4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx] ###
28 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
29 have been converted to Markdown with the goal to produce documents
30 which not only look pretty when viewed online in the browser, but
31 remain well readable inside a plain text editor.
33 To achieve this goal, a 'minimalistic' Markdown style has been applied
34 which avoids formatting elements that interfere too much with the
35 reading flow in the text file. For example, it
37 * avoids [ATX headings][] and uses [setext headings][] instead
38 (which works for `<h1>` and `<h2>` headings only).
39 * avoids [inline links][] and uses [reference links][] instead.
40 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
42 [ATX headings]: https://github.github.com/gfm/#atx-headings
43 [setext headings]: https://github.github.com/gfm/#setext-headings
44 [inline links]: https://github.github.com/gfm/#inline-link
45 [reference links]: https://github.github.com/gfm/#reference-link
46 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
47 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
52 * The test suite is changed to preserve results of each test recipe.
53 A new directory test-runs/ with subdirectories named like the
54 test recipes are created in the build tree for this purpose.
58 * The command line utilities ecparam and ec have been deprecated. Instead
59 use the pkeyparam, pkey and genpkey programs.
63 * All of the low level RSA functions have been deprecated including:
65 RSA_new_method, RSA_bits, RSA_size, RSA_security_bits,
66 RSA_get0_pss_params, RSA_get_version, RSA_get0_engine,
67 RSA_generate_key_ex, RSA_generate_multi_prime_key,
68 RSA_X931_derive_ex, RSA_X931_generate_key_ex, RSA_check_key,
69 RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
70 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
71 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
72 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
73 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING,
74 RSA_blinding_on, RSA_blinding_off, RSA_setup_blinding,
75 RSA_padding_add_PKCS1_type_1, RSA_padding_check_PKCS1_type_1,
76 RSA_padding_add_PKCS1_type_2, RSA_padding_check_PKCS1_type_2,
77 PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP, RSA_padding_check_PKCS1_OAEP,
78 RSA_padding_add_PKCS1_OAEP_mgf1, RSA_padding_check_PKCS1_OAEP_mgf1,
79 RSA_padding_add_SSLv23, RSA_padding_check_SSLv23,
80 RSA_padding_add_none, RSA_padding_check_none, RSA_padding_add_X931,
81 RSA_padding_check_X931, RSA_X931_hash_id, RSA_verify_PKCS1_PSS,
82 RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
83 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
84 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
85 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
86 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
87 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
88 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
89 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
90 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
91 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
92 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
93 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
94 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
96 Use of these low level functions has been informally discouraged for a long
97 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
98 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
99 L<EVP_PKEY_decrypt(3)>.
103 * X509 certificates signed using SHA1 are no longer allowed at security
105 In TLS/SSL the default security level is 1. It can be set either
106 using the cipher string with @SECLEVEL, or calling
107 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
108 a call to SSL_CTX_use_certificate() will fail if the security level is not
110 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
111 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
116 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
117 deprecated. Instead use the pkeyparam, pkey, genpkey and pkeyparam
118 programs respectively.
122 * All of the low level DH functions have been deprecated including:
124 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
125 DH_new_method, DH_bits, DH_size, DH_security_bits, DH_get_ex_new_index,
126 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
127 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
128 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
129 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
130 DH_KDF_X9_42, DH_get0_engine, DH_get_length, DH_set_length, DH_meth_new,
131 DH_meth_free, DH_meth_dup, DH_meth_get0_name, DH_meth_set1_name,
132 DH_meth_get_flags, DH_meth_set_flags, DH_meth_get0_app_data,
133 DH_meth_set0_app_data, DH_meth_get_generate_key,
134 DH_meth_set_generate_key, DH_meth_get_compute_key,
135 DH_meth_set_compute_key, DH_meth_get_bn_mod_exp,
136 DH_meth_set_bn_mod_exp, DH_meth_get_init, DH_meth_set_init,
137 DH_meth_get_finish, DH_meth_set_finish, DH_meth_get_generate_params
138 and DH_meth_set_generate_params.
140 Use of these low level functions has been informally discouraged for a long
141 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
142 and L<EVP_PKEY_derive(3)>.
146 * All of the low level DSA functions have been deprecated including:
148 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
149 DSA_get_default_method, DSA_set_method, DSA_get_method, DSA_new_method,
150 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
151 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
152 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
153 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
154 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
155 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
156 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
157 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
158 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
159 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
160 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
162 Use of these low level functions has been informally discouraged for a long
163 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
164 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
168 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
169 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
170 This means that applications don't have to look at the curve NID and
171 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
172 However, they still can, that EVP_PKEY_set_alias_type() call acts as
173 a no-op when the EVP_PKEY is already of the given type.
175 Parameter and key generation is also reworked to make it possible
176 to generate EVP_PKEY_SM2 parameters and keys without having to go
177 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
178 However, code that does the latter will still work as before.
182 * Deprecated low level ECDH and ECDSA functions. These include:
184 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
185 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
188 Use of these low level functions has been informally discouraged for a long
189 time. Instead applications should use the EVP_PKEY_derive(3),
190 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
194 * Deprecated the EC_KEY_METHOD functions. These include:
196 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
197 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
198 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
199 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
200 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
201 EC_KEY_METHOD_get_verify.
203 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
207 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
208 and EVP_PKEY_decrypt() instead.
209 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
210 and EVP_PKEY_encrypt() instead.
214 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
215 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
216 a new formulation to include all the things it can be used for,
217 as well as words of caution.
221 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
222 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
226 * All of the low level HMAC functions have been deprecated including:
228 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
229 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
232 Use of these low level functions has been informally discouraged for a long
233 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
234 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
235 and L<EVP_MAC_final(3)>.
239 * Over two thousand fixes were made to the documentation, including:
240 - Common options (such as -rand/-writerand, TLS version control, etc)
241 were refactored and point to newly-enhanced descriptions in openssl.pod.
242 - Added style conformance for all options (with help from Richard Levitte),
243 documented all reported missing options, added a CI build to check
244 that all options are documented and that no unimplemented options
246 - Documented some internals, such as all use of environment variables.
247 - Addressed all internal broken L<> references.
251 * All of the low level CMAC functions have been deprecated including:
253 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
254 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
256 Use of these low level functions has been informally discouraged for a long
257 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
258 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
259 and L<EVP_MAC_final(3)>.
263 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
264 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
267 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
268 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
269 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
270 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
271 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
272 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
273 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
274 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
275 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
276 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
278 Use of these low level functions has been informally discouraged
279 for a long time. Applications should use the EVP_DigestInit_ex(3),
280 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
284 * Corrected the documentation of the return values from the EVP_DigestSign*
285 set of functions. The documentation mentioned negative values for some
286 errors, but this was never the case, so the mention of negative values
289 Code that followed the documentation and thereby check with something
290 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
294 * All of the low level cipher functions have been deprecated including:
296 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
297 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
298 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
299 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
300 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
301 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
302 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
303 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
304 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
305 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
306 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
307 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
308 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
309 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
310 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
311 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
312 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
313 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
314 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
315 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
316 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
317 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
318 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
319 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
320 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
321 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
322 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
323 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
324 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
326 Use of these low level functions has been informally discouraged for
327 a long time. Applications should use the high level EVP APIs, e.g.
328 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
329 equivalently named decrypt functions instead.
331 *Matt Caswell and Paul Dale*
333 * Removed include/openssl/opensslconf.h.in and replaced it with
334 include/openssl/configuration.h.in, which differs in not including
335 <openssl/macros.h>. A short header include/openssl/opensslconf.h
336 was added to include both.
338 This allows internal hacks where one might need to modify the set
339 of configured macros, for example this if deprecated symbols are
340 still supposed to be available internally:
342 #include <openssl/configuration.h>
344 #undef OPENSSL_NO_DEPRECATED
345 #define OPENSSL_SUPPRESS_DEPRECATED
347 #include <openssl/macros.h>
349 This should not be used by applications that use the exported
350 symbols, as that will lead to linking errors.
354 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
355 used in exponentiation with 512-bit moduli. No EC algorithms are
356 affected. Analysis suggests that attacks against 2-prime RSA1024,
357 3-prime RSA1536, and DSA1024 as a result of this defect would be very
358 difficult to perform and are not believed likely. Attacks against DH512
359 are considered just feasible. However, for an attack the target would
360 have to re-use the DH512 private key, which is not recommended anyway.
361 Also applications directly using the low level API BN_mod_exp may be
362 affected if they use BN_FLG_CONSTTIME.
367 * Most memory-debug features have been deprecated, and the functionality
368 replaced with no-ops.
372 * Introduced a new method type and API, OSSL_SERIALIZER, to
373 represent generic serializers. An implementation is expected to
374 be able to serialize an object associated with a given name (such
375 as an algorithm name for an asymmetric key) into forms given by
376 implementation properties.
378 Serializers are primarily used from inside libcrypto, through
379 calls to functions like EVP_PKEY_print_private(),
380 PEM_write_bio_PrivateKey() and similar.
382 Serializers are specified in such a way that they can be made to
383 directly handle the provider side portion of an object, if this
384 provider side part comes from the same provider as the serializer
385 itself, but can also be made to handle objects in parametrized
386 form (as an OSSL_PARAM array of data). This allows a provider to
387 offer generic serializers as a service for any other provider.
391 * Added a .pragma directive to the syntax of configuration files, to
392 allow varying behavior in a supported and predictable manner.
393 Currently added pragma:
397 This allows dollar signs to be a keyword character unless it's
398 followed by a opening brace or parenthesis. This is useful for
399 platforms where dollar signs are commonly used in names, such as
400 volume names and system directory names on VMS.
404 * Added functionality to create an EVP_PKEY from user data. This
405 is effectively the same as creating a RSA, DH or DSA object and
406 then assigning them to an EVP_PKEY, but directly using algorithm
407 agnostic EVP functions. A benefit is that this should be future
408 proof for public key algorithms to come.
412 * Change the interpretation of the '--api' configuration option to
413 mean that this is a desired API compatibility level with no
414 further meaning. The previous interpretation, that this would
415 also mean to remove all deprecated symbols up to and including
416 the given version, no requires that 'no-deprecated' is also used
417 in the configuration.
419 When building applications, the desired API compatibility level
420 can be set with the OPENSSL_API_COMPAT macro like before. For
421 API compatibility version below 3.0, the old style numerical
422 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
423 For version 3.0 and on, the value is expected to be the decimal
424 value calculated from the major and minor version like this:
426 MAJOR * 10000 + MINOR * 100
430 -DOPENSSL_API_COMPAT=30000 For 3.0
431 -DOPENSSL_API_COMPAT=30200 For 3.2
433 To hide declarations that are deprecated up to and including the
434 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
435 given when building the application as well.
439 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
440 access to certificate and CRL stores via URIs and OSSL_STORE
443 This adds the following functions:
445 - X509_LOOKUP_store()
446 - X509_STORE_load_file()
447 - X509_STORE_load_path()
448 - X509_STORE_load_store()
449 - SSL_add_store_cert_subjects_to_stack()
450 - SSL_CTX_set_default_verify_store()
451 - SSL_CTX_load_verify_file()
452 - SSL_CTX_load_verify_dir()
453 - SSL_CTX_load_verify_store()
455 Also, the following functions are now deprecated:
457 - X509_STORE_load_locations() (use X509_STORE_load_file(),
458 X509_STORE_load_path() or X509_STORE_load_store() instead)
459 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
460 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
464 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
465 The presence of this system service is determined at run-time.
469 * Added functionality to create an EVP_PKEY context based on data
470 for methods from providers. This takes an algorithm name and a
471 property query string and simply stores them, with the intent
472 that any operation that uses this context will use those strings
473 to fetch the needed methods implicitly, thereby making the port
474 of application written for pre-3.0 OpenSSL easier.
478 * The undocumented function NCONF_WIN32() has been deprecated; for
479 conversion details see the HISTORY section of doc/man5/config.pod
483 * Introduced the new functions EVP_DigestSignInit_ex() and
484 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
485 EVP_DigestVerifyUpdate() have been converted to functions. See the man
486 pages for further details.
490 * Over two thousand fixes were made to the documentation, including:
491 adding missing command flags, better style conformance, documentation
494 *Rich Salz, Richard Levitte*
496 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
497 X25519, X448, Ed25519 and Ed448.
501 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
506 * Deprecated the public definition of ERR_STATE as well as the function
507 ERR_get_state(). This is done in preparation of making ERR_STATE an
512 * Added ERR functionality to give callers access to the stored function
513 names that have replaced the older function code based functions.
515 New functions are ERR_get_error_func(), ERR_peek_error_func(),
516 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
517 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
518 and ERR_peek_last_error_all().
520 These functions have become deprecated: ERR_get_error_line_data(),
521 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
522 ERR_func_error_string().
526 * Extended testing to be verbose for failing tests only. The make variables
527 VERBOSE_FAILURE or VF can be used to enable this:
529 $ make VF=1 test # Unix
530 $ mms /macro=(VF=1) test ! OpenVMS
531 $ nmake VF=1 test # Windows
536 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
537 used even when parsing explicit parameters, when loading a serialized key
538 or calling `EC_GROUP_new_from_ecpkparameters()`/
539 `EC_GROUP_new_from_ecparameters()`.
540 This prevents bypass of security hardening and performance gains,
541 especially for curves with specialized EC_METHODs.
542 By default, if a key encoded with explicit parameters is loaded and later
543 serialized, the output is still encoded with explicit parameters, even if
544 internally a "named" EC_GROUP is used for computation.
548 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
549 this change, EC_GROUP_set_generator would accept order and/or cofactor as
550 NULL. After this change, only the cofactor parameter can be NULL. It also
551 does some minimal sanity checks on the passed order.
556 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
557 An attack is simple, if the first CMS_recipientInfo is valid but the
558 second CMS_recipientInfo is chosen ciphertext. If the second
559 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
560 encryption key will be replaced by garbage, and the message cannot be
561 decoded, but if the RSA decryption fails, the correct encryption key is
562 used and the recipient will not notice the attack.
563 As a work around for this potential attack the length of the decrypted
564 key must be equal to the cipher default key length, in case the
565 certifiate is not given and all recipientInfo are tried out.
566 The old behaviour can be re-enabled in the CMS code by setting the
567 CMS_DEBUG_DECRYPT flag.
571 * Early start up entropy quality from the DEVRANDOM seed source has been
572 improved for older Linux systems. The RAND subsystem will wait for
573 /dev/random to be producing output before seeding from /dev/urandom.
574 The seeded state is stored for future library initialisations using
575 a system global shared memory segment. The shared memory identifier
576 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
577 the desired value. The default identifier is 114.
581 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
582 when primes for RSA keys are computed.
583 Since we previously always generated primes == 2 (mod 3) for RSA keys,
584 the 2-prime and 3-prime RSA modules were easy to distinguish, since
585 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
586 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
587 This avoids possible fingerprinting of newly generated RSA modules.
591 * Correct the extended master secret constant on EBCDIC systems. Without this
592 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
593 negotiate EMS will fail. Unfortunately this also means that TLS connections
594 between EBCDIC systems with this fix, and EBCDIC systems without this
595 fix will fail if they negotiate EMS.
599 * Changed the library initialisation so that the config file is now loaded
600 by default. This was already the case for libssl. It now occurs for both
601 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
602 OPENSSL_init_crypto() to suppress automatic loading of a config file.
606 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
607 where the former acts as a replacement for ERR_put_error(), and the
608 latter replaces the combination ERR_put_error()+ERR_add_error_data().
609 ERR_raise_data() adds more flexibility by taking a format string and
610 an arbitrary number of arguments following it, to be processed with
615 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
616 to check if a named provider is loaded and available. When called, it
617 will also activate all fallback providers if such are still present.
621 * Enforce a minimum DH modulus size of 512 bits.
625 * Changed DH parameters to generate the order q subgroup instead of 2q.
626 Previously generated DH parameters are still accepted by DH_check
627 but DH_generate_key works around that by clearing bit 0 of the
628 private key for those. This avoids leaking bit 0 of the private key.
632 * Significantly reduce secure memory usage by the randomness pools.
636 * {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
641 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
642 algorithms. An implementation of a key exchange algorithm can be obtained
643 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
644 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
645 the older EVP_PKEY_derive_init() function. See the man pages for the new
646 functions for further details.
650 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
654 * Removed the function names from error messages and deprecated the
657 * Removed NextStep support and the macro OPENSSL_UNISTD
661 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
662 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
663 Also removed "export var as function" capability; we do not export
664 variables, only functions.
668 * RC5_32_set_key has been changed to return an int type, with 0 indicating
669 an error and 1 indicating success. In previous versions of OpenSSL this
670 was a void type. If a key was set longer than the maximum possible this
675 * Support SM2 signing and verification schemes with X509 certificate.
679 * Use SHA256 as the default digest for TS query in the ts app.
683 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
684 This checks that the salt length is at least 128 bits, the derived key
685 length is at least 112 bits, and that the iteration count is at least 1000.
686 For backwards compatibility these checks are disabled by default in the
687 default provider, but are enabled by default in the fips provider.
688 To enable or disable these checks use the control
689 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
693 * Default cipher lists/suites are now available via a function, the
694 #defines are deprecated.
698 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
699 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
700 for Windows Store apps easier. Also, the "no-uplink" option has been added.
704 * Join the directories crypto/x509 and crypto/x509v3
708 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
709 This changes the size when using the genpkey app when no size is given. It
710 fixes an omission in earlier changes that changed all RSA, DSA and DH
711 generation apps to use 2048 bits by default.
715 * Added command 'openssl kdf' that uses the EVP_KDF API.
