S. Engelschall]
13119 *) Fix the generation of two part addresses in perl.
13120 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
13122 *) Add config entry for Linux on MIPS.
13123 [John Tobey <jtobey@channel1.com>]
13125 *) Make links whenever Configure is run, unless we are on Windoze.
13128 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
13129 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
13133 *) Add a useful kludge to allow package maintainers to specify compiler and
13134 other platforms details on the command line without having to patch the
13135 Configure script every time: One now can use ``perl Configure
13136 <id>:<details>'', i.e. platform ids are allowed to have details appended
13137 to them (separated by colons). This is treated as there would be a static
13138 pre-configured entry in Configure's %table under key <id> with value
13139 <details> and ``perl Configure <id>'' is called. So, when you want to
13140 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
13141 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
13142 now, which overrides the FreeBSD-elf entry on-the-fly.
13143 [Ralf S. Engelschall]
13145 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
13148 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
13149 on the `perl Configure ...' command line. This way one can compile
13150 OpenSSL libraries with Position Independent Code (PIC) which is needed
13151 for linking it into DSOs.
13152 [Ralf S. Engelschall]
13154 *) Remarkably, export ciphers were totally broken and no-one had noticed!
13158 *) Cleaned up the LICENSE document: The official contact for any license
13159 questions now is the OpenSSL core team under openssl-core@openssl.org.
13160 And add a paragraph about the dual-license situation to make sure people
13161 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
13162 to the OpenSSL toolkit.
13163 [Ralf S. Engelschall]
13165 *) General source tree makefile cleanups: Made `making xxx in yyy...'
13166 display consistent in the source tree and replaced `/bin/rm' by `rm'.
13167 Additionally cleaned up the `make links' target: Remove unnecessary
13168 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
13169 to speed processing and no longer clutter the display with confusing
13170 stuff. Instead only the actually done links are displayed.
13171 [Ralf S. Engelschall]
13173 *) Permit null encryption ciphersuites, used for authentication only. It used
13174 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
13175 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
13179 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
13180 signed attributes when verifying signatures (this would break them),
13181 the detached data encoding was wrong and public keys obtained using
13182 X509_get_pubkey() weren't freed.
13185 *) Add text documentation for the BUFFER functions. Also added a work around
13186 to a Win95 console bug. This was triggered by the password read stuff: the
13187 last character typed gets carried over to the next fread(). If you were
13188 generating a new cert request using 'req' for example then the last
13189 character of the passphrase would be CR which would then enter the first
13193 *) Added the new `Includes OpenSSL Cryptography Software' button as
13194 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
13195 button and can be used by applications based on OpenSSL to show the
13196 relationship to the OpenSSL project.
13197 [Ralf S. Engelschall]
13199 *) Remove confusing variables in function signatures in files
13200 ssl/ssl_lib.c and ssl/ssl.h.
13201 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13203 *) Don't install bss_file.c under PREFIX/include/
13204 [Lennart Bong <lob@kulthea.stacken.kth.se>]
13206 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
13207 functions that return function pointers and has support for NT specific
13208 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
13209 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
13210 unsigned to signed types: this was killing the Win32 compile.
13213 *) Add new certificate file to stack functions,
13214 SSL_add_dir_cert_subjects_to_stack() and
13215 SSL_add_file_cert_subjects_to_stack(). These largely supplant
13216 SSL_load_client_CA_file(), and can be used to add multiple certs easily
13217 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
13218 This means that Apache-SSL and similar packages don't have to mess around
13219 to add as many CAs as they want to the preferred list.
13222 *) Experiment with doxygen documentation. Currently only partially applied to
13224 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
13225 openssl.doxy as the configuration file.
13228 *) Get rid of remaining C++-style comments which strict C compilers hate.
13229 [Ralf S. Engelschall, pointed out by Carlos Amengual]
13231 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
13232 compiled in by default: it has problems with large keys.
13235 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
13236 DH private keys and/or callback functions which directly correspond to
13237 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
13238 is needed for applications which have to configure certificates on a
13239 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
13241 For the RSA certificate situation is makes no difference, but
13242 for the DSA certificate situation this fixes the "no shared cipher"
13243 problem where the OpenSSL cipher selection procedure failed because the
13244 temporary keys were not overtaken from the context and the API provided
13245 no way to reconfigure them.
13246 The new functions now let applications reconfigure the stuff and they
13247 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
13248 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
13249 non-public-API function ssl_cert_instantiate() is used as a helper
13250 function and also to reduce code redundancy inside ssl_rsa.c.
13251 [Ralf S. Engelschall]
13253 *) Move s_server -dcert and -dkey options out of the undocumented feature
13254 area because they are useful for the DSA situation and should be
13255 recognized by the users.
13256 [Ralf S. Engelschall]
13258 *) Fix the cipher decision scheme for export ciphers: the export bits are
13259 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
13260 SSL_EXP_MASK. So, the original variable has to be used instead of the
13261 already masked variable.
13262 [Richard Levitte <levitte@stacken.kth.se>]
13264 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
13265 [Richard Levitte <levitte@stacken.kth.se>]
13267 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
13268 from `int' to `unsigned int' because it's a length and initialized by
13269 EVP_DigestFinal() which expects an `unsigned int *'.
13270 [Richard Levitte <levitte@stacken.kth.se>]
13272 *) Don't hard-code path to Perl interpreter on shebang line of Configure
13273 script. Instead use the usual Shell->Perl transition trick.
13274 [Ralf S. Engelschall]
13276 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
13277 (in addition to RSA certificates) to match the behaviour of `openssl dsa
13278 -noout -modulus' as it's already the case for `openssl rsa -noout
13279 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
13280 currently the public key is printed (a decision which was already done by
13281 `openssl dsa -modulus' in the past) which serves a similar purpose.
13282 Additionally the NO_RSA no longer completely removes the whole -modulus
13283 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
13285 [Ralf S. Engelschall]
13287 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
13288 BIO. See the source (crypto/evp/bio_ok.c) for more info.
13289 [Arne Ansper <arne@ats.cyber.ee>]
13291 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
13292 to be added. Now both 'req' and 'ca' can use new objects defined in the
13296 *) Add cool BIO that does syslog (or event log on NT).
13297 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
13299 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
13300 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
13301 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
13302 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
13305 *) Add preliminary config info for new extension code.
13308 *) Make RSA_NO_PADDING really use no padding.
13309 [Ulf Moeller <ulf@fitug.de>]
13311 *) Generate errors when private/public key check is done.
13314 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
13315 for some CRL extensions and new objects added.
13318 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13319 key usage extension and fuller support for authority key id.
13322 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13323 padding method for RSA, which is recommended for new applications in PKCS
13324 #1 v2.0 (RFC 2437, October 1998).
13325 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13326 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13327 against Bleichbacher's attack on RSA.
13328 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13331 *) Updates to the new SSL compression code
13332 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13334 *) Fix so that the version number in the master secret, when passed
13335 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13336 (because the server will not accept higher), that the version number
13337 is 0x03,0x01, not 0x03,0x00
13338 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13340 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13341 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13342 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13345 *) Support for RAW extensions where an arbitrary extension can be
13346 created by including its DER encoding. See apps/openssl.cnf for
13350 *) Make sure latest Perl versions don't interpret some generated C array
13351 code as Perl array code in the crypto/err/err_genc.pl script.
13352 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13354 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13355 not many people have the assembler. Various Win32 compilation fixes and
13356 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13357 build instructions.
13360 *) Modify configure script 'Configure' to automatically create crypto/date.h
13361 file under Win32 and also build pem.h from pem.org. New script
13362 util/mkfiles.pl to create the MINFO file on environments that can't do a
13363 'make files': perl util/mkfiles.pl >MINFO should work.
13366 *) Major rework of DES function declarations, in the pursuit of correctness
13367 and purity. As a result, many evil casts evaporated, and some weirdness,
13368 too. You may find this causes warnings in your code. Zapping your evil
13369 casts will probably fix them. Mostly.