719 * Added command 'openssl mac' that uses the EVP_MAC API.
723 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
724 as default directories. Also added the command 'openssl info'
725 for scripting purposes.
729 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
730 deprecated. These undocumented functions were never integrated into the EVP
731 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
732 Bi-directional IGE mode. These modes were never formally standardised and
733 usage of these functions is believed to be very small. In particular
734 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
735 is ever used. The security implications are believed to be minimal, but
736 this issue was never fixed for backwards compatibility reasons. New code
737 should not use these modes.
741 * Add prediction resistance to the DRBG reseeding process.
745 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
746 mandated by IEEE Std 1619-2018.
750 * Added newline escaping functionality to a filename when using openssl dgst.
751 This output format is to replicate the output format found in the '*sum'
752 checksum programs. This aims to preserve backward compatibility.
754 *Matt Eaton, Richard Levitte, and Paul Dale*
756 * Removed the heartbeat message in DTLS feature, as it has very
757 little usage and doesn't seem to fulfill a valuable purpose.
758 The configuration option is now deprecated.
762 * Changed the output of 'openssl {digestname} < file' to display the
763 digest name in its output.
767 * Added a new generic trace API which provides support for enabling
768 instrumentation through trace output. This feature is mainly intended
769 as an aid for developers and is disabled by default. To utilize it,
770 OpenSSL needs to be configured with the `enable-trace` option.
772 If the tracing API is enabled, the application can activate trace output
773 by registering BIOs as trace channels for a number of tracing and debugging
776 The 'openssl' application has been expanded to enable any of the types
777 available via environment variables defined by the user, and serves as
778 one possible example on how to use this functionality.
780 *Richard Levitte & Matthias St. Pierre*
782 * Added build tests for C++. These are generated files that only do one
783 thing, to include one public OpenSSL head file each. This tests that
784 the public header files can be usefully included in a C++ application.
786 This test isn't enabled by default. It can be enabled with the option
787 'enable-buildtest-c++'.
791 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
795 * Add KMAC to EVP_MAC.
799 * Added property based algorithm implementation selection framework to
804 * Added SCA hardening for modular field inversion in EC_GROUP through
805 a new dedicated field_inv() pointer in EC_METHOD.
806 This also addresses a leakage affecting conversions from projective
807 to affine coordinates.
809 *Billy Bob Brumley, Nicola Tuveri*
811 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
812 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
813 those algorithms that were already supported through the EVP_PKEY API
814 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
815 and scrypt are now wrappers that call EVP_KDF.
819 * Build devcrypto engine as a dynamic engine.
823 * Add keyed BLAKE2 to EVP_MAC.
827 * Fix a bug in the computation of the endpoint-pair shared secret used
828 by DTLS over SCTP. This breaks interoperability with older versions
829 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
830 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
831 interoperability with such broken implementations. However, enabling
832 this switch breaks interoperability with correct implementations.
834 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
835 re-used X509_PUBKEY object if the second PUBKEY is malformed.
839 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
843 * Change the license to the Apache License v2.0.
847 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
849 - Major releases (indicated by incrementing the MAJOR release number)
850 may introduce incompatible API/ABI changes.
851 - Minor releases (indicated by incrementing the MINOR release number)
852 may introduce new features but retain API/ABI compatibility.
853 - Patch releases (indicated by incrementing the PATCH number)
854 are intended for bug fixes and other improvements of existing
855 features only (like improving performance or adding documentation)
856 and retain API/ABI compatibility.
860 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
864 * Remove the 'dist' target and add a tarball building script. The
865 'dist' target has fallen out of use, and it shouldn't be
866 necessary to configure just to create a source distribution.
870 * Recreate the OS390-Unix config target. It no longer relies on a
871 special script like it did for OpenSSL pre-1.1.0.
875 * Instead of having the source directories listed in Configure, add
876 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
881 * Add GMAC to EVP_MAC.
885 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
889 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
890 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
891 to facilitate the continued use of MACs through raw private keys in
892 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
896 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
897 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
901 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
902 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
903 are retained for backwards compatibility.
907 * AES-XTS mode now enforces that its two keys are different to mitigate
908 the attacked described in "Efficient Instantiations of Tweakable
909 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
910 Details of this attack can be obtained from:
911 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
915 * Rename the object files, i.e. give them other names than in previous
916 versions. Their names now include the name of the final product, as
917 well as its type mnemonic (bin, lib, shlib).
921 * Added new option for 'openssl list', '-objects', which will display the
922 list of built in objects, i.e. OIDs with names.
926 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
927 improves application performance by removing data copies and providing
928 applications with zero-copy system calls such as sendfile and splice.
936 ### Changes between 1.1.1d and 1.1.1e [xx XXX xxxx] ###
938 * *[VMS only]* The header files that the VMS compilers include automatically,
939 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
940 that the C++ compiler doesn't understand. This is a shortcoming in the
941 compiler, but can be worked around with `__cplusplus` guards.
943 C++ applications that use OpenSSL libraries must be compiled using the
944 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
945 functions. Otherwise, only functions with symbols of less than 31
946 characters can be used, as the linker will not be able to successfully
947 resolve symbols with longer names.
951 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
952 The presence of this system service is determined at run-time.
956 * Added newline escaping functionality to a filename when using openssl dgst.
957 This output format is to replicate the output format found in the '*sum'
958 checksum programs. This aims to preserve backward compatibility.
960 *Matt Eaton, Richard Levitte, and Paul Dale*
962 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
967 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019] ###
969 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
970 number generator (RNG). This was intended to include protection in the
971 event of a fork() system call in order to ensure that the parent and child
972 processes did not share the same RNG state. However this protection was not
973 being used in the default case.
975 A partial mitigation for this issue is that the output from a high
976 precision timer is mixed into the RNG state so the likelihood of a parent
977 and child process sharing state is significantly reduced.
979 If an application already calls OPENSSL_init_crypto() explicitly using
980 OPENSSL_INIT_ATFORK then this problem does not occur at all.
983 *Matthias St. Pierre*
985 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
986 used even when parsing explicit parameters, when loading a serialized key
987 or calling `EC_GROUP_new_from_ecpkparameters()`/
988 `EC_GROUP_new_from_ecparameters()`.
989 This prevents bypass of security hardening and performance gains,
990 especially for curves with specialized EC_METHODs.
991 By default, if a key encoded with explicit parameters is loaded and later
992 serialized, the output is still encoded with explicit parameters, even if
993 internally a "named" EC_GROUP is used for computation.
997 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
998 this change, EC_GROUP_set_generator would accept order and/or cofactor as
999 NULL. After this change, only the cofactor parameter can be NULL. It also
1000 does some minimal sanity checks on the passed order.
1005 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1006 An attack is simple, if the first CMS_recipientInfo is valid but the
1007 second CMS_recipientInfo is chosen ciphertext. If the second
1008 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1009 encryption key will be replaced by garbage, and the message cannot be
1010 decoded, but if the RSA decryption fails, the correct encryption key is
1011 used and the recipient will not notice the attack.
1012 As a work around for this potential attack the length of the decrypted
1013 key must be equal to the cipher default key length, in case the
1014 certifiate is not given and all recipientInfo are tried out.
1015 The old behaviour can be re-enabled in the CMS code by setting the
1016 CMS_DEBUG_DECRYPT flag.
1021 * Early start up entropy quality from the DEVRANDOM seed source has been
1022 improved for older Linux systems. The RAND subsystem will wait for
1023 /dev/random to be producing output before seeding from /dev/urandom.
1024 The seeded state is stored for future library initialisations using
1025 a system global shared memory segment. The shared memory identifier
1026 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1027 the desired value. The default identifier is 114.
1031 * Correct the extended master secret constant on EBCDIC systems. Without this
1032 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1033 negotiate EMS will fail. Unfortunately this also means that TLS connections
1034 between EBCDIC systems with this fix, and EBCDIC systems without this
1035 fix will fail if they negotiate EMS.
1039 * Use Windows installation paths in the mingw builds
1041 Mingw isn't a POSIX environment per se, which means that Windows
1042 paths should be used for installation.
1047 * Changed DH_check to accept parameters with order q and 2q subgroups.
1048 With order 2q subgroups the bit 0 of the private key is not secret
1049 but DH_generate_key works around that by clearing bit 0 of the
1050 private key for those. This avoids leaking bit 0 of the private key.
1054 * Significantly reduce secure memory usage by the randomness pools.
1058 * Revert the DEVRANDOM_WAIT feature for Linux systems
1060 The DEVRANDOM_WAIT feature added a select() call to wait for the
1061 /dev/random device to become readable before reading from the
1062 /dev/urandom device.
1064 It turned out that this change had negative side effects on
1065 performance which were not acceptable. After some discussion it
1066 was decided to revert this feature and leave it up to the OS
1067 resp. the platform maintainer to ensure a proper initialization
1068 during early boot time.
1070 *Matthias St. Pierre*
1072 ### Changes between 1.1.1b and 1.1.1c [28 May 2019] ###
1074 * Add build tests for C++. These are generated files that only do one
1075 thing, to include one public OpenSSL head file each. This tests that
1076 the public header files can be usefully included in a C++ application.
1078 This test isn't enabled by default. It can be enabled with the option
1079 'enable-buildtest-c++'.
1083 * Enable SHA3 pre-hashing for ECDSA and DSA.
1087 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1088 This changes the size when using the genpkey app when no size is given. It
1089 fixes an omission in earlier changes that changed all RSA, DSA and DH
1090 generation apps to use 2048 bits by default.
1094 * Reorganize the manual pages to consistently have RETURN VALUES,
1095 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1096 util/fix-doc-nits accordingly.
1098 *Paul Yang, Joshua Lock*
1100 * Add the missing accessor EVP_PKEY_get0_engine()
1104 * Have apps like 's_client' and 's_server' output the signature scheme
1105 along with other cipher suite parameters when debugging.
1109 * Make OPENSSL_config() error agnostic again.
1113 * Do the error handling in RSA decryption constant time.
1117 * Prevent over long nonces in ChaCha20-Poly1305.
1119 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1120 for every encryption operation. RFC 7539 specifies that the nonce value
1121 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1122 and front pads the nonce with 0 bytes if it is less than 12
1123 bytes. However it also incorrectly allows a nonce to be set of up to 16
1124 bytes. In this case only the last 12 bytes are significant and any
1125 additional leading bytes are ignored.
1127 It is a requirement of using this cipher that nonce values are
1128 unique. Messages encrypted using a reused nonce value are susceptible to
1129 serious confidentiality and integrity attacks. If an application changes
1130 the default nonce length to be longer than 12 bytes and then makes a
1131 change to the leading bytes of the nonce expecting the new value to be a
1132 new unique nonce then such an application could inadvertently encrypt
1133 messages with a reused nonce.
1135 Additionally the ignored bytes in a long nonce are not covered by the
1136 integrity guarantee of this cipher. Any application that relies on the
1137 integrity of these ignored leading bytes of a long nonce may be further
1138 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1139 is safe because no such use sets such a long nonce value. However user
1140 applications that use this cipher directly and set a non-default nonce
1141 length to be longer than 12 bytes may be vulnerable.
1143 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1149 * Add DEVRANDOM_WAIT feature for Linux systems
1151 On older Linux systems where the getrandom() system call is not available,
1152 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1153 Contrary to getrandom(), the /dev/urandom device will not block during
1154 early boot when the kernel CSPRNG has not been seeded yet.
1156 To mitigate this known weakness, use select() to wait for /dev/random to
1157 become readable before reading from /dev/urandom.
1159 * Ensure that SM2 only uses SM3 as digest algorithm
1163 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019] ###
1165 * Change the info callback signals for the start and end of a post-handshake
1166 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1167 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1168 confused by this and assume that a TLSv1.2 renegotiation has started. This
1169 can break KeyUpdate handling. Instead we no longer signal the start and end
1170 of a post handshake message exchange (although the messages themselves are
1171 still signalled). This could break some applications that were expecting
1172 the old signals. However without this KeyUpdate is not usable for many
1177 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018] ###
1179 * Timing vulnerability in DSA signature generation
1181 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1182 timing side channel attack. An attacker could use variations in the signing
1183 algorithm to recover the private key.
1185 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1190 * Timing vulnerability in ECDSA signature generation
1192 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1193 timing side channel attack. An attacker could use variations in the signing
1194 algorithm to recover the private key.
1196 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1201 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1202 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1203 of two gigabytes and the error handling improved.
1205 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1206 categorized as a normal bug, not a security issue, because the DRBG reseeds
1207 automatically and is fully functional even without additional randomness
1208 provided by the application.
1210 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018] ###
1212 * Add a new ClientHello callback. Provides a callback interface that gives
1213 the application the ability to adjust the nascent SSL object at the
1214 earliest stage of ClientHello processing, immediately after extensions have
1215 been collected but before they have been processed. In particular, this
1216 callback can adjust the supported TLS versions in response to the contents
1221 * Add SM2 base algorithm support.
1225 * s390x assembly pack: add (improved) hardware-support for the following
1226 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1227 aes-cfb/cfb8, aes-ecb.
1231 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1232 parameter is no longer accepted, as it leads to a corrupt table. NULL
1233 pem_str is reserved for alias entries only.
1237 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1238 step for prime curves. The new implementation is based on formulae from
1239 differential addition-and-doubling in homogeneous projective coordinates
1240 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1241 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1242 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1243 to work in projective coordinates.
1245 *Billy Bob Brumley, Nicola Tuveri*
1247 * Change generating and checking of primes so that the error rate of not
1248 being prime depends on the intended use based on the size of the input.
1249 For larger primes this will result in more rounds of Miller-Rabin.
1250 The maximal error rate for primes with more than 1080 bits is lowered
1253 *Kurt Roeckx, Annie Yousar*
1255 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1259 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1260 moving between systems, and to avoid confusion when a Windows build is
1261 done with mingw vs with MSVC. For POSIX installs, there's still a
1262 symlink or copy named 'tsget' to avoid that confusion as well.
1266 * Revert blinding in ECDSA sign and instead make problematic addition
1267 length-invariant. Switch even to fixed-length Montgomery multiplication.
1271 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1272 step for binary curves. The new implementation is based on formulae from
1273 differential addition-and-doubling in mixed Lopez-Dahab projective
1274 coordinates, modified to independently blind the operands.
1276 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1278 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1279 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1280 EC_METHODs to implement their own specialized "ladder step", to take
1281 advantage of more favorable coordinate systems or more efficient
1282 differential addition-and-doubling algorithms.
1284 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1286 * Modified the random device based seed sources to keep the relevant
1287 file descriptors open rather than reopening them on each access.
1288 This allows such sources to operate in a chroot() jail without
1289 the associated device nodes being available. This behaviour can be
1290 controlled using RAND_keep_random_devices_open().
1294 * Numerous side-channel attack mitigations have been applied. This may have
1295 performance impacts for some algorithms for the benefit of improved
1296 security. Specific changes are noted in this change log by their respective
1301 * AIX shared library support overhaul. Switch to AIX "natural" way of
1302 handling shared libraries, which means collecting shared objects of
1303 different versions and bitnesses in one common archive. This allows to
1304 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1305 doesn't affect the way 3rd party applications are linked, only how
1306 multi-version installation is managed.
1310 * Make ec_group_do_inverse_ord() more robust and available to other
1311 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1312 mitigations are applied to the fallback BN_mod_inverse().
1313 When using this function rather than BN_mod_inverse() directly, new
1314 EC cryptosystem implementations are then safer-by-default.
1318 * Add coordinate blinding for EC_POINT and implement projective
1319 coordinate blinding for generic prime curves as a countermeasure to
1320 chosen point SCA attacks.
1322 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1324 * Add blinding to ECDSA and DSA signatures to protect against side channel
1325 attacks discovered by Keegan Ryan (NCC Group).
1329 * Enforce checking in the pkeyutl command line app to ensure that the input
1330 length does not exceed the maximum supported digest length when performing
1331 a sign, verify or verifyrecover operation.
1335 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1336 I/O in combination with something like select() or poll() will hang. This
1337 can be turned off again using SSL_CTX_clear_mode().
1338 Many applications do not properly handle non-application data records, and
1339 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1340 around the problems in those applications, but can also break some.
1341 It's recommended to read the manpages about SSL_read(), SSL_write(),
1342 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1343 SSL_CTX_set_read_ahead() again.
1347 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1348 now allow empty (zero character) pass phrases.
1352 * Apply blinding to binary field modular inversion and remove patent
1353 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1357 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1358 binary and prime elliptic curves.
1362 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1363 constant time fixed point multiplication.
1367 * Revise elliptic curve scalar multiplication with timing attack
1368 defenses: ec_wNAF_mul redirects to a constant time implementation
1369 when computing fixed point and variable point multiplication (which
1370 in OpenSSL are mostly used with secret scalars in keygen, sign,
1371 ECDH derive operations).
1372 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1375 * Updated CONTRIBUTING
1379 * Updated DRBG / RAND to request nonce and additional low entropy
1380 randomness from the system.
1382 *Matthias St. Pierre*
1384 * Updated 'openssl rehash' to use OpenSSL consistent default.
1388 * Moved the load of the ssl_conf module to libcrypto, which helps
1389 loading engines that libssl uses before libssl is initialised.
1393 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1397 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1399 *Ingo Schwarze, Rich Salz*
1401 * Added output of accepting IP address and port for 'openssl s_server'
1405 * Added a new API for TLSv1.3 ciphersuites:
1406 SSL_CTX_set_ciphersuites()
1407 SSL_set_ciphersuites()
1411 * Memory allocation failures consistently add an error to the error
1416 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1417 in libcrypto when run as setuid/setgid.
1421 * Load any config file by default when libssl is used.
1425 * Added new public header file <openssl/rand_drbg.h> and documentation
1426 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1428 *Matthias St. Pierre*
1430 * QNX support removed (cannot find contributors to get their approval
1431 for the license change).
1435 * TLSv1.3 replay protection for early data has been implemented. See the
1436 SSL_read_early_data() man page for further details.
1440 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1441 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1442 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1443 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1444 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1445 configuration has been separated out. See the ciphers man page or the
1446 SSL_CTX_set_ciphersuites() man page for more information.
1450 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1451 in responder mode now supports the new "-multi" option, which
1452 spawns the specified number of child processes to handle OCSP
1453 requests. The "-timeout" option now also limits the OCSP
1454 responder's patience to wait to receive the full client request
1455 on a newly accepted connection. Child processes are respawned
1456 as needed, and the CA index file is automatically reloaded
1457 when changed. This makes it possible to run the "ocsp" responder
1458 as a long-running service, making the OpenSSL CA somewhat more
1459 feature-complete. In this mode, most diagnostic messages logged
1460 after entering the event loop are logged via syslog(3) rather than
1465 * Added support for X448 and Ed448. Heavily based on original work by
1470 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1471 objects loaded. This adds the functions OSSL_STORE_expect() and
1472 OSSL_STORE_find() as well as needed tools to construct searches and
1473 get the search data out of them.
1477 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1478 version of OpenSSL should review their configuration settings to ensure
1479 that they are still appropriate for TLSv1.3. For further information see:
1480 https://wiki.openssl.org/index.php/TLS1.3
1484 * Grand redesign of the OpenSSL random generator
1486 The default RAND method now utilizes an AES-CTR DRBG according to
1487 NIST standard SP 800-90Ar1. The new random generator is essentially
1488 a port of the default random generator from the OpenSSL FIPS 2.0
1489 object module. It is a hybrid deterministic random bit generator
1490 using an AES-CTR bit stream and which seeds and reseeds itself
1491 automatically using trusted system entropy sources.
1493 Some of its new features are:
1494 - Support for multiple DRBG instances with seed chaining.
1495 - The default RAND method makes use of a DRBG.
1496 - There is a public and private DRBG instance.
1497 - The DRBG instances are fork-safe.
1498 - Keep all global DRBG instances on the secure heap if it is enabled.
1499 - The public and private DRBG instance are per thread for lock free
1502 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1504 * Changed Configure so it only says what it does and doesn't dump
1505 so much data. Instead, ./configdata.pm should be used as a script
1506 to display all sorts of configuration data.
1510 * Added processing of "make variables" to Configure.
1514 * Added SHA512/224 and SHA512/256 algorithm support.
1518 * The last traces of Netware support, first removed in 1.1.0, have
1523 * Get rid of Makefile.shared, and in the process, make the processing
1524 of certain files (rc.obj, or the .def/.map/.opt files produced from
1525 the ordinal files) more visible and hopefully easier to trace and
1526 debug (or make silent).
1530 * Make it possible to have environment variable assignments as
1531 arguments to config / Configure.
1535 * Add multi-prime RSA (RFC 8017) support.
1539 * Add SM3 implemented according to GB/T 32905-2016
1540 * Jack Lloyd <jack.lloyd@ribose.com>,
1541 Ronald Tse <ronald.tse@ribose.com>,
1542 Erick Borsboom <erick.borsboom@ribose.com> *
1544 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1545 as documented in RFC6066.
1546 Based on a patch from Tomasz Moń
1548 *Filipe Raimundo da Silva*
1550 * Add SM4 implemented according to GB/T 32907-2016.
1551 * Jack Lloyd <jack.lloyd@ribose.com>,
1552 Ronald Tse <ronald.tse@ribose.com>,
1553 Erick Borsboom <erick.borsboom@ribose.com> *
1555 * Reimplement -newreq-nodes and ERR_error_string_n; the
1556 original author does not agree with the license change.
1560 * Add ARIA AEAD TLS support.
1564 * Some macro definitions to support VS6 have been removed. Visual
1565 Studio 6 has not worked since 1.1.0
1569 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1570 without clearing the errors.
1574 * Add "atfork" functions. If building on a system that without
1575 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1576 requirements. The RAND facility now uses/requires this.
1584 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1585 not possible to disable entirely. However, it's still possible to
1586 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1589 To disable, configure with 'no-ui-console'. 'no-ui' is still
1590 possible to use as an alias. Check at compile time with the
1591 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1592 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1596 * Add a STORE module, which implements a uniform and URI based reader of
1597 stores that can contain keys, certificates, CRLs and numerous other
1598 objects. The main API is loosely based on a few stdio functions,
1599 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1600 OSSL_STORE_error and OSSL_STORE_close.
1601 The implementation uses backends called "loaders" to implement arbitrary
1602 URI schemes. There is one built in "loader" for the 'file' scheme.
1606 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1607 then adjusted to work on FreeBSD 8.4 as well.
1608 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1609 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1613 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1614 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1615 error code calls like this:
1617 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1619 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1620 that can be encoded in C. For the foreseeable future, this will only
1623 *Richard Levitte and Tim Hudson*
1625 * Removed BSD cryptodev engine.
1629 * Add a build target 'build_all_generated', to build all generated files
1630 and only that. This can be used to prepare everything that requires
1631 things like perl for a system that lacks perl and then move everything
1632 to that system and do the rest of the build there.
1636 * In the UI interface, make it possible to duplicate the user data. This
1637 can be used by engines that need to retain the data for a longer time
1638 than just the call where this user data is passed.
1642 * Ignore the '-named_curve auto' value for compatibility of applications
1645 *Tomas Mraz <tmraz@fedoraproject.org>*
1647 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1648 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1649 alerts across multiple records (some of which could be empty). In practice
1650 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1651 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1652 support this at all. Supporting it adds significant complexity to the
1653 record layer, and its removal is unlikely to cause interoperability
1658 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1659 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1660 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1665 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1666 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1668 *Richard Levitte, Andy Polyakov*
1670 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1675 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1676 platform rather than 'mingw'.
1680 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1681 success if they are asked to add an object which already exists
1682 in the store. This change cascades to other functions which load
1683 certificates and CRLs.
1687 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1688 facilitate stack unwinding even from assembly subroutines.
1692 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1693 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1697 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1698 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1699 which is the minimum version we support.
1703 * Certificate time validation (X509_cmp_time) enforces stricter
1704 compliance with RFC 5280. Fractional seconds and timezone offsets
1705 are no longer allowed.
1709 * Add support for ARIA
1713 * s_client will now send the Server Name Indication (SNI) extension by
1714 default unless the new "-noservername" option is used. The server name is
1715 based on the host provided to the "-connect" option unless overridden by
1716 using "-servername".
1720 * Add support for SipHash
1724 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1725 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1726 prevent issues where no progress is being made and the peer continually
1727 sends unrecognised record types, using up resources processing them.
1731 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1732 using the algorithm defined in
1733 https://www.akkadia.org/drepper/SHA-crypt.txt
1737 * Heartbeat support has been removed; the ABI is changed for now.
1739 *Richard Levitte, Rich Salz*
1741 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1745 * The RSA "null" method, which was partially supported to avoid patent
1746 issues, has been replaced to always returns NULL.
1754 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019] ###
1756 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1757 used even when parsing explicit parameters, when loading a serialized key
1758 or calling `EC_GROUP_new_from_ecpkparameters()`/
1759 `EC_GROUP_new_from_ecparameters()`.
1760 This prevents bypass of security hardening and performance gains,
1761 especially for curves with specialized EC_METHODs.
1762 By default, if a key encoded with explicit parameters is loaded and later
1763 serialized, the output is still encoded with explicit parameters, even if
1764 internally a "named" EC_GROUP is used for computation.
1768 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1769 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1770 NULL. After this change, only the cofactor parameter can be NULL. It also
1771 does some minimal sanity checks on the passed order.
1776 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1777 An attack is simple, if the first CMS_recipientInfo is valid but the
1778 second CMS_recipientInfo is chosen ciphertext. If the second
1779 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1780 encryption key will be replaced by garbage, and the message cannot be
1781 decoded, but if the RSA decryption fails, the correct encryption key is
1782 used and the recipient will not notice the attack.
1783 As a work around for this potential attack the length of the decrypted
1784 key must be equal to the cipher default key length, in case the
1785 certifiate is not given and all recipientInfo are tried out.
1786 The old behaviour can be re-enabled in the CMS code by setting the
1787 CMS_DEBUG_DECRYPT flag.
1792 * Use Windows installation paths in the mingw builds
1794 Mingw isn't a POSIX environment per se, which means that Windows
1795 paths should be used for installation.
1800 ### Changes between 1.1.0j and 1.1.0k [28 May 2019] ###
1802 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1803 This changes the size when using the genpkey app when no size is given. It
1804 fixes an omission in earlier changes that changed all RSA, DSA and DH
1805 generation apps to use 2048 bits by default.
1809 * Prevent over long nonces in ChaCha20-Poly1305.
1811 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1812 for every encryption operation. RFC 7539 specifies that the nonce value
1813 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1814 and front pads the nonce with 0 bytes if it is less than 12
1815 bytes. However it also incorrectly allows a nonce to be set of up to 16
1816 bytes. In this case only the last 12 bytes are significant and any
1817 additional leading bytes are ignored.
1819 It is a requirement of using this cipher that nonce values are
1820 unique. Messages encrypted using a reused nonce value are susceptible to
1821 serious confidentiality and integrity attacks. If an application changes
1822 the default nonce length to be longer than 12 bytes and then makes a
1823 change to the leading bytes of the nonce expecting the new value to be a
1824 new unique nonce then such an application could inadvertently encrypt
1825 messages with a reused nonce.
1827 Additionally the ignored bytes in a long nonce are not covered by the
1828 integrity guarantee of this cipher. Any application that relies on the
1829 integrity of these ignored leading bytes of a long nonce may be further
1830 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1831 is safe because no such use sets such a long nonce value. However user
1832 applications that use this cipher directly and set a non-default nonce
1833 length to be longer than 12 bytes may be vulnerable.
1835 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1841 * Added SCA hardening for modular field inversion in EC_GROUP through
1842 a new dedicated field_inv() pointer in EC_METHOD.
1843 This also addresses a leakage affecting conversions from projective
1844 to affine coordinates.
1846 *Billy Bob Brumley, Nicola Tuveri*
1848 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1849 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1853 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1857 * Remove the 'dist' target and add a tarball building script. The
1858 'dist' target has fallen out of use, and it shouldn't be
1859 necessary to configure just to create a source distribution.
1863 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018] ###
1865 * Timing vulnerability in DSA signature generation
1867 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1868 timing side channel attack. An attacker could use variations in the signing
1869 algorithm to recover the private key.
1871 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1876 * Timing vulnerability in ECDSA signature generation
1878 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1879 timing side channel attack. An attacker could use variations in the signing
1880 algorithm to recover the private key.
1882 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1887 * Add coordinate blinding for EC_POINT and implement projective
1888 coordinate blinding for generic prime curves as a countermeasure to
1889 chosen point SCA attacks.
1891 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1893 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018] ###
1895 * Client DoS due to large DH parameter
1897 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
1898 malicious server can send a very large prime value to the client. This will
1899 cause the client to spend an unreasonably long period of time generating a
1900 key for this prime resulting in a hang until the client has finished. This
1901 could be exploited in a Denial Of Service attack.
1903 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
1908 * Cache timing vulnerability in RSA Key Generation
1910 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
1911 a cache timing side channel attack. An attacker with sufficient access to
1912 mount cache timing attacks during the RSA key generation process could
1913 recover the private key.
1915 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
1916 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
1921 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1922 parameter is no longer accepted, as it leads to a corrupt table. NULL
1923 pem_str is reserved for alias entries only.
1927 * Revert blinding in ECDSA sign and instead make problematic addition
1928 length-invariant. Switch even to fixed-length Montgomery multiplication.
1932 * Change generating and checking of primes so that the error rate of not
1933 being prime depends on the intended use based on the size of the input.
1934 For larger primes this will result in more rounds of Miller-Rabin.
1935 The maximal error rate for primes with more than 1080 bits is lowered
1938 *Kurt Roeckx, Annie Yousar*
1940 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1944 * Add blinding to ECDSA and DSA signatures to protect against side channel
1945 attacks discovered by Keegan Ryan (NCC Group).
1949 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1950 now allow empty (zero character) pass phrases.
1954 * Certificate time validation (X509_cmp_time) enforces stricter
1955 compliance with RFC 5280. Fractional seconds and timezone offsets
1956 are no longer allowed.
1960 * Fixed a text canonicalisation bug in CMS
1962 Where a CMS detached signature is used with text content the text goes
1963 through a canonicalisation process first prior to signing or verifying a
1964 signature. This process strips trailing space at the end of lines, converts
1965 line terminators to CRLF and removes additional trailing line terminators
1966 at the end of a file. A bug in the canonicalisation process meant that
1967 some characters, such as form-feed, were incorrectly treated as whitespace
1968 and removed. This is contrary to the specification (RFC5485). This fix
1969 could mean that detached text data signed with an earlier version of
1970 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1971 signed with a fixed OpenSSL may fail to verify with an earlier version of
1972 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1973 and use the "-binary" flag (for the "cms" command line application) or set
1974 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1978 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018] ###
1980 * Constructed ASN.1 types with a recursive definition could exceed the stack
1982 Constructed ASN.1 types with a recursive definition (such as can be found
1983 in PKCS7) could eventually exceed the stack given malicious input with
1984 excessive recursion. This could result in a Denial Of Service attack. There
1985 are no such structures used within SSL/TLS that come from untrusted sources
1986 so this is considered safe.
1988 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1994 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1996 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1997 effectively reduced to only comparing the least significant bit of each
1998 byte. This allows an attacker to forge messages that would be considered as
1999 authenticated in an amount of tries lower than that guaranteed by the
2000 security claims of the scheme. The module can only be compiled by the
2001 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2003 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2009 * Add a build target 'build_all_generated', to build all generated files
2010 and only that. This can be used to prepare everything that requires
2011 things like perl for a system that lacks perl and then move everything
2012 to that system and do the rest of the build there.
2016 * Backport SSL_OP_NO_RENGOTIATION
2018 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2019 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2020 changes this is no longer possible in 1.1.0. Therefore the new
2021 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2022 1.1.0 to provide equivalent functionality.
2024 Note that if an application built against 1.1.0h headers (or above) is run
2025 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2026 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2030 * Removed the OS390-Unix config target. It relied on a script that doesn't
2035 * rsaz_1024_mul_avx2 overflow bug on x86_64
2037 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2038 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2039 Analysis suggests that attacks against RSA and DSA as a result of this
2040 defect would be very difficult to perform and are not believed likely.
2041 Attacks against DH1024 are considered just feasible, because most of the
2042 work necessary to deduce information about a private key may be performed
2043 offline. The amount of resources required for such an attack would be
2044 significant. However, for an attack on TLS to be meaningful, the server
2045 would have to share the DH1024 private key among multiple clients, which is
2046 no longer an option since CVE-2016-0701.
2048 This only affects processors that support the AVX2 but not ADX extensions
2049 like Intel Haswell (4th generation).
2051 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2052 was originally found via the OSS-Fuzz project.
2057 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017] ###
2059 * bn_sqrx8x_internal carry bug on x86_64
2061 There is a carry propagating bug in the x86_64 Montgomery squaring
2062 procedure. No EC algorithms are affected. Analysis suggests that attacks
2063 against RSA and DSA as a result of this defect would be very difficult to
2064 perform and are not believed likely. Attacks against DH are considered just
2065 feasible (although very difficult) because most of the work necessary to
2066 deduce information about a private key may be performed offline. The amount
2067 of resources required for such an attack would be very significant and
2068 likely only accessible to a limited number of attackers. An attacker would
2069 additionally need online access to an unpatched system using the target
2070 private key in a scenario with persistent DH parameters and a private
2071 key that is shared between multiple clients.
2073 This only affects processors that support the BMI1, BMI2 and ADX extensions
2074 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2076 This issue was reported to OpenSSL by the OSS-Fuzz project.
2081 * Malformed X.509 IPAddressFamily could cause OOB read
2083 If an X.509 certificate has a malformed IPAddressFamily extension,
2084 OpenSSL could do a one-byte buffer overread. The most likely result
2085 would be an erroneous display of the certificate in text format.
2087 This issue was reported to OpenSSL by the OSS-Fuzz project.
2092 ### Changes between 1.1.0e and 1.1.0f [25 May 2017] ###
2094 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2095 platform rather than 'mingw'.
2099 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2100 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2101 which is the minimum version we support.
2105 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017] ###
2107 * Encrypt-Then-Mac renegotiation crash
2109 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2110 negotiated where it was not in the original handshake (or vice-versa) then
2111 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2112 and servers are affected.
2114 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2119 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017] ###
2121 * Truncated packet could crash via OOB read
2123 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2124 cipher is being used, then a truncated packet can cause that host to
2125 perform an out-of-bounds read, usually resulting in a crash.