13372 *) Fix for a typo in asn1.h. Bug fix to object creation script
13373 obj_dat.pl. It considered a zero in an object definition to mean
13374 "end of object": none of the objects in objects.h have any zeros
13375 so it wasn't spotted.
13376 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13378 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13379 Masking (CBCM). In the absence of test vectors, the best I have been able
13380 to do is check that the decrypt undoes the encrypt, so far. Send me test
13381 vectors if you have them.
13384 *) Correct calculation of key length for export ciphers (too much space was
13385 allocated for null ciphers). This has not been tested!
13388 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13389 message is now correct (it understands "crypto" and "ssl" on its
13390 command line). There is also now an "update" option. This will update
13391 the util/ssleay.num and util/libeay.num files with any new functions.
13393 perl util/mkdef.pl crypto ssl update
13394 it will update them.
13397 *) Overhauled the Perl interface (perl/*):
13398 - ported BN stuff to OpenSSL's different BN library
13399 - made the perl/ source tree CVS-aware
13400 - renamed the package from SSLeay to OpenSSL (the files still contain
13401 their history because I've copied them in the repository)
13402 - removed obsolete files (the test scripts will be replaced
13403 by better Test::Harness variants in the future)
13404 [Ralf S. Engelschall]
13406 *) First cut for a very conservative source tree cleanup:
13407 1. merge various obsolete readme texts into doc/ssleay.txt
13408 where we collect the old documents and readme texts.
13409 2. remove the first part of files where I'm already sure that we no
13410 longer need them because of three reasons: either they are just temporary
13411 files which were left by Eric or they are preserved original files where
13412 I've verified that the diff is also available in the CVS via "cvs diff
13413 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13414 the crypto/md/ stuff).
13415 [Ralf S. Engelschall]
13417 *) More extension code. Incomplete support for subject and issuer alt
13418 name, issuer and authority key id. Change the i2v function parameters
13419 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13420 what that's for :-) Fix to ASN1 macro which messed up
13421 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13424 *) Preliminary support for ENUMERATED type. This is largely copied from the
13428 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13429 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13431 *) Make sure `make rehash' target really finds the `openssl' program.
13432 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13434 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13435 like to hear about it if this slows down other processors.
13438 *) Add CygWin32 platform information to Configure script.
13439 [Alan Batie <batie@aahz.jf.intel.com>]
13441 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13442 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13444 *) New program nseq to manipulate netscape certificate sequences
13447 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13451 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13452 but the BN code had some problems that would cause failures when
13453 doing certificate verification and some other functions.
13454 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13456 *) Add ASN1 and PEM code to support netscape certificate sequences.
13459 *) Add ASN1 and PEM code to support netscape certificate sequences.
13462 *) Add several PKIX and private extended key usage OIDs.
13465 *) Modify the 'ca' program to handle the new extension code. Modify
13466 openssl.cnf for new extension format, add comments.
13469 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13470 and add a sample to openssl.cnf so req -x509 now adds appropriate
13474 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13475 error code, add initial support to X509_print() and x509 application.
13478 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13479 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13480 stuff is currently isolated and isn't even compiled yet.
13483 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13484 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13485 Removed the versions check from X509 routines when loading extensions:
13486 this allows certain broken certificates that don't set the version
13487 properly to be processed.
13490 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13491 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13492 can still be regenerated with "make depend".
13495 *) Spelling mistake in C version of CAST-128.
13496 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13498 *) Changes to the error generation code. The perl script err-code.pl
13499 now reads in the old error codes and retains the old numbers, only
13500 adding new ones if necessary. It also only changes the .err files if new
13501 codes are added. The makefiles have been modified to only insert errors
13502 when needed (to avoid needlessly modifying header files). This is done
13503 by only inserting errors if the .err file is newer than the auto generated
13504 C file. To rebuild all the error codes from scratch (the old behaviour)
13505 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13506 or delete all the .err files.
13509 *) CAST-128 was incorrectly implemented for short keys. The C version has
13510 been fixed, but is untested. The assembler versions are also fixed, but
13511 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13512 to regenerate it if needed.