2127 This issue was reported to OpenSSL by Robert Święcki of Google.
2132 * Bad (EC)DHE parameters cause a client crash
2134 If a malicious server supplies bad parameters for a DHE or ECDHE key
2135 exchange then this can result in the client attempting to dereference a
2136 NULL pointer leading to a client crash. This could be exploited in a Denial
2139 This issue was reported to OpenSSL by Guido Vranken.
2144 * BN_mod_exp may produce incorrect results on x86_64
2146 There is a carry propagating bug in the x86_64 Montgomery squaring
2147 procedure. No EC algorithms are affected. Analysis suggests that attacks
2148 against RSA and DSA as a result of this defect would be very difficult to
2149 perform and are not believed likely. Attacks against DH are considered just
2150 feasible (although very difficult) because most of the work necessary to
2151 deduce information about a private key may be performed offline. The amount
2152 of resources required for such an attack would be very significant and
2153 likely only accessible to a limited number of attackers. An attacker would
2154 additionally need online access to an unpatched system using the target
2155 private key in a scenario with persistent DH parameters and a private
2156 key that is shared between multiple clients. For example this can occur by
2157 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2158 similar to CVE-2015-3193 but must be treated as a separate problem.
2160 This issue was reported to OpenSSL by the OSS-Fuzz project.
2165 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016] ###
2167 * ChaCha20/Poly1305 heap-buffer-overflow
2169 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
2170 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2171 crash. This issue is not considered to be exploitable beyond a DoS.
2173 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2178 * CMS Null dereference
2180 Applications parsing invalid CMS structures can crash with a NULL pointer
2181 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2182 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2183 structure callback if an attempt is made to free certain invalid encodings.
2184 Only CHOICE structures using a callback which do not handle NULL value are
2187 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2192 * Montgomery multiplication may produce incorrect results
2194 There is a carry propagating bug in the Broadwell-specific Montgomery
2195 multiplication procedure that handles input lengths divisible by, but
2196 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2197 and DH private keys are impossible. This is because the subroutine in
2198 question is not used in operations with the private key itself and an input
2199 of the attacker's direct choice. Otherwise the bug can manifest itself as
2200 transient authentication and key negotiation failures or reproducible
2201 erroneous outcome of public-key operations with specially crafted input.
2202 Among EC algorithms only Brainpool P-512 curves are affected and one
2203 presumably can attack ECDH key negotiation. Impact was not analyzed in
2204 detail, because pre-requisites for attack are considered unlikely. Namely
2205 multiple clients have to choose the curve in question and the server has to
2206 share the private key among them, neither of which is default behaviour.
2207 Even then only clients that chose the curve will be affected.
2209 This issue was publicly reported as transient failures and was not
2210 initially recognized as a security issue. Thanks to Richard Morgan for
2211 providing reproducible case.
2216 * Removed automatic addition of RPATH in shared libraries and executables,
2217 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2221 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016] ###
2223 * Fix Use After Free for large message sizes
2225 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2226 message larger than approx 16k is received then the underlying buffer to
2227 store the incoming message is reallocated and moved. Unfortunately a
2228 dangling pointer to the old location is left which results in an attempt to
2229 write to the previously freed location. This is likely to result in a
2230 crash, however it could potentially lead to execution of arbitrary code.
2232 This issue only affects OpenSSL 1.1.0a.
2234 This issue was reported to OpenSSL by Robert Święcki.
2239 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016] ###
2241 * OCSP Status Request extension unbounded memory growth
2243 A malicious client can send an excessively large OCSP Status Request
2244 extension. If that client continually requests renegotiation, sending a
2245 large OCSP Status Request extension each time, then there will be unbounded
2246 memory growth on the server. This will eventually lead to a Denial Of
2247 Service attack through memory exhaustion. Servers with a default
2248 configuration are vulnerable even if they do not support OCSP. Builds using
2249 the "no-ocsp" build time option are not affected.
2251 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2256 * SSL_peek() hang on empty record
2258 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2259 sends an empty record. This could be exploited by a malicious peer in a
2260 Denial Of Service attack.
2262 This issue was reported to OpenSSL by Alex Gaynor.
2267 * Excessive allocation of memory in tls_get_message_header() and
2268 dtls1_preprocess_fragment()
2270 A (D)TLS message includes 3 bytes for its length in the header for the
2271 message. This would allow for messages up to 16Mb in length. Messages of
2272 this length are excessive and OpenSSL includes a check to ensure that a
2273 peer is sending reasonably sized messages in order to avoid too much memory
2274 being consumed to service a connection. A flaw in the logic of version
2275 1.1.0 means that memory for the message is allocated too early, prior to
2276 the excessive message length check. Due to way memory is allocated in
2277 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2278 to service a connection. This could lead to a Denial of Service through
2279 memory exhaustion. However, the excessive message length check still takes
2280 place, and this would cause the connection to immediately fail. Assuming
2281 that the application calls SSL_free() on the failed connection in a timely
2282 manner then the 21Mb of allocated memory will then be immediately freed
2283 again. Therefore the excessive memory allocation will be transitory in
2284 nature. This then means that there is only a security impact if:
2286 1) The application does not call SSL_free() in a timely manner in the event
2287 that the connection fails
2289 2) The application is working in a constrained environment where there is
2290 very little free memory
2292 3) The attacker initiates multiple connection attempts such that there are
2293 multiple connections in a state where memory has been allocated for the
2294 connection; SSL_free() has not yet been called; and there is insufficient
2295 memory to service the multiple requests.
2297 Except in the instance of (1) above any Denial Of Service is likely to be
2298 transitory because as soon as the connection fails the memory is
2299 subsequently freed again in the SSL_free() call. However there is an
2300 increased risk during this period of application crashes due to the lack of
2301 memory - which would then mean a more serious Denial of Service.
2303 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2304 (CVE-2016-6307 and CVE-2016-6308)
2308 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2309 had to be removed. Primary reason is that vendor assembler can't
2310 assemble our modules with -KPIC flag. As result it, assembly
2311 support, was not even available as option. But its lack means
2312 lack of side-channel resistant code, which is incompatible with
2313 security by todays standards. Fortunately gcc is readily available
2314 prepackaged option, which we firmly point at...
2318 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016] ###
2320 * Windows command-line tool supports UTF-8 opt-in option for arguments
2321 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2322 (to any value) allows Windows user to access PKCS#12 file generated
2323 with Windows CryptoAPI and protected with non-ASCII password, as well
2324 as files generated under UTF-8 locale on Linux also protected with
2329 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2330 have been disabled by default and removed from DEFAULT, just like RC4.
2331 See the RC4 item below to re-enable both.
2335 * The method for finding the storage location for the Windows RAND seed file
2336 has changed. First we check %RANDFILE%. If that is not set then we check
2337 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2338 all else fails we fall back to C:\.
2342 * The EVP_EncryptUpdate() function has had its return type changed from void
2343 to int. A return of 0 indicates and error while a return of 1 indicates
2348 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2349 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2350 off the constant time implementation for RSA, DSA and DH have been made
2351 no-ops and deprecated.
2355 * Windows RAND implementation was simplified to only get entropy by
2356 calling CryptGenRandom(). Various other RAND-related tickets
2359 *Joseph Wylie Yandle, Rich Salz*
2361 * The stack and lhash API's were renamed to start with OPENSSL_SK_
2362 and OPENSSL_LH_, respectively. The old names are available
2363 with API compatibility. They new names are now completely documented.
2367 * Unify TYPE_up_ref(obj) methods signature.
2368 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2369 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2370 int (instead of void) like all others TYPE_up_ref() methods.
2371 So now these methods also check the return value of CRYPTO_atomic_add(),
2372 and the validity of object reference counter.
2374 *fdasilvayy@gmail.com*
2376 * With Windows Visual Studio builds, the .pdb files are installed
2377 alongside the installed libraries and executables. For a static
2378 library installation, ossl_static.pdb is the associate compiler
2379 generated .pdb file to be used when linking programs.
2383 * Remove openssl.spec. Packaging files belong with the packagers.
2387 * Automatic Darwin/OSX configuration has had a refresh, it will now
2388 recognise x86_64 architectures automatically. You can still decide
2389 to build for a different bitness with the environment variable
2390 KERNEL_BITS (can be 32 or 64), for example:
2392 KERNEL_BITS=32 ./config
2396 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2397 256 bit AES and HMAC with SHA256.
2401 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2405 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2409 * To enable users to have their own config files and build file templates,
2410 Configure looks in the directory indicated by the environment variable
2411 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2412 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2413 name and is used as is.
2417 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2418 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2419 X509_CERT_FILE_CTX was removed.
2423 * "shared" builds are now the default. To create only static libraries use
2424 the "no-shared" Configure option.
2428 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2429 All of these option have not worked for some while and are fundamental
2434 * Make various cleanup routines no-ops and mark them as deprecated. Most
2435 global cleanup functions are no longer required because they are handled
2436 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2437 Explicitly de-initing can cause problems (e.g. where a library that uses
2438 OpenSSL de-inits, but an application is still using it). The affected
2439 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2440 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2441 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2442 COMP_zlib_cleanup().
2446 * --strict-warnings no longer enables runtime debugging options
2447 such as REF_DEBUG. Instead, debug options are automatically
2448 enabled with '--debug' builds.
2450 *Andy Polyakov, Emilia Käsper*
2452 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2453 have been moved out of the public header files. New functions for managing
2454 these have been added.
2458 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2459 objects have been moved out of the public header files. New
2460 functions for managing these have been added.
2464 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2465 have been moved out of the public header files. New functions for managing
2466 these have been added.
2470 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2471 moved out of the public header files. New functions for managing these
2476 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2480 * Removed the mk1mf build scripts.
2484 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2485 it is always safe to #include a header now.
2489 * Removed the aged BC-32 config and all its supporting scripts
2493 * Removed support for Ultrix, Netware, and OS/2.
2497 * Add support for HKDF.
2499 *Alessandro Ghedini*
2501 * Add support for blake2b and blake2s
2505 * Added support for "pipelining". Ciphers that have the
2506 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2507 encryptions/decryptions simultaneously. There are currently no built-in
2508 ciphers with this property but the expectation is that engines will be able
2509 to offer it to significantly improve throughput. Support has been extended
2510 into libssl so that multiple records for a single connection can be
2511 processed in one go (for >=TLS 1.1).
2515 * Added the AFALG engine. This is an async capable engine which is able to
2516 offload work to the Linux kernel. In this initial version it only supports
2517 AES128-CBC. The kernel must be version 4.1.0 or greater.
2521 * OpenSSL now uses a new threading API. It is no longer necessary to
2522 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2523 are two supported threading models: pthreads and windows threads. It is
2524 also possible to configure OpenSSL at compile time for "no-threads". The
2525 old threading API should no longer be used. The functions have been
2526 replaced with "no-op" compatibility macros.
2528 *Alessandro Ghedini, Matt Caswell*
2530 * Modify behavior of ALPN to invoke callback after SNI/servername
2531 callback, such that updates to the SSL_CTX affect ALPN.
2535 * Add SSL_CIPHER queries for authentication and key-exchange.
2539 * Changes to the DEFAULT cipherlist:
2540 - Prefer (EC)DHE handshakes over plain RSA.
2541 - Prefer AEAD ciphers over legacy ciphers.
2542 - Prefer ECDSA over RSA when both certificates are available.
2543 - Prefer TLSv1.2 ciphers/PRF.
2544 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2549 * Change the ECC default curve list to be this, in order: x25519,
2550 secp256r1, secp521r1, secp384r1.
2554 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2555 disabled by default. They can be re-enabled using the
2556 enable-weak-ssl-ciphers option to Configure.
2560 * If the server has ALPN configured, but supports no protocols that the
2561 client advertises, send a fatal "no_application_protocol" alert.
2562 This behaviour is SHALL in RFC 7301, though it isn't universally
2563 implemented by other servers.
2567 * Add X25519 support.
2568 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2569 for public and private key encoding using the format documented in
2570 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2571 key generation and key derivation.
2573 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2578 * Deprecate SRP_VBASE_get_by_user.
2579 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2580 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2581 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2582 seed, even if the seed is configured.
2584 Users should use SRP_VBASE_get1_by_user instead. Note that in
2585 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2586 also that even though configuring the SRP seed attempts to hide
2587 invalid usernames by continuing the handshake with fake
2588 credentials, this behaviour is not constant time and no strong
2589 guarantees are made that the handshake is indistinguishable from
2590 that of a valid user.
2594 * Configuration change; it's now possible to build dynamic engines
2595 without having to build shared libraries and vice versa. This
2596 only applies to the engines in engines/, those in crypto/engine/
2597 will always be built into libcrypto (i.e. "static").
2599 Building dynamic engines is enabled by default; to disable, use
2600 the configuration option "disable-dynamic-engine".
2602 The only requirements for building dynamic engines are the
2603 presence of the DSO module and building with position independent
2604 code, so they will also automatically be disabled if configuring
2605 with "disable-dso" or "disable-pic".
2607 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2608 are also taken away from openssl/opensslconf.h, as they are
2613 * Configuration change; if there is a known flag to compile
2614 position independent code, it will always be applied on the
2615 libcrypto and libssl object files, and never on the application
2616 object files. This means other libraries that use routines from
2617 libcrypto / libssl can be made into shared libraries regardless
2618 of how OpenSSL was configured.
2620 If this isn't desirable, the configuration options "disable-pic"
2621 or "no-pic" can be used to disable the use of PIC. This will
2622 also disable building shared libraries and dynamic engines.
2626 * Removed JPAKE code. It was experimental and has no wide use.
2630 * The INSTALL_PREFIX Makefile variable has been renamed to
2631 DESTDIR. That makes for less confusion on what this variable
2632 is for. Also, the configuration option --install_prefix is
2637 * Heartbeat for TLS has been removed and is disabled by default
2638 for DTLS; configure with enable-heartbeats. Code that uses the
2639 old #define's might need to be updated.
2641 *Emilia Käsper, Rich Salz*
2643 * Rename REF_CHECK to REF_DEBUG.
2647 * New "unified" build system
2649 The "unified" build system is aimed to be a common system for all
2650 platforms we support. With it comes new support for VMS.
2652 This system builds supports building in a different directory tree
2653 than the source tree. It produces one Makefile (for unix family
2654 or lookalikes), or one descrip.mms (for VMS).
2656 The source of information to make the Makefile / descrip.mms is
2657 small files called 'build.info', holding the necessary
2658 information for each directory with source to compile, and a
2659 template in Configurations, like unix-Makefile.tmpl or
2662 With this change, the library names were also renamed on Windows
2663 and on VMS. They now have names that are closer to the standard
2664 on Unix, and include the major version number, and in certain
2665 cases, the architecture they are built for. See "Notes on shared
2666 libraries" in INSTALL.
2668 We rely heavily on the perl module Text::Template.
2672 * Added support for auto-initialisation and de-initialisation of the library.
2673 OpenSSL no longer requires explicit init or deinit routines to be called,
2674 except in certain circumstances. See the OPENSSL_init_crypto() and
2675 OPENSSL_init_ssl() man pages for further information.
2679 * The arguments to the DTLSv1_listen function have changed. Specifically the
2680 "peer" argument is now expected to be a BIO_ADDR object.
2682 * Rewrite of BIO networking library. The BIO library lacked consistent
2683 support of IPv6, and adding it required some more extensive
2684 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2685 which hold all types of addresses and chains of address information.
2686 It also introduces a new API, with functions like BIO_socket,
2687 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2688 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2689 have been adapted accordingly.
2693 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2698 * CRIME protection: disable compression by default, even if OpenSSL is
2699 compiled with zlib enabled. Applications can still enable compression
2700 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2701 using the SSL_CONF library to configure compression.
2705 * The signature of the session callback configured with
2706 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2707 was explicitly marked as 'const unsigned char*' instead of
2712 * Always DPURIFY. Remove the use of uninitialized memory in the
2713 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2717 * Removed many obsolete configuration items, including
2718 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2719 MD2_CHAR, MD2_INT, MD2_LONG
2721 IDEA_SHORT, IDEA_LONG
2722 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2724 *Rich Salz, with advice from Andy Polyakov*
2726 * Many BN internals have been moved to an internal header file.
2728 *Rich Salz with help from Andy Polyakov*
2730 * Configuration and writing out the results from it has changed.
2731 Files such as Makefile include/openssl/opensslconf.h and are now
2732 produced through general templates, such as Makefile.in and
2733 crypto/opensslconf.h.in and some help from the perl module
2736 Also, the center of configuration information is no longer
2737 Makefile. Instead, Configure produces a perl module in
2738 configdata.pm which holds most of the config data (in the hash
2739 table %config), the target data that comes from the target
2740 configuration in one of the Configurations/*.conf files (in
2745 * To clarify their intended purposes, the Configure options
2746 --prefix and --openssldir change their semantics, and become more
2747 straightforward and less interdependent.
2749 --prefix shall be used exclusively to give the location INSTALLTOP
2750 where programs, scripts, libraries, include files and manuals are
2751 going to be installed. The default is now /usr/local.
2753 --openssldir shall be used exclusively to give the default
2754 location OPENSSLDIR where certificates, private keys, CRLs are
2755 managed. This is also where the default openssl.cnf gets
2757 If the directory given with this option is a relative path, the
2758 values of both the --prefix value and the --openssldir value will
2759 be combined to become OPENSSLDIR.
2760 The default for --openssldir is INSTALLTOP/ssl.
2762 Anyone who uses --openssldir to specify where OpenSSL is to be
2763 installed MUST change to use --prefix instead.
2767 * The GOST engine was out of date and therefore it has been removed. An up
2768 to date GOST engine is now being maintained in an external repository.
2769 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
2770 support for GOST ciphersuites (these are only activated if a GOST engine
2775 * EGD is no longer supported by default; use enable-egd when
2778 *Ben Kaduk and Rich Salz*
2780 * The distribution now has Makefile.in files, which are used to
2781 create Makefile's when Configure is run. *Configure must be run
2782 before trying to build now.*
2786 * The return value for SSL_CIPHER_description() for error conditions
2791 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2793 Obtaining and performing DNSSEC validation of TLSA records is
2794 the application's responsibility. The application provides
2795 the TLSA records of its choice to OpenSSL, and these are then
2796 used to authenticate the peer.
2798 The TLSA records need not even come from DNS. They can, for
2799 example, be used to implement local end-entity certificate or
2800 trust-anchor "pinning", where the "pin" data takes the form
2801 of TLSA records, which can augment or replace verification
2802 based on the usual WebPKI public certification authorities.
2806 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
2807 continues to support deprecated interfaces in default builds.
2808 However, applications are strongly advised to compile their
2809 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
2810 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
2811 or the 1.1.0 releases.
2813 In environments in which all applications have been ported to
2814 not use any deprecated interfaces OpenSSL's Configure script
2815 should be used with the --api=1.1.0 option to entirely remove
2816 support for the deprecated features from the library and
2817 unconditionally disable them in the installed headers.
2818 Essentially the same effect can be achieved with the "no-deprecated"
2819 argument to Configure, except that this will always restrict
2820 the build to just the latest API, rather than a fixed API
2823 As applications are ported to future revisions of the API,
2824 they should update their compile-time OPENSSL_API_COMPAT define
2825 accordingly, but in most cases should be able to continue to
2826 compile with later releases.
2828 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
2829 0x10000000L and 0x00908000L, respectively. However those
2830 versions did not support the OPENSSL_API_COMPAT feature, and
2831 so applications are not typically tested for explicit support
2832 of just the undeprecated features of either release.
2836 * Add support for setting the minimum and maximum supported protocol.
2837 It can bet set via the SSL_set_min_proto_version() and
2838 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
2839 MaxProtocol. It's recommended to use the new APIs to disable
2840 protocols instead of disabling individual protocols using
2841 SSL_set_options() or SSL_CONF's Protocol. This change also
2842 removes support for disabling TLS 1.2 in the OpenSSL TLS
2843 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
2847 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
2851 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
2852 and integrates ECDSA and ECDH functionality into EC. Implementations can
2853 now redirect key generation and no longer need to convert to or from
2856 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
2857 include the ec.h header file instead.
2861 * Remove support for all 40 and 56 bit ciphers. This includes all the export
2862 ciphers who are no longer supported and drops support the ephemeral RSA key
2863 exchange. The LOW ciphers currently doesn't have any ciphers in it.
2867 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
2868 opaque. For HMAC_CTX, the following constructors and destructors
2871 HMAC_CTX *HMAC_CTX_new(void);
2872 void HMAC_CTX_free(HMAC_CTX *ctx);
2874 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
2875 destroy such methods has been added. See EVP_MD_meth_new(3) and
2876 EVP_CIPHER_meth_new(3) for documentation.
2879 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
2880 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
2881 EVP_MD_CTX_reset() should be called instead to reinitialise
2882 an already created structure.
2883 2) For consistency with the majority of our object creators and
2884 destructors, EVP_MD_CTX_(create|destroy) were renamed to
2885 EVP_MD_CTX_(new|free). The old names are retained as macros
2886 for deprecated builds.
2890 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
2891 cryptographic operations to be performed asynchronously as long as an
2892 asynchronous capable engine is used. See the ASYNC_start_job() man page for
2893 further details. Libssl has also had this capability integrated with the
2894 introduction of the new mode SSL_MODE_ASYNC and associated error
2895 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
2896 pages. This work was developed in partnership with Intel Corp.
2900 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
2901 always enabled now. If you want to disable the support you should
2902 exclude it using the list of supported ciphers. This also means that the
2903 "-no_ecdhe" option has been removed from s_server.
2907 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
2908 SSL_{CTX_}set1_curves() which can set a list.
2912 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
2913 curve you want to support using SSL_{CTX_}set1_curves().
2917 * State machine rewrite. The state machine code has been significantly
2918 refactored in order to remove much duplication of code and solve issues
2919 with the old code (see ssl/statem/README for further details). This change
2920 does have some associated API changes. Notably the SSL_state() function
2921 has been removed and replaced by SSL_get_state which now returns an
2922 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
2923 altogether. The previous handshake states defined in ssl.h and ssl3.h have
2928 * All instances of the string "ssleay" in the public API were replaced
2929 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
2930 Some error codes related to internal RSA_eay API's were renamed.
2934 * The demo files in crypto/threads were moved to demo/threads.
2938 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
2941 *Matt Caswell, Rich Salz*
2943 * New ASN.1 embed macro.
2945 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
2946 structure is not allocated: it is part of the parent. That is instead of
2954 This reduces memory fragmentation and make it impossible to accidentally
2955 set a mandatory field to NULL.
2957 This currently only works for some fields specifically a SEQUENCE, CHOICE,
2958 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
2959 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
2964 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
2968 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
2969 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
2970 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
2971 DES and RC4 ciphersuites.
2975 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2976 This changes the decoding behaviour for some invalid messages,
2977 though the change is mostly in the more lenient direction, and
2978 legacy behaviour is preserved as much as possible.
2982 * Fix no-stdio build.
2983 * David Woodhouse <David.Woodhouse@intel.com> and also
2984 Ivan Nestlerode <ivan.nestlerode@sonos.com> *
2986 * New testing framework
2987 The testing framework has been largely rewritten and is now using
2988 perl and the perl modules Test::Harness and an extended variant of
2989 Test::More called OpenSSL::Test to do its work. All test scripts in
2990 test/ have been rewritten into test recipes, and all direct calls to
2991 executables in test/Makefile have become individual recipes using the
2992 simplified testing OpenSSL::Test::Simple.
2994 For documentation on our testing modules, do:
2996 perldoc test/testlib/OpenSSL/Test/Simple.pm
2997 perldoc test/testlib/OpenSSL/Test.pm
3001 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3002 are used; the latter aborts on memory leaks (usually checked on exit).
3003 Some undocumented "set malloc, etc., hooks" functions were removed
3004 and others were changed. All are now documented.
3008 * In DSA_generate_parameters_ex, if the provided seed is too short,
3011 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3013 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3014 from RFC4279, RFC4785, RFC5487, RFC5489.
3016 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3017 original RSA_PSK patch.
3021 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3022 era flag was never set throughout the codebase (only read). Also removed
3023 SSL3_FLAGS_POP_BUFFER which was only used if
3024 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3028 * Changed the default name options in the "ca", "crl", "req" and "x509"
3029 to be "oneline" instead of "compat".
3033 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3034 not aware of clients that still exhibit this bug, and the workaround
3035 hasn't been working properly for a while.
3039 * The return type of BIO_number_read() and BIO_number_written() as well as
3040 the corresponding num_read and num_write members in the BIO structure has
3041 changed from unsigned long to uint64_t. On platforms where an unsigned
3042 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3047 * Given the pervasive nature of TLS extensions it is inadvisable to run
3048 OpenSSL without support for them. It also means that maintaining
3049 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3050 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3054 * Removed support for the two export grade static DH ciphersuites
3055 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3056 were newly added (along with a number of other static DH ciphersuites) to
3057 1.0.2. However the two export ones have *never* worked since they were
3058 introduced. It seems strange in any case to be adding new export
3059 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3063 * Version negotiation has been rewritten. In particular SSLv23_method(),
3064 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3065 and turned into macros which simply call the new preferred function names
3066 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3067 should use the new names instead. Also as part of this change the ssl23.h
3068 header file has been removed.
3072 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3073 code and the associated standard is no longer considered fit-for-purpose.
3077 * RT2547 was closed. When generating a private key, try to make the
3078 output file readable only by the owner. This behavior change might
3079 be noticeable when interacting with other software.
3081 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3086 * Added HTTP GET support to the ocsp command.
3090 * Changed default digest for the dgst and enc commands from MD5 to
3095 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3099 * Added support for TLS extended master secret from
3100 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3101 initial patch which was a great help during development.
3105 * All libssl internal structures have been removed from the public header
3106 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3107 now redundant). Users should not attempt to access internal structures
3108 directly. Instead they should use the provided API functions.
3112 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3113 Access to deprecated functions can be re-enabled by running config with
3114 "enable-deprecated". In addition applications wishing to use deprecated
3115 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3116 will, by default, disable some transitive includes that previously existed
3117 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3121 * Added support for OCB mode. OpenSSL has been granted a patent license
3122 compatible with the OpenSSL license for use of OCB. Details are available
3123 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
3124 for OCB can be removed by calling config with no-ocb.
3128 * SSLv2 support has been removed. It still supports receiving a SSLv2
3129 compatible client hello.
3133 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3134 done while fixing the error code for the key-too-small case.
3136 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3138 * CA.sh has been removed; use CA.pl instead.
3142 * Removed old DES API.
3146 * Remove various unsupported platforms:
3152 Sinix/ReliantUNIX RM400
3157 16-bit platforms such as WIN16
3161 * Clean up OPENSSL_NO_xxx #define's
3162 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3163 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3164 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3165 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3166 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3167 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3168 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3169 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3170 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3171 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3175 * Cleaned up dead code
3176 Remove all but one '#ifdef undef' which is to be looked at.
3180 * Clean up calling of xxx_free routines.
3181 Just like free(), fix most of the xxx_free routines to accept
3182 NULL. Remove the non-null checks from callers. Save much code.
3186 * Add secure heap for storage of private keys (when possible).
3187 Add BIO_s_secmem(), CBIGNUM, etc.
3188 Contributed by Akamai Technologies under our Corporate CLA.
3192 * Experimental support for a new, fast, unbiased prime candidate generator,
3193 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3195 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3197 * New output format NSS in the sess_id command line tool. This allows
3198 exporting the session id and the master key in NSS keylog format.
3200 *Martin Kaiser <martin@kaiser.cx>*
3202 * Harmonize version and its documentation. -f flag is used to display
3205 *mancha <mancha1@zoho.com>*
3207 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3208 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3210 *mancha <mancha1@zoho.com>*
3212 * Fix some double frees. These are not thought to be exploitable.
3214 *mancha <mancha1@zoho.com>*
3216 * A missing bounds check in the handling of the TLS heartbeat extension
3217 can be used to reveal up to 64k of memory to a connected client or
3220 Thanks for Neel Mehta of Google Security for discovering this bug and to
3221 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3222 preparing the fix [CVE-2014-0160][]
3224 *Adam Langley, Bodo Moeller*
3226 * Fix for the attack described in the paper "Recovering OpenSSL
3227 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3228 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3229 http://eprint.iacr.org/2014/140
3231 Thanks to Yuval Yarom and Naomi Benger for discovering this
3232 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3234 *Yuval Yarom and Naomi Benger*
3236 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3237 this fixes a limitation in previous versions of OpenSSL.
3241 * Experimental encrypt-then-mac support.
3243 Experimental support for encrypt then mac from
3244 draft-gutmann-tls-encrypt-then-mac-02.txt
3246 To enable it set the appropriate extension number (0x42 for the test
3247 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3249 For non-compliant peers (i.e. just about everything) this should have no
3252 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3257 * Add EVP support for key wrapping algorithms, to avoid problems with
3258 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3259 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3260 algorithms and include tests cases.
3264 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3269 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3270 MGF1 digest and OAEP label.
3274 * Make openssl verify return errors.
3276 *Chris Palmer <palmer@google.com> and Ben Laurie*
3278 * New function ASN1_TIME_diff to calculate the difference between two
3279 ASN1_TIME structures or one structure and the current time.
3283 * Update fips_test_suite to support multiple command line options. New
3284 test to induce all self test errors in sequence and check expected
3289 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3290 sign or verify all in one operation.
3294 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3295 test programs and fips_test_suite. Includes functionality to parse
3296 the minimal script output of fipsalgest.pl directly.
3300 * Add authorisation parameter to FIPS_module_mode_set().
3304 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3308 * Use separate DRBG fields for internal and external flags. New function
3309 FIPS_drbg_health_check() to perform on demand health checking. Add
3310 generation tests to fips_test_suite with reduced health check interval to
3311 demonstrate periodic health checking. Add "nodh" option to
3312 fips_test_suite to skip very slow DH test.
3316 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3321 * More extensive health check for DRBG checking many more failure modes.
3322 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3323 combination: call this in fips_test_suite.
3327 * Add support for canonical generation of DSA parameter 'g'. See
3330 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3331 POST to handle HMAC cases.
3335 * Add functions FIPS_module_version() and FIPS_module_version_text()
3336 to return numerical and string versions of the FIPS module number.
3340 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3341 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3342 outside the validated module in the FIPS capable OpenSSL.
3346 * Minor change to DRBG entropy callback semantics. In some cases
3347 there is no multiple of the block length between min_len and
3348 max_len. Allow the callback to return more than max_len bytes
3349 of entropy but discard any extra: it is the callback's responsibility
3350 to ensure that the extra data discarded does not impact the
3351 requested amount of entropy.
3355 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3356 information in FIPS186-3, SP800-57 and SP800-131A.
3360 * CCM support via EVP. Interface is very similar to GCM case except we
3361 must supply all data in one chunk (i.e. no update, final) and the
3362 message length must be supplied if AAD is used. Add algorithm test
3367 * Initial version of POST overhaul. Add POST callback to allow the status
3368 of POST to be monitored and/or failures induced. Modify fips_test_suite
3369 to use callback. Always run all selftests even if one fails.
3373 * XTS support including algorithm test driver in the fips_gcmtest program.
3374 Note: this does increase the maximum key length from 32 to 64 bytes but
3375 there should be no binary compatibility issues as existing applications
3376 will never use XTS mode.
3380 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3381 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3382 performs algorithm blocking for unapproved PRNG types. Also do not
3383 set PRNG type in FIPS_mode_set(): leave this to the application.
3384 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3385 the standard OpenSSL PRNG: set additional data to a date time vector.
3389 * Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
3390 This shouldn't present any incompatibility problems because applications
3391 shouldn't be using these directly and any that are will need to rethink
3392 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3396 * Extensive self tests and health checking required by SP800-90 DRBG.
3397 Remove strength parameter from FIPS_drbg_instantiate and always
3398 instantiate at maximum supported strength.
3402 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3406 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3410 * New function DH_compute_key_padded() to compute a DH key and pad with
3411 leading zeroes if needed: this complies with SP800-56A et al.
3415 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3416 anything, incomplete, subject to change and largely untested at present.
3420 * Modify fipscanisteronly build option to only build the necessary object
3421 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3425 * Add experimental option FIPSSYMS to give all symbols in
3426 fipscanister.o and FIPS or fips prefix. This will avoid
3427 conflicts with future versions of OpenSSL. Add perl script
3428 util/fipsas.pl to preprocess assembly language source files
3429 and rename any affected symbols.
3433 * Add selftest checks and algorithm block of non-fips algorithms in
3434 FIPS mode. Remove DES2 from selftests.
3438 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3439 return internal method without any ENGINE dependencies. Add new
3440 tiny fips sign and verify functions.
3444 * New build option no-ec2m to disable characteristic 2 code.
3448 * New build option "fipscanisteronly". This only builds fipscanister.o
3449 and (currently) associated fips utilities. Uses the file Makefile.fips
3450 instead of Makefile.org as the prototype.
3454 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3455 Update fips_gcmtest to use IV generator.
3459 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3460 setting output buffer to NULL. The *Final function must be
3461 called although it will not retrieve any additional data. The tag
3462 can be set or retrieved with a ctrl. The IV length is by default 12
3463 bytes (96 bits) but can be set to an alternative value. If the IV
3464 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3469 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3470 underlying do_cipher function handles all cipher semantics itself
3471 including padding and finalisation. This is useful if (for example)
3472 an ENGINE cipher handles block padding itself. The behaviour of
3473 do_cipher is subtly changed if this flag is set: the return value
3474 is the number of characters written to the output buffer (zero is
3475 no longer an error code) or a negative error code. Also if the
3476 input buffer is NULL and length 0 finalisation should be performed.
3480 * If a candidate issuer certificate is already part of the constructed
3481 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3485 * Improve forward-security support: add functions
3487 void SSL_CTX_set_not_resumable_session_callback(
3488 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3489 void SSL_set_not_resumable_session_callback(
3490 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3492 for use by SSL/TLS servers; the callback function will be called whenever a
3493 new session is created, and gets to decide whether the session may be
3494 cached to make it resumable (return 0) or not (return 1). (As by the
3495 SSL/TLS protocol specifications, the session_id sent by the server will be
3496 empty to indicate that the session is not resumable; also, the server will
3497 not generate RFC 4507 (RFC 5077) session tickets.)
3499 A simple reasonable callback implementation is to return is_forward_secure.
3500 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3501 by the SSL/TLS server library, indicating whether it can provide forward
3504 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3506 * New -verify_name option in command line utilities to set verification
3511 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3512 Add CMAC pkey methods.
3516 * Experimental renegotiation in s_server -www mode. If the client
3517 browses /reneg connection is renegotiated. If /renegcert it is
3518 renegotiated requesting a certificate.
3522 * Add an "external" session cache for debugging purposes to s_server. This
3523 should help trace issues which normally are only apparent in deployed
3524 multi-process servers.
3528 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3529 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3530 BIO_set_cipher() and some obscure PEM functions were changed so they
3531 can now return an error. The RAND changes required a change to the
3532 RAND_METHOD structure.
3536 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3537 a gcc attribute to warn if the result of a function is ignored. This
3538 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3539 whose return value is often ignored.
3543 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3544 These allow SCTs (signed certificate timestamps) to be requested and
3545 validated when establishing a connection.
3547 *Rob Percival <robpercival@google.com>*
3552 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019] ###
3554 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3555 used even when parsing explicit parameters, when loading a serialized key
3556 or calling `EC_GROUP_new_from_ecpkparameters()`/
3557 `EC_GROUP_new_from_ecparameters()`.