13513 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13514 Hagino <itojun@kame.net>]
13516 *) File was opened incorrectly in randfile.c.
13517 [Ulf Möller <ulf@fitug.de>]
13519 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13520 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13521 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13522 al: it's just almost always a UTCTime. Note this patch adds new error
13523 codes so do a "make errors" if there are problems.
13526 *) Correct Linux 1 recognition in config.
13527 [Ulf Möller <ulf@fitug.de>]
13529 *) Remove pointless MD5 hash when using DSA keys in ca.
13530 [Anonymous <nobody@replay.com>]
13532 *) Generate an error if given an empty string as a cert directory. Also
13533 generate an error if handed NULL (previously returned 0 to indicate an
13534 error, but didn't set one).
13535 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13537 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13540 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13541 parameters. This was causing a warning which killed off the Win32 compile.
13544 *) Remove C++ style comments from crypto/bn/bn_local.h.
13545 [Neil Costigan <neil.costigan@celocom.com>]
13547 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13548 based on a text string, looking up short and long names and finally
13549 "dot" format. The "dot" format stuff didn't work. Added new function
13550 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13551 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13552 OID is not part of the table.
13555 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13556 X509_LOOKUP_by_alias().
13559 *) Sort openssl functions by name.
13562 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13563 encryption from sample DSA keys (in case anyone is interested the password
13567 *) Make _all_ *_free functions accept a NULL pointer.
13568 [Frans Heymans <fheymans@isaserver.be>]
13570 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13572 [Anonymous <nobody@replay.com>]
13574 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13575 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13577 *) Don't blow it for numeric -newkey arguments to apps/req.
13578 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13580 *) Temp key "for export" tests were wrong in s3_srvr.c.
13581 [Anonymous <nobody@replay.com>]
13583 *) Add prototype for temp key callback functions
13584 SSL_CTX_set_tmp_{rsa,dh}_callback().
13587 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13588 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13591 *) X509_name_add_entry() freed the wrong thing after an error.
13592 [Arne Ansper <arne@ats.cyber.ee>]
13594 *) rsa_eay.c would attempt to free a NULL context.
13595 [Arne Ansper <arne@ats.cyber.ee>]
13597 *) BIO_s_socket() had a broken should_retry() on Windoze.
13598 [Arne Ansper <arne@ats.cyber.ee>]
13600 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13601 [Arne Ansper <arne@ats.cyber.ee>]
13603 *) Make sure the already existing X509_STORE->depth variable is initialized
13604 in X509_STORE_new(), but document the fact that this variable is still
13605 unused in the certificate verification process.
13606 [Ralf S. Engelschall]
13608 *) Fix the various library and apps files to free up pkeys obtained from
13609 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13612 *) Fix reference counting in X509_PUBKEY_get(). This makes
13613 demos/maurice/example2.c work, amongst others, probably.
13614 [Steve Henson and Ben Laurie]
13616 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13617 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13618 are no longer created. This way we have a single and consistent command
13619 line interface `openssl <command>', similar to `cvs <command>'.
13620 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13622 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13623 BIT STRING wrapper always have zero unused bits.
13626 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13629 *) Make the top-level INSTALL documentation easier to understand.
13632 *) Makefiles updated to exit if an error occurs in a sub-directory
13633 make (including if user presses ^C) [Paul Sutton]
13635 *) Make Montgomery context stuff explicit in RSA data structure.
13638 *) Fix build order of pem and err to allow for generated pem.h.
13641 *) Fix renumbering bug in X509_NAME_delete_entry().
13644 *) Enhanced the err-ins.pl script so it makes the error library number
13645 global and can add a library name. This is needed for external ASN1 and
13646 other error libraries.
13649 *) Fixed sk_insert which never worked properly.
13652 *) Fix ASN1 macros so they can handle indefinite length constructed
13653 EXPLICIT tags. Some non standard certificates use these: they can now
13657 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13658 into a single doc/ssleay.txt bundle. This way the information is still
13659 preserved but no longer messes up this directory. Now it's new room for
13660 the new set of documentation files.