3558 This prevents bypass of security hardening and performance gains,
3559 especially for curves with specialized EC_METHODs.
3560 By default, if a key encoded with explicit parameters is loaded and later
3561 serialized, the output is still encoded with explicit parameters, even if
3562 internally a "named" EC_GROUP is used for computation.
3566 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3567 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3568 NULL. After this change, only the cofactor parameter can be NULL. It also
3569 does some minimal sanity checks on the passed order.
3574 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3575 An attack is simple, if the first CMS_recipientInfo is valid but the
3576 second CMS_recipientInfo is chosen ciphertext. If the second
3577 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3578 encryption key will be replaced by garbage, and the message cannot be
3579 decoded, but if the RSA decryption fails, the correct encryption key is
3580 used and the recipient will not notice the attack.
3581 As a work around for this potential attack the length of the decrypted
3582 key must be equal to the cipher default key length, in case the
3583 certifiate is not given and all recipientInfo are tried out.
3584 The old behaviour can be re-enabled in the CMS code by setting the
3585 CMS_DEBUG_DECRYPT flag.
3590 * Document issue with installation paths in diverse Windows builds
3592 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3593 binaries and run-time config file.
3598 ### Changes between 1.0.2r and 1.0.2s [28 May 2019] ###
3600 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3601 This changes the size when using the genpkey app when no size is given. It
3602 fixes an omission in earlier changes that changed all RSA, DSA and DH
3603 generation apps to use 2048 bits by default.
3607 * Add FIPS support for Android Arm 64-bit
3609 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3610 Module in Version 2.0.10. For some reason, the corresponding target
3611 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3612 built with FIPS support on Android Arm 64-bit. This omission has been
3615 *Matthias St. Pierre*
3617 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019] ###
3619 * 0-byte record padding oracle
3621 If an application encounters a fatal protocol error and then calls
3622 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3623 then OpenSSL can respond differently to the calling application if a 0 byte
3624 record is received with invalid padding compared to if a 0 byte record is
3625 received with an invalid MAC. If the application then behaves differently
3626 based on that in a way that is detectable to the remote peer, then this
3627 amounts to a padding oracle that could be used to decrypt data.
3629 In order for this to be exploitable "non-stitched" ciphersuites must be in
3630 use. Stitched ciphersuites are optimised implementations of certain
3631 commonly used ciphersuites. Also the application must call SSL_shutdown()
3632 twice even if a protocol error has occurred (applications should not do
3633 this but some do anyway).
3635 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3636 Aviram, with additional investigation by Steven Collison and Andrew
3637 Hourselt. It was reported to OpenSSL on 10th December 2018.
3642 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3646 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018] ###
3648 * Microarchitecture timing vulnerability in ECC scalar multiplication
3650 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3651 shown to be vulnerable to a microarchitecture timing side channel attack.
3652 An attacker with sufficient access to mount local timing attacks during
3653 ECDSA signature generation could recover the private key.
3655 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3656 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3662 * Timing vulnerability in DSA signature generation
3664 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3665 timing side channel attack. An attacker could use variations in the signing
3666 algorithm to recover the private key.
3668 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3673 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3674 Module, accidentally introduced while backporting security fixes from the
3675 development branch and hindering the use of ECC in FIPS mode.
3679 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018] ###
3681 * Client DoS due to large DH parameter
3683 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3684 malicious server can send a very large prime value to the client. This will
3685 cause the client to spend an unreasonably long period of time generating a
3686 key for this prime resulting in a hang until the client has finished. This
3687 could be exploited in a Denial Of Service attack.
3689 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3694 * Cache timing vulnerability in RSA Key Generation
3696 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3697 a cache timing side channel attack. An attacker with sufficient access to
3698 mount cache timing attacks during the RSA key generation process could
3699 recover the private key.
3701 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3702 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3707 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3708 parameter is no longer accepted, as it leads to a corrupt table. NULL
3709 pem_str is reserved for alias entries only.
3713 * Revert blinding in ECDSA sign and instead make problematic addition
3714 length-invariant. Switch even to fixed-length Montgomery multiplication.
3718 * Change generating and checking of primes so that the error rate of not
3719 being prime depends on the intended use based on the size of the input.
3720 For larger primes this will result in more rounds of Miller-Rabin.
3721 The maximal error rate for primes with more than 1080 bits is lowered
3724 *Kurt Roeckx, Annie Yousar*
3726 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3730 * Add blinding to ECDSA and DSA signatures to protect against side channel
3731 attacks discovered by Keegan Ryan (NCC Group).
3735 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3736 now allow empty (zero character) pass phrases.
3740 * Certificate time validation (X509_cmp_time) enforces stricter
3741 compliance with RFC 5280. Fractional seconds and timezone offsets
3742 are no longer allowed.
3746 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018] ###
3748 * Constructed ASN.1 types with a recursive definition could exceed the stack
3750 Constructed ASN.1 types with a recursive definition (such as can be found
3751 in PKCS7) could eventually exceed the stack given malicious input with
3752 excessive recursion. This could result in a Denial Of Service attack. There
3753 are no such structures used within SSL/TLS that come from untrusted sources
3754 so this is considered safe.
3756 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3762 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017] ###
3764 * Read/write after SSL object in error state
3766 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3767 mechanism. The intent was that if a fatal error occurred during a handshake
3768 then OpenSSL would move into the error state and would immediately fail if
3769 you attempted to continue the handshake. This works as designed for the
3770 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3771 SSL_connect()), however due to a bug it does not work correctly if
3772 SSL_read() or SSL_write() is called directly. In that scenario, if the
3773 handshake fails then a fatal error will be returned in the initial function
3774 call. If SSL_read()/SSL_write() is subsequently called by the application
3775 for the same SSL object then it will succeed and the data is passed without
3776 being decrypted/encrypted directly from the SSL/TLS record layer.
3778 In order to exploit this issue an application bug would have to be present
3779 that resulted in a call to SSL_read()/SSL_write() being issued after having
3780 already received a fatal error.
3782 This issue was reported to OpenSSL by David Benjamin (Google).
3787 * rsaz_1024_mul_avx2 overflow bug on x86_64
3789 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3790 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3791 Analysis suggests that attacks against RSA and DSA as a result of this
3792 defect would be very difficult to perform and are not believed likely.
3793 Attacks against DH1024 are considered just feasible, because most of the
3794 work necessary to deduce information about a private key may be performed
3795 offline. The amount of resources required for such an attack would be
3796 significant. However, for an attack on TLS to be meaningful, the server
3797 would have to share the DH1024 private key among multiple clients, which is
3798 no longer an option since CVE-2016-0701.
3800 This only affects processors that support the AVX2 but not ADX extensions
3801 like Intel Haswell (4th generation).
3803 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3804 was originally found via the OSS-Fuzz project.
3809 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017] ###
3811 * bn_sqrx8x_internal carry bug on x86_64
3813 There is a carry propagating bug in the x86_64 Montgomery squaring
3814 procedure. No EC algorithms are affected. Analysis suggests that attacks
3815 against RSA and DSA as a result of this defect would be very difficult to
3816 perform and are not believed likely. Attacks against DH are considered just
3817 feasible (although very difficult) because most of the work necessary to
3818 deduce information about a private key may be performed offline. The amount
3819 of resources required for such an attack would be very significant and
3820 likely only accessible to a limited number of attackers. An attacker would
3821 additionally need online access to an unpatched system using the target
3822 private key in a scenario with persistent DH parameters and a private
3823 key that is shared between multiple clients.
3825 This only affects processors that support the BMI1, BMI2 and ADX extensions
3826 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3828 This issue was reported to OpenSSL by the OSS-Fuzz project.
3833 * Malformed X.509 IPAddressFamily could cause OOB read
3835 If an X.509 certificate has a malformed IPAddressFamily extension,
3836 OpenSSL could do a one-byte buffer overread. The most likely result
3837 would be an erroneous display of the certificate in text format.
3839 This issue was reported to OpenSSL by the OSS-Fuzz project.
3844 ### Changes between 1.0.2k and 1.0.2l [25 May 2017] ###
3846 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3847 platform rather than 'mingw'.
3851 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017] ###
3853 * Truncated packet could crash via OOB read
3855 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3856 cipher is being used, then a truncated packet can cause that host to
3857 perform an out-of-bounds read, usually resulting in a crash.
3859 This issue was reported to OpenSSL by Robert Święcki of Google.
3864 * BN_mod_exp may produce incorrect results on x86_64
3866 There is a carry propagating bug in the x86_64 Montgomery squaring
3867 procedure. No EC algorithms are affected. Analysis suggests that attacks
3868 against RSA and DSA as a result of this defect would be very difficult to
3869 perform and are not believed likely. Attacks against DH are considered just
3870 feasible (although very difficult) because most of the work necessary to
3871 deduce information about a private key may be performed offline. The amount
3872 of resources required for such an attack would be very significant and
3873 likely only accessible to a limited number of attackers. An attacker would
3874 additionally need online access to an unpatched system using the target
3875 private key in a scenario with persistent DH parameters and a private
3876 key that is shared between multiple clients. For example this can occur by
3877 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3878 similar to CVE-2015-3193 but must be treated as a separate problem.
3880 This issue was reported to OpenSSL by the OSS-Fuzz project.
3885 * Montgomery multiplication may produce incorrect results
3887 There is a carry propagating bug in the Broadwell-specific Montgomery
3888 multiplication procedure that handles input lengths divisible by, but
3889 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3890 and DH private keys are impossible. This is because the subroutine in
3891 question is not used in operations with the private key itself and an input
3892 of the attacker's direct choice. Otherwise the bug can manifest itself as
3893 transient authentication and key negotiation failures or reproducible
3894 erroneous outcome of public-key operations with specially crafted input.
3895 Among EC algorithms only Brainpool P-512 curves are affected and one
3896 presumably can attack ECDH key negotiation. Impact was not analyzed in
3897 detail, because pre-requisites for attack are considered unlikely. Namely
3898 multiple clients have to choose the curve in question and the server has to
3899 share the private key among them, neither of which is default behaviour.
3900 Even then only clients that chose the curve will be affected.
3902 This issue was publicly reported as transient failures and was not
3903 initially recognized as a security issue. Thanks to Richard Morgan for
3904 providing reproducible case.
3909 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3910 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3911 prevent issues where no progress is being made and the peer continually
3912 sends unrecognised record types, using up resources processing them.
3916 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016] ###
3918 * Missing CRL sanity check
3920 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
3921 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
3922 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
3924 This issue only affects the OpenSSL 1.0.2i
3929 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016] ###
3931 * OCSP Status Request extension unbounded memory growth
3933 A malicious client can send an excessively large OCSP Status Request
3934 extension. If that client continually requests renegotiation, sending a
3935 large OCSP Status Request extension each time, then there will be unbounded
3936 memory growth on the server. This will eventually lead to a Denial Of
3937 Service attack through memory exhaustion. Servers with a default
3938 configuration are vulnerable even if they do not support OCSP. Builds using
3939 the "no-ocsp" build time option are not affected.
3941 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3946 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
3949 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
3955 * OOB write in MDC2_Update()
3957 An overflow can occur in MDC2_Update() either if called directly or
3958 through the EVP_DigestUpdate() function using MDC2. If an attacker
3959 is able to supply very large amounts of input data after a previous
3960 call to EVP_EncryptUpdate() with a partial block then a length check
3961 can overflow resulting in a heap corruption.
3963 The amount of data needed is comparable to SIZE_MAX which is impractical
3966 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3971 * Malformed SHA512 ticket DoS
3973 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
3974 DoS attack where a malformed ticket will result in an OOB read which will
3977 The use of SHA512 in TLS session tickets is comparatively rare as it requires
3978 a custom server callback and ticket lookup mechanism.
3980 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3985 * OOB write in BN_bn2dec()
3987 The function BN_bn2dec() does not check the return value of BN_div_word().
3988 This can cause an OOB write if an application uses this function with an
3989 overly large BIGNUM. This could be a problem if an overly large certificate
3990 or CRL is printed out from an untrusted source. TLS is not affected because
3991 record limits will reject an oversized certificate before it is parsed.
3993 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3998 * OOB read in TS_OBJ_print_bio()
4000 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4001 the total length the OID text representation would use and not the amount
4002 of data written. This will result in OOB reads when large OIDs are
4005 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4010 * Pointer arithmetic undefined behaviour
4012 Avoid some undefined pointer arithmetic
4014 A common idiom in the codebase is to check limits in the following manner:
4017 Where "p" points to some malloc'd data of SIZE bytes and
4020 "len" here could be from some externally supplied data (e.g. from a TLS
4023 The rules of C pointer arithmetic are such that "p + len" is only well
4024 defined where len <= SIZE. Therefore the above idiom is actually
4025 undefined behaviour.
4027 For example this could cause problems if some malloc implementation
4028 provides an address for "p" such that "p + len" actually overflows for
4029 values of len that are too big and therefore p + len < limit.
4031 This issue was reported to OpenSSL by Guido Vranken
4036 * Constant time flag not preserved in DSA signing
4038 Operations in the DSA signing algorithm should run in constant time in
4039 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4040 implementation means that a non-constant time codepath is followed for
4041 certain operations. This has been demonstrated through a cache-timing
4042 attack to be sufficient for an attacker to recover the private DSA key.
4044 This issue was reported by César Pereida (Aalto University), Billy Brumley
4045 (Tampere University of Technology), and Yuval Yarom (The University of
4046 Adelaide and NICTA).
4051 * DTLS buffered message DoS
4053 In a DTLS connection where handshake messages are delivered out-of-order
4054 those messages that OpenSSL is not yet ready to process will be buffered
4055 for later use. Under certain circumstances, a flaw in the logic means that
4056 those messages do not get removed from the buffer even though the handshake
4057 has been completed. An attacker could force up to approx. 15 messages to
4058 remain in the buffer when they are no longer required. These messages will
4059 be cleared when the DTLS connection is closed. The default maximum size for
4060 a message is 100k. Therefore the attacker could force an additional 1500k
4061 to be consumed per connection. By opening many simulataneous connections an
4062 attacker could cause a DoS attack through memory exhaustion.
4064 This issue was reported to OpenSSL by Quan Luo.
4069 * DTLS replay protection DoS
4071 A flaw in the DTLS replay attack protection mechanism means that records
4072 that arrive for future epochs update the replay protection "window" before
4073 the MAC for the record has been validated. This could be exploited by an
4074 attacker by sending a record for the next epoch (which does not have to
4075 decrypt or have a valid MAC), with a very large sequence number. This means
4076 that all subsequent legitimate packets are dropped causing a denial of
4077 service for a specific DTLS connection.
4079 This issue was reported to OpenSSL by the OCAP audit team.
4084 * Certificate message OOB reads
4086 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4087 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4088 theoretical DoS risk but this has not been observed in practice on common
4091 The messages affected are client certificate, client certificate request
4092 and server certificate. As a result the attack can only be performed
4093 against a client or a server which enables client authentication.
4095 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4100 ### Changes between 1.0.2g and 1.0.2h [3 May 2016] ###
4102 * Prevent padding oracle in AES-NI CBC MAC check
4104 A MITM attacker can use a padding oracle attack to decrypt traffic
4105 when the connection uses an AES CBC cipher and the server support
4108 This issue was introduced as part of the fix for Lucky 13 padding
4109 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4110 constant time by making sure that always the same bytes are read and
4111 compared against either the MAC or padding bytes. But it no longer
4112 checked that there was enough data to have both the MAC and padding
4115 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4120 * Fix EVP_EncodeUpdate overflow
4122 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4123 Base64 encoding of binary data. If an attacker is able to supply very large
4124 amounts of input data then a length check can overflow resulting in a heap
4127 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4128 the PEM_write_bio* family of functions. These are mainly used within the
4129 OpenSSL command line applications, so any application which processes data
4130 from an untrusted source and outputs it as a PEM file should be considered
4131 vulnerable to this issue. User applications that call these APIs directly
4132 with large amounts of untrusted data may also be vulnerable.
4134 This issue was reported by Guido Vranken.
4139 * Fix EVP_EncryptUpdate overflow
4141 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4142 is able to supply very large amounts of input data after a previous call to
4143 EVP_EncryptUpdate() with a partial block then a length check can overflow
4144 resulting in a heap corruption. Following an analysis of all OpenSSL
4145 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4146 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4147 the first called function after an EVP_EncryptInit(), and therefore that
4148 specific call must be safe. The second form is where the length passed to
4149 EVP_EncryptUpdate() can be seen from the code to be some small value and
4150 therefore there is no possibility of an overflow. Since all instances are
4151 one of these two forms, it is believed that there can be no overflows in
4152 internal code due to this problem. It should be noted that
4153 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4154 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4155 of these calls have also been analysed too and it is believed there are no
4156 instances in internal usage where an overflow could occur.
4158 This issue was reported by Guido Vranken.
4163 * Prevent ASN.1 BIO excessive memory allocation
4165 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4166 a short invalid encoding can cause allocation of large amounts of memory
4167 potentially consuming excessive resources or exhausting memory.
4169 Any application parsing untrusted data through d2i BIO functions is
4170 affected. The memory based functions such as d2i_X509() are *not* affected.
4171 Since the memory based functions are used by the TLS library, TLS
4172 applications are not affected.
4174 This issue was reported by Brian Carpenter.
4181 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4182 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4183 in arbitrary stack data being returned in the buffer.
4185 This issue was reported by Guido Vranken.
4190 * Modify behavior of ALPN to invoke callback after SNI/servername
4191 callback, such that updates to the SSL_CTX affect ALPN.
4195 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4200 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4201 methods are enabled and ssl2 is disabled the methods return NULL.
4205 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016] ###
4207 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4208 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4209 provide any "EXPORT" or "LOW" strength ciphers.
4213 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4214 is by default disabled at build-time. Builds that are not configured with
4215 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4216 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4217 will need to explicitly call either of:
4219 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4221 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4223 as appropriate. Even if either of those is used, or the application
4224 explicitly uses the version-specific SSLv2_method() or its client and
4225 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4226 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4227 ciphers, and SSLv2 56-bit DES are no longer available.
4232 * Fix a double-free in DSA code
4234 A double free bug was discovered when OpenSSL parses malformed DSA private
4235 keys and could lead to a DoS attack or memory corruption for applications
4236 that receive DSA private keys from untrusted sources. This scenario is
4239 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4245 * Disable SRP fake user seed to address a server memory leak.
4247 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4249 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4250 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4251 was changed to ignore the "fake user" SRP seed, even if the seed
4254 Users should use SRP_VBASE_get1_by_user instead. Note that in
4255 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4256 also that even though configuring the SRP seed attempts to hide
4257 invalid usernames by continuing the handshake with fake
4258 credentials, this behaviour is not constant time and no strong
4259 guarantees are made that the handshake is indistinguishable from
4260 that of a valid user.
4265 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4267 In the BN_hex2bn function the number of hex digits is calculated using an
4268 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
4269 large values of |i| this can result in |bn_expand| not allocating any
4270 memory because |i * 4| is negative. This can leave the internal BIGNUM data
4271 field as NULL leading to a subsequent NULL ptr deref. For very large values
4272 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
4273 In this case memory is allocated to the internal BIGNUM data field, but it
4274 is insufficiently sized leading to heap corruption. A similar issue exists
4275 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4276 is ever called by user applications with very large untrusted hex/dec data.
4277 This is anticipated to be a rare occurrence.
4279 All OpenSSL internal usage of these functions use data that is not expected
4280 to be untrusted, e.g. config file data or application command line
4281 arguments. If user developed applications generate config file data based
4282 on untrusted data then it is possible that this could also lead to security
4283 consequences. This is also anticipated to be rare.
4285 This issue was reported to OpenSSL by Guido Vranken.
4290 * Fix memory issues in BIO_*printf functions
4292 The internal |fmtstr| function used in processing a "%s" format string in
4293 the BIO_*printf functions could overflow while calculating the length of a
4294 string and cause an OOB read when printing very long strings.
4296 Additionally the internal |doapr_outch| function can attempt to write to an
4297 OOB memory location (at an offset from the NULL pointer) in the event of a
4298 memory allocation failure. In 1.0.2 and below this could be caused where
4299 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4300 could be in processing a very long "%s" format string. Memory leaks can
4303 The first issue may mask the second issue dependent on compiler behaviour.
4304 These problems could enable attacks where large amounts of untrusted data
4305 is passed to the BIO_*printf functions. If applications use these functions
4306 in this way then they could be vulnerable. OpenSSL itself uses these
4307 functions when printing out human-readable dumps of ASN.1 data. Therefore
4308 applications that print this data could be vulnerable if the data is from
4309 untrusted sources. OpenSSL command line applications could also be
4310 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4311 as command line arguments.
4313 Libssl is not considered directly vulnerable. Additionally certificates etc
4314 received via remote connections via libssl are also unlikely to be able to
4315 trigger these issues because of message size limits enforced within libssl.
4317 This issue was reported to OpenSSL Guido Vranken.
4322 * Side channel attack on modular exponentiation
4324 A side-channel attack was found which makes use of cache-bank conflicts on
4325 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4326 of RSA keys. The ability to exploit this issue is limited as it relies on
4327 an attacker who has control of code in a thread running on the same
4328 hyper-threaded core as the victim thread which is performing decryptions.
4330 This issue was reported to OpenSSL by Yuval Yarom, The University of
4331 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4332 Nadia Heninger, University of Pennsylvania with more information at
4333 http://cachebleed.info.
4338 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4339 if no keysize is specified with default_bits. This fixes an
4340 omission in an earlier change that changed all RSA/DSA key generation
4341 apps to use 2048 bits by default.
4345 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016] ###
4346 * DH small subgroups
4348 Historically OpenSSL only ever generated DH parameters based on "safe"
4349 primes. More recently (in version 1.0.2) support was provided for
4350 generating X9.42 style parameter files such as those required for RFC 5114
4351 support. The primes used in such files may not be "safe". Where an
4352 application is using DH configured with parameters based on primes that are
4353 not "safe" then an attacker could use this fact to find a peer's private
4354 DH exponent. This attack requires that the attacker complete multiple
4355 handshakes in which the peer uses the same private DH exponent. For example
4356 this could be used to discover a TLS server's private DH exponent if it's
4357 reusing the private DH exponent or it's using a static DH ciphersuite.
4359 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4360 TLS. It is not on by default. If the option is not set then the server
4361 reuses the same private DH exponent for the life of the server process and
4362 would be vulnerable to this attack. It is believed that many popular
4363 applications do set this option and would therefore not be at risk.
4365 The fix for this issue adds an additional check where a "q" parameter is
4366 available (as is the case in X9.42 based parameters). This detects the
4367 only known attack, and is the only possible defense for static DH
4368 ciphersuites. This could have some performance impact.
4370 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4371 default and cannot be disabled. This could have some performance impact.
4373 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4378 * SSLv2 doesn't block disabled ciphers
4380 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4381 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4382 been disabled, provided that the SSLv2 protocol was not also disabled via
4385 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4386 and Sebastian Schinzel.
4391 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015] ###
4393 * BN_mod_exp may produce incorrect results on x86_64
4395 There is a carry propagating bug in the x86_64 Montgomery squaring
4396 procedure. No EC algorithms are affected. Analysis suggests that attacks
4397 against RSA and DSA as a result of this defect would be very difficult to
4398 perform and are not believed likely. Attacks against DH are considered just
4399 feasible (although very difficult) because most of the work necessary to
4400 deduce information about a private key may be performed offline. The amount
4401 of resources required for such an attack would be very significant and
4402 likely only accessible to a limited number of attackers. An attacker would
4403 additionally need online access to an unpatched system using the target
4404 private key in a scenario with persistent DH parameters and a private
4405 key that is shared between multiple clients. For example this can occur by
4406 default in OpenSSL DHE based SSL/TLS ciphersuites.
4408 This issue was reported to OpenSSL by Hanno Böck.
4413 * Certificate verify crash with missing PSS parameter
4415 The signature verification routines will crash with a NULL pointer
4416 dereference if presented with an ASN.1 signature using the RSA PSS
4417 algorithm and absent mask generation function parameter. Since these
4418 routines are used to verify certificate signature algorithms this can be
4419 used to crash any certificate verification operation and exploited in a
4420 DoS attack. Any application which performs certificate verification is
4421 vulnerable including OpenSSL clients and servers which enable client
4424 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4429 * X509_ATTRIBUTE memory leak
4431 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4432 memory. This structure is used by the PKCS#7 and CMS routines so any
4433 application which reads PKCS#7 or CMS data from untrusted sources is
4434 affected. SSL/TLS is not affected.
4436 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4442 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4443 This changes the decoding behaviour for some invalid messages,
4444 though the change is mostly in the more lenient direction, and
4445 legacy behaviour is preserved as much as possible.
4449 * In DSA_generate_parameters_ex, if the provided seed is too short,
4452 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4454 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015] ###
4456 * Alternate chains certificate forgery
4458 During certificate verification, OpenSSL will attempt to find an
4459 alternative certificate chain if the first attempt to build such a chain
4460 fails. An error in the implementation of this logic can mean that an
4461 attacker could cause certain checks on untrusted certificates to be
4462 bypassed, such as the CA flag, enabling them to use a valid leaf
4463 certificate to act as a CA and "issue" an invalid certificate.
4465 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4470 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015] ###
4472 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4473 incompatibility in the handling of HMAC. The previous ABI has now been
4478 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015] ###
4480 * Malformed ECParameters causes infinite loop
4482 When processing an ECParameters structure OpenSSL enters an infinite loop
4483 if the curve specified is over a specially malformed binary polynomial
4486 This can be used to perform denial of service against any
4487 system which processes public keys, certificate requests or
4488 certificates. This includes TLS clients and TLS servers with
4489 client authentication enabled.
4491 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4496 * Exploitable out-of-bounds read in X509_cmp_time
4498 X509_cmp_time does not properly check the length of the ASN1_TIME
4499 string and can read a few bytes out of bounds. In addition,
4500 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4503 An attacker can use this to craft malformed certificates and CRLs of
4504 various sizes and potentially cause a segmentation fault, resulting in
4505 a DoS on applications that verify certificates or CRLs. TLS clients
4506 that verify CRLs are affected. TLS clients and servers with client
4507 authentication enabled may be affected if they use custom verification
4510 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4511 independently by Hanno Böck.
4516 * PKCS7 crash with missing EnvelopedContent
4518 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4519 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4520 with missing content and trigger a NULL pointer dereference on parsing.
4522 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4523 structures from untrusted sources are affected. OpenSSL clients and
4524 servers are not affected.
4526 This issue was reported to OpenSSL by Michal Zalewski (Google).
4531 * CMS verify infinite loop with unknown hash function
4533 When verifying a signedData message the CMS code can enter an infinite loop
4534 if presented with an unknown hash function OID. This can be used to perform
4535 denial of service against any system which verifies signedData messages using
4537 This issue was reported to OpenSSL by Johannes Bauer.
4542 * Race condition handling NewSessionTicket
4544 If a NewSessionTicket is received by a multi-threaded client when attempting to
4545 reuse a previous ticket then a race condition can occur potentially leading to
4546 a double free of the ticket data.
4551 * Only support 256-bit or stronger elliptic curves with the
4552 'ecdh_auto' setting (server) or by default (client). Of supported
4553 curves, prefer P-256 (both).
4557 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015] ###
4559 * ClientHello sigalgs DoS fix
4561 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4562 invalid signature algorithms extension a NULL pointer dereference will
4563 occur. This can be exploited in a DoS attack against the server.
4565 This issue was was reported to OpenSSL by David Ramos of Stanford
4569 *Stephen Henson and Matt Caswell*
4571 * Multiblock corrupted pointer fix
4573 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4574 feature only applies on 64 bit x86 architecture platforms that support AES
4575 NI instructions. A defect in the implementation of "multiblock" can cause
4576 OpenSSL's internal write buffer to become incorrectly set to NULL when
4577 using non-blocking IO. Typically, when the user application is using a
4578 socket BIO for writing, this will only result in a failed connection.
4579 However if some other BIO is used then it is likely that a segmentation
4580 fault will be triggered, thus enabling a potential DoS attack.
4582 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4587 * Segmentation fault in DTLSv1_listen fix
4589 The DTLSv1_listen function is intended to be stateless and processes the
4590 initial ClientHello from many peers. It is common for user code to loop
4591 over the call to DTLSv1_listen until a valid ClientHello is received with
4592 an associated cookie. A defect in the implementation of DTLSv1_listen means
4593 that state is preserved in the SSL object from one invocation to the next
4594 that can lead to a segmentation fault. Errors processing the initial
4595 ClientHello can trigger this scenario. An example of such an error could be
4596 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4599 This issue was reported to OpenSSL by Per Allansson.
4604 * Segmentation fault in ASN1_TYPE_cmp fix
4606 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4607 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4608 certificate signature algorithm consistency this can be used to crash any
4609 certificate verification operation and exploited in a DoS attack. Any
4610 application which performs certificate verification is vulnerable including
4611 OpenSSL clients and servers which enable client authentication.
4616 * Segmentation fault for invalid PSS parameters fix
4618 The signature verification routines will crash with a NULL pointer
4619 dereference if presented with an ASN.1 signature using the RSA PSS
4620 algorithm and invalid parameters. Since these routines are used to verify
4621 certificate signature algorithms this can be used to crash any
4622 certificate verification operation and exploited in a DoS attack. Any
4623 application which performs certificate verification is vulnerable including
4624 OpenSSL clients and servers which enable client authentication.
4626 This issue was was reported to OpenSSL by Brian Carpenter.
4631 * ASN.1 structure reuse memory corruption fix
4633 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4634 memory corruption via an invalid write. Such reuse is and has been
4635 strongly discouraged and is believed to be rare.
4637 Applications that parse structures containing CHOICE or ANY DEFINED BY
4638 components may be affected. Certificate parsing (d2i_X509 and related
4639 functions) are however not affected. OpenSSL clients and servers are
4645 * PKCS7 NULL pointer dereferences fix
4647 The PKCS#7 parsing code does not handle missing outer ContentInfo
4648 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4649 missing content and trigger a NULL pointer dereference on parsing.
4651 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4652 otherwise parse PKCS#7 structures from untrusted sources are
4653 affected. OpenSSL clients and servers are not affected.
4655 This issue was reported to OpenSSL by Michal Zalewski (Google).
4660 * DoS via reachable assert in SSLv2 servers fix
4662 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4663 servers that both support SSLv2 and enable export cipher suites by sending
4664 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4666 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4667 (OpenSSL development team).
4672 * Empty CKE with client auth and DHE fix
4674 If client auth is used then a server can seg fault in the event of a DHE
4675 ciphersuite being selected and a zero length ClientKeyExchange message
4676 being sent by the client. This could be exploited in a DoS attack.
4681 * Handshake with unseeded PRNG fix
4683 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4684 with an unseeded PRNG. The conditions are:
4685 - The client is on a platform where the PRNG has not been seeded
4686 automatically, and the user has not seeded manually
4687 - A protocol specific client method version has been used (i.e. not
4688 SSL_client_methodv23)
4689 - A ciphersuite is used that does not require additional random data from
4690 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4692 If the handshake succeeds then the client random that has been used will
4693 have been generated from a PRNG with insufficient entropy and therefore the
4694 output may be predictable.
4696 For example using the following command with an unseeded openssl will
4697 succeed on an unpatched platform:
4699 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4704 * Use After Free following d2i_ECPrivatekey error fix
4706 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4707 could cause a use after free condition. This, in turn, could cause a double
4708 free in several private key parsing functions (such as d2i_PrivateKey
4709 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4710 for applications that receive EC private keys from untrusted
4711 sources. This scenario is considered rare.
4713 This issue was discovered by the BoringSSL project and fixed in their
4719 * X509_to_X509_REQ NULL pointer deref fix
4721 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4722 the certificate key is invalid. This function is rarely used in practice.
4724 This issue was discovered by Brian Carpenter.
4729 * Removed the export ciphers from the DEFAULT ciphers
4733 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015] ###
4735 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4736 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4737 So far those who have to target multiple platforms would compromise
4738 and argue that binary targeting say ARMv5 would still execute on
4739 ARMv8. "Universal" build resolves this compromise by providing
4740 near-optimal performance even on newer platforms.
4744 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4745 (other platforms pending).
4747 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4749 * Add support for the SignedCertificateTimestampList certificate and
4750 OCSP response extensions from RFC6962.
4754 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4755 for corner cases. (Certain input points at infinity could lead to
4756 bogus results, with non-infinity inputs mapped to infinity too.)
4760 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4761 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4762 common cases are optimized and there still is room for further
4763 improvements. Vector Permutation AES for Altivec is also added.
4767 * Add support for little-endian ppc64 Linux target.
4769 *Marcelo Cerri (IBM)*
4771 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4772 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4773 are optimized and there still is room for further improvements.
4774 Both 32- and 64-bit modes are supported.
4776 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4778 * Improved ARMv7 NEON support.
4782 * Support for SPARC Architecture 2011 crypto extensions, first
4783 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4784 SHA256/512, MD5, GHASH and modular exponentiation.
4786 *Andy Polyakov, David Miller*
4788 * Accelerated modular exponentiation for Intel processors, a.k.a.
4791 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4793 * Support for new and upcoming Intel processors, including AVX2,
4794 BMI and SHA ISA extensions. This includes additional "stitched"
4795 implementations, AESNI-SHA256 and GCM, and multi-buffer support
4798 This work was sponsored by Intel Corp.
4802 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
4803 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
4804 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
4808 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4809 this fixes a limitation in previous versions of OpenSSL.
4813 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4814 MGF1 digest and OAEP label.
4818 * Add EVP support for key wrapping algorithms, to avoid problems with
4819 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4820 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4821 algorithms and include tests cases.
4825 * Add functions to allocate and set the fields of an ECDSA_METHOD
4828 *Douglas E. Engert, Steve Henson*
4830 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
4831 difference in days and seconds between two tm or ASN1_TIME structures.
4835 * Add -rev test option to s_server to just reverse order of characters
4836 received by client and send back to server. Also prints an abbreviated
4837 summary of the connection parameters.
4841 * New option -brief for s_client and s_server to print out a brief summary
4842 of connection parameters.
4846 * Add callbacks for arbitrary TLS extensions.
4848 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
4850 * New option -crl_download in several openssl utilities to download CRLs
4851 from CRLDP extension in certificates.
4855 * New options -CRL and -CRLform for s_client and s_server for CRLs.
4859 * New function X509_CRL_diff to generate a delta CRL from the difference
4860 of two full CRLs. Add support to "crl" utility.
4864 * New functions to set lookup_crls function and to retrieve
4865 X509_STORE from X509_STORE_CTX.
4869 * Print out deprecated issuer and subject unique ID fields in
4874 * Extend OCSP I/O functions so they can be used for simple general purpose
4875 HTTP as well as OCSP. New wrapper function which can be used to download
4876 CRLs using the OCSP API.
4880 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
4884 * SSL_CONF* functions. These provide a common framework for application
4885 configuration using configuration files or command lines.