13661 [Ralf S. Engelschall]
13663 *) SETs were incorrectly DER encoded. This was a major pain, because they
13664 shared code with SEQUENCEs, which aren't coded the same. This means that
13665 almost everything to do with SETs or SEQUENCEs has either changed name or
13666 number of arguments.
13667 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13669 *) Fix test data to work with the above.
13672 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13673 was already fixed by Eric for 0.9.1 it seems.
13674 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13676 *) Autodetect FreeBSD3.
13679 *) Fix various bugs in Configure. This affects the following platforms:
13683 unixware-2.0-pentium
13687 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13688 before they are needed.
13691 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13695 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13697 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13698 changed SSLeay to OpenSSL in version strings.
13699 [Ralf S. Engelschall]
13701 *) Some fixups to the top-level documents.
13704 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13705 because the symlink to include/ was missing.
13706 [Ralf S. Engelschall]
13708 *) Incorporated the popular no-RSA/DSA-only patches
13709 which allow to compile a RSA-free SSLeay.
13710 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13712 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13713 when "ssleay" is still not found.
13714 [Ralf S. Engelschall]
13716 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13717 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13719 *) Updated the README file.
13720 [Ralf S. Engelschall]
13722 *) Added various .cvsignore files in the CVS repository subdirs
13723 to make a "cvs update" really silent.
13724 [Ralf S. Engelschall]
13726 *) Recompiled the error-definition header files and added
13727 missing symbols to the Win32 linker tables.
13728 [Ralf S. Engelschall]
13730 *) Cleaned up the top-level documents;
13731 o new files: CHANGES and LICENSE
13732 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13733 o merged COPYRIGHT into LICENSE
13734 o removed obsolete TODO file
13735 o renamed MICROSOFT to INSTALL.W32
13736 [Ralf S. Engelschall]
13738 *) Removed dummy files from the 0.9.1b source tree:
13739 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13740 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13741 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13742 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13743 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13744 [Ralf S. Engelschall]
13746 *) Added various platform portability fixes.
13749 *) The Genesis of the OpenSSL rpject:
13750 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13751 Young and Tim J. Hudson created while they were working for C2Net until
13753 [The OpenSSL Project]
13756 Changes between 0.9.0b and 0.9.1b [not released]
13758 *) Updated a few CA certificates under certs/
13761 *) Changed some BIGNUM api stuff.
13764 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13765 DGUX x86, Linux Alpha, etc.
13768 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13769 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13773 *) Add -strparse option to asn1pars program which parses nested
13775 [Dr Stephen Henson <shenson@bigfoot.com>]
13777 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13780 *) DSA fix for "ca" program.
13783 *) Added "-genkey" option to "dsaparam" program.
13786 *) Added RIPE MD160 (rmd160) message digest.
13789 *) Added -a (all) option to "ssleay version" command.
13792 *) Added PLATFORM define which is the id given to Configure.
13795 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13798 *) Extended the ASN.1 parser routines.
13801 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13804 *) Added a BN_CTX to the BN library.
13807 *) Fixed the weak key values in DES library
13810 *) Changed API in EVP library for cipher aliases.
13813 *) Added support for RC2/64bit cipher.
13816 *) Converted the lhash library to the crypto/mem.c functions.
13819 *) Added more recognized ASN.1 object ids.
13822 *) Added more RSA padding checks for SSL/TLS.
13825 *) Added BIO proxy/filter functionality.
13828 *) Added extra_certs to SSL_CTX which can be used
13829 send extra CA certificates to the client in the CA cert chain sending
13830 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13833 *) Now Fortezza is denied in the authentication phase because
13834 this is key exchange mechanism is not supported by SSLeay at all.
13837 *) Additional PKCS1 checks.
13840 *) Support the string "TLSv1" for all TLS v1 ciphers.
13843 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13844 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13847 *) Fixed a few memory leaks.
13850 *) Fixed various code and comment typos.
13853 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13854 bytes sent in the client random.
13855 [Edward Bishop <ebishop@spyglass.com>]