4889 * SSL/TLS tracing code. This parses out SSL/TLS records using the
4890 message callback and prints the results. Needs compile time option
4891 "enable-ssl-trace". New options to s_client and s_server to enable
4896 * New ctrl and macro to retrieve supported points extensions.
4897 Print out extension in s_server and s_client.
4901 * New functions to retrieve certificate signature and signature
4906 * Add functions to retrieve and manipulate the raw cipherlist sent by a
4911 * New Suite B modes for TLS code. These use and enforce the requirements
4912 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
4913 only use Suite B curves. The Suite B modes can be set by using the
4914 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
4918 * New chain verification flags for Suite B levels of security. Check
4919 algorithms are acceptable when flags are set in X509_verify_cert.
4923 * Make tls1_check_chain return a set of flags indicating checks passed
4924 by a certificate chain. Add additional tests to handle client
4925 certificates: checks for matching certificate type and issuer name
4930 * If an attempt is made to use a signature algorithm not in the peer
4931 preference list abort the handshake. If client has no suitable
4932 signature algorithms in response to a certificate request do not
4933 use the certificate.
4937 * If server EC tmp key is not in client preference list abort handshake.
4941 * Add support for certificate stores in CERT structure. This makes it
4942 possible to have different stores per SSL structure or one store in
4943 the parent SSL_CTX. Include distinct stores for certificate chain
4944 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
4945 to build and store a certificate chain in CERT structure: returning
4946 an error if the chain cannot be built: this will allow applications
4947 to test if a chain is correctly configured.
4949 Note: if the CERT based stores are not set then the parent SSL_CTX
4950 store is used to retain compatibility with existing behaviour.
4955 * New function ssl_set_client_disabled to set a ciphersuite disabled
4956 mask based on the current session, check mask when sending client
4957 hello and checking the requested ciphersuite.
4961 * New ctrls to retrieve and set certificate types in a certificate
4962 request message. Print out received values in s_client. If certificate
4963 types is not set with custom values set sensible values based on
4964 supported signature algorithms.
4968 * Support for distinct client and server supported signature algorithms.
4972 * Add certificate callback. If set this is called whenever a certificate
4973 is required by client or server. An application can decide which
4974 certificate chain to present based on arbitrary criteria: for example
4975 supported signature algorithms. Add very simple example to s_server.
4976 This fixes many of the problems and restrictions of the existing client
4977 certificate callback: for example you can now clear an existing
4978 certificate and specify the whole chain.
4982 * Add new "valid_flags" field to CERT_PKEY structure which determines what
4983 the certificate can be used for (if anything). Set valid_flags field
4984 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
4985 to have similar checks in it.
4987 Add new "cert_flags" field to CERT structure and include a "strict mode".
4988 This enforces some TLS certificate requirements (such as only permitting
4989 certificate signature algorithms contained in the supported algorithms
4990 extension) which some implementations ignore: this option should be used
4991 with caution as it could cause interoperability issues.
4995 * Update and tidy signature algorithm extension processing. Work out
4996 shared signature algorithms based on preferences and peer algorithms
4997 and print them out in s_client and s_server. Abort handshake if no
4998 shared signature algorithms.
5002 * Add new functions to allow customised supported signature algorithms
5003 for SSL and SSL_CTX structures. Add options to s_client and s_server
5008 * New function SSL_certs_clear() to delete all references to certificates
5009 from an SSL structure. Before this once a certificate had been added
5010 it couldn't be removed.
5014 * Integrate hostname, email address and IP address checking with certificate
5015 verification. New verify options supporting checking in openssl utility.
5019 * Fixes and wildcard matching support to hostname and email checking
5020 functions. Add manual page.
5022 *Florian Weimer (Red Hat Product Security Team)*
5024 * New functions to check a hostname email or IP address against a
5025 certificate. Add options x509 utility to print results of checks against
5030 * Fix OCSP checking.
5032 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5034 * Initial experimental support for explicitly trusted non-root CAs.
5035 OpenSSL still tries to build a complete chain to a root but if an
5036 intermediate CA has a trust setting included that is used. The first
5037 setting is used: whether to trust (e.g., -addtrust option to the x509
5042 * Add -trusted_first option which attempts to find certificates in the
5043 trusted store even if an untrusted chain is also supplied.
5047 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5048 platform support for Linux and Android.
5052 * Support for linux-x32, ILP32 environment in x86_64 framework.
5056 * Experimental multi-implementation support for FIPS capable OpenSSL.
5057 When in FIPS mode the approved implementations are used as normal,
5058 when not in FIPS mode the internal unapproved versions are used instead.
5059 This means that the FIPS capable OpenSSL isn't forced to use the
5060 (often lower performance) FIPS implementations outside FIPS mode.
5064 * Transparently support X9.42 DH parameters when calling
5065 PEM_read_bio_DHparameters. This means existing applications can handle
5066 the new parameter format automatically.
5070 * Initial experimental support for X9.42 DH parameter format: mainly
5071 to support use of 'q' parameter for RFC5114 parameters.
5075 * Add DH parameters from RFC5114 including test data to dhtest.
5079 * Support for automatic EC temporary key parameter selection. If enabled
5080 the most preferred EC parameters are automatically used instead of
5081 hardcoded fixed parameters. Now a server just has to call:
5082 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5083 support ECDH and use the most appropriate parameters.
5087 * Enhance and tidy EC curve and point format TLS extension code. Use
5088 static structures instead of allocation if default values are used.
5089 New ctrls to set curves we wish to support and to retrieve shared curves.
5090 Print out shared curves in s_server. New options to s_server and s_client
5091 to set list of supported curves.
5095 * New ctrls to retrieve supported signature algorithms and
5096 supported curve values as an array of NIDs. Extend openssl utility
5097 to print out received values.
5101 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5102 between NIDs and the more common NIST names such as "P-256". Enhance
5103 ecparam utility and ECC method to recognise the NIST names for curves.
5107 * Enhance SSL/TLS certificate chain handling to support different
5108 chains for each certificate instead of one chain in the parent SSL_CTX.
5112 * Support for fixed DH ciphersuite client authentication: where both
5113 server and client use DH certificates with common parameters.
5117 * Support for fixed DH ciphersuites: those requiring DH server
5122 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5124 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5125 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5126 X509_CINF_get_signature were reverted post internal team review.
5132 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016] ###
5134 * OCSP Status Request extension unbounded memory growth
5136 A malicious client can send an excessively large OCSP Status Request
5137 extension. If that client continually requests renegotiation, sending a
5138 large OCSP Status Request extension each time, then there will be unbounded
5139 memory growth on the server. This will eventually lead to a Denial Of
5140 Service attack through memory exhaustion. Servers with a default
5141 configuration are vulnerable even if they do not support OCSP. Builds using
5142 the "no-ocsp" build time option are not affected.
5144 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5149 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5152 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5158 * OOB write in MDC2_Update()
5160 An overflow can occur in MDC2_Update() either if called directly or
5161 through the EVP_DigestUpdate() function using MDC2. If an attacker
5162 is able to supply very large amounts of input data after a previous
5163 call to EVP_EncryptUpdate() with a partial block then a length check
5164 can overflow resulting in a heap corruption.
5166 The amount of data needed is comparable to SIZE_MAX which is impractical
5169 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5174 * Malformed SHA512 ticket DoS
5176 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5177 DoS attack where a malformed ticket will result in an OOB read which will
5180 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5181 a custom server callback and ticket lookup mechanism.
5183 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5188 * OOB write in BN_bn2dec()
5190 The function BN_bn2dec() does not check the return value of BN_div_word().
5191 This can cause an OOB write if an application uses this function with an
5192 overly large BIGNUM. This could be a problem if an overly large certificate
5193 or CRL is printed out from an untrusted source. TLS is not affected because
5194 record limits will reject an oversized certificate before it is parsed.
5196 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5201 * OOB read in TS_OBJ_print_bio()
5203 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5204 the total length the OID text representation would use and not the amount
5205 of data written. This will result in OOB reads when large OIDs are
5208 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5213 * Pointer arithmetic undefined behaviour
5215 Avoid some undefined pointer arithmetic
5217 A common idiom in the codebase is to check limits in the following manner:
5220 Where "p" points to some malloc'd data of SIZE bytes and
5223 "len" here could be from some externally supplied data (e.g. from a TLS
5226 The rules of C pointer arithmetic are such that "p + len" is only well
5227 defined where len <= SIZE. Therefore the above idiom is actually
5228 undefined behaviour.
5230 For example this could cause problems if some malloc implementation
5231 provides an address for "p" such that "p + len" actually overflows for
5232 values of len that are too big and therefore p + len < limit.
5234 This issue was reported to OpenSSL by Guido Vranken
5239 * Constant time flag not preserved in DSA signing
5241 Operations in the DSA signing algorithm should run in constant time in
5242 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5243 implementation means that a non-constant time codepath is followed for
5244 certain operations. This has been demonstrated through a cache-timing
5245 attack to be sufficient for an attacker to recover the private DSA key.
5247 This issue was reported by César Pereida (Aalto University), Billy Brumley
5248 (Tampere University of Technology), and Yuval Yarom (The University of
5249 Adelaide and NICTA).
5254 * DTLS buffered message DoS
5256 In a DTLS connection where handshake messages are delivered out-of-order
5257 those messages that OpenSSL is not yet ready to process will be buffered
5258 for later use. Under certain circumstances, a flaw in the logic means that
5259 those messages do not get removed from the buffer even though the handshake
5260 has been completed. An attacker could force up to approx. 15 messages to
5261 remain in the buffer when they are no longer required. These messages will
5262 be cleared when the DTLS connection is closed. The default maximum size for
5263 a message is 100k. Therefore the attacker could force an additional 1500k
5264 to be consumed per connection. By opening many simulataneous connections an
5265 attacker could cause a DoS attack through memory exhaustion.
5267 This issue was reported to OpenSSL by Quan Luo.
5272 * DTLS replay protection DoS
5274 A flaw in the DTLS replay attack protection mechanism means that records
5275 that arrive for future epochs update the replay protection "window" before
5276 the MAC for the record has been validated. This could be exploited by an
5277 attacker by sending a record for the next epoch (which does not have to
5278 decrypt or have a valid MAC), with a very large sequence number. This means
5279 that all subsequent legitimate packets are dropped causing a denial of
5280 service for a specific DTLS connection.
5282 This issue was reported to OpenSSL by the OCAP audit team.
5287 * Certificate message OOB reads
5289 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5290 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5291 theoretical DoS risk but this has not been observed in practice on common
5294 The messages affected are client certificate, client certificate request
5295 and server certificate. As a result the attack can only be performed
5296 against a client or a server which enables client authentication.
5298 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5303 ### Changes between 1.0.1s and 1.0.1t [3 May 2016] ###
5305 * Prevent padding oracle in AES-NI CBC MAC check
5307 A MITM attacker can use a padding oracle attack to decrypt traffic
5308 when the connection uses an AES CBC cipher and the server support
5311 This issue was introduced as part of the fix for Lucky 13 padding
5312 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5313 constant time by making sure that always the same bytes are read and
5314 compared against either the MAC or padding bytes. But it no longer
5315 checked that there was enough data to have both the MAC and padding
5318 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5323 * Fix EVP_EncodeUpdate overflow
5325 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5326 Base64 encoding of binary data. If an attacker is able to supply very large
5327 amounts of input data then a length check can overflow resulting in a heap
5330 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5331 the PEM_write_bio* family of functions. These are mainly used within the
5332 OpenSSL command line applications, so any application which processes data
5333 from an untrusted source and outputs it as a PEM file should be considered
5334 vulnerable to this issue. User applications that call these APIs directly
5335 with large amounts of untrusted data may also be vulnerable.
5337 This issue was reported by Guido Vranken.
5342 * Fix EVP_EncryptUpdate overflow
5344 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5345 is able to supply very large amounts of input data after a previous call to
5346 EVP_EncryptUpdate() with a partial block then a length check can overflow
5347 resulting in a heap corruption. Following an analysis of all OpenSSL
5348 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5349 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5350 the first called function after an EVP_EncryptInit(), and therefore that
5351 specific call must be safe. The second form is where the length passed to
5352 EVP_EncryptUpdate() can be seen from the code to be some small value and
5353 therefore there is no possibility of an overflow. Since all instances are
5354 one of these two forms, it is believed that there can be no overflows in
5355 internal code due to this problem. It should be noted that
5356 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5357 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5358 of these calls have also been analysed too and it is believed there are no
5359 instances in internal usage where an overflow could occur.
5361 This issue was reported by Guido Vranken.
5366 * Prevent ASN.1 BIO excessive memory allocation
5368 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5369 a short invalid encoding can casuse allocation of large amounts of memory
5370 potentially consuming excessive resources or exhausting memory.
5372 Any application parsing untrusted data through d2i BIO functions is
5373 affected. The memory based functions such as d2i_X509() are *not* affected.
5374 Since the memory based functions are used by the TLS library, TLS
5375 applications are not affected.
5377 This issue was reported by Brian Carpenter.
5384 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5385 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5386 in arbitrary stack data being returned in the buffer.
5388 This issue was reported by Guido Vranken.
5393 * Modify behavior of ALPN to invoke callback after SNI/servername
5394 callback, such that updates to the SSL_CTX affect ALPN.
5398 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5403 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5404 methods are enabled and ssl2 is disabled the methods return NULL.
5408 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016] ###
5410 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5411 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5412 provide any "EXPORT" or "LOW" strength ciphers.
5416 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5417 is by default disabled at build-time. Builds that are not configured with
5418 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5419 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5420 will need to explicitly call either of:
5422 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5424 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5426 as appropriate. Even if either of those is used, or the application
5427 explicitly uses the version-specific SSLv2_method() or its client and
5428 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5429 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5430 ciphers, and SSLv2 56-bit DES are no longer available.
5435 * Fix a double-free in DSA code
5437 A double free bug was discovered when OpenSSL parses malformed DSA private
5438 keys and could lead to a DoS attack or memory corruption for applications
5439 that receive DSA private keys from untrusted sources. This scenario is
5442 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5448 * Disable SRP fake user seed to address a server memory leak.
5450 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5452 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5453 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5454 was changed to ignore the "fake user" SRP seed, even if the seed
5457 Users should use SRP_VBASE_get1_by_user instead. Note that in
5458 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5459 also that even though configuring the SRP seed attempts to hide
5460 invalid usernames by continuing the handshake with fake
5461 credentials, this behaviour is not constant time and no strong
5462 guarantees are made that the handshake is indistinguishable from
5463 that of a valid user.
5468 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5470 In the BN_hex2bn function the number of hex digits is calculated using an
5471 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
5472 large values of |i| this can result in |bn_expand| not allocating any
5473 memory because |i * 4| is negative. This can leave the internal BIGNUM data
5474 field as NULL leading to a subsequent NULL ptr deref. For very large values
5475 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
5476 In this case memory is allocated to the internal BIGNUM data field, but it
5477 is insufficiently sized leading to heap corruption. A similar issue exists
5478 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5479 is ever called by user applications with very large untrusted hex/dec data.
5480 This is anticipated to be a rare occurrence.
5482 All OpenSSL internal usage of these functions use data that is not expected
5483 to be untrusted, e.g. config file data or application command line
5484 arguments. If user developed applications generate config file data based
5485 on untrusted data then it is possible that this could also lead to security
5486 consequences. This is also anticipated to be rare.
5488 This issue was reported to OpenSSL by Guido Vranken.
5493 * Fix memory issues in BIO_*printf functions
5495 The internal |fmtstr| function used in processing a "%s" format string in
5496 the BIO_*printf functions could overflow while calculating the length of a
5497 string and cause an OOB read when printing very long strings.
5499 Additionally the internal |doapr_outch| function can attempt to write to an
5500 OOB memory location (at an offset from the NULL pointer) in the event of a
5501 memory allocation failure. In 1.0.2 and below this could be caused where
5502 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5503 could be in processing a very long "%s" format string. Memory leaks can
5506 The first issue may mask the second issue dependent on compiler behaviour.
5507 These problems could enable attacks where large amounts of untrusted data
5508 is passed to the BIO_*printf functions. If applications use these functions
5509 in this way then they could be vulnerable. OpenSSL itself uses these
5510 functions when printing out human-readable dumps of ASN.1 data. Therefore
5511 applications that print this data could be vulnerable if the data is from
5512 untrusted sources. OpenSSL command line applications could also be
5513 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5514 as command line arguments.
5516 Libssl is not considered directly vulnerable. Additionally certificates etc
5517 received via remote connections via libssl are also unlikely to be able to
5518 trigger these issues because of message size limits enforced within libssl.
5520 This issue was reported to OpenSSL Guido Vranken.
5525 * Side channel attack on modular exponentiation
5527 A side-channel attack was found which makes use of cache-bank conflicts on
5528 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5529 of RSA keys. The ability to exploit this issue is limited as it relies on
5530 an attacker who has control of code in a thread running on the same
5531 hyper-threaded core as the victim thread which is performing decryptions.
5533 This issue was reported to OpenSSL by Yuval Yarom, The University of
5534 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5535 Nadia Heninger, University of Pennsylvania with more information at
5536 http://cachebleed.info.
5541 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5542 if no keysize is specified with default_bits. This fixes an
5543 omission in an earlier change that changed all RSA/DSA key generation
5544 apps to use 2048 bits by default.
5548 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016] ###
5550 * Protection for DH small subgroup attacks
5552 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5553 switched on by default and cannot be disabled. This could have some
5558 * SSLv2 doesn't block disabled ciphers
5560 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5561 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5562 been disabled, provided that the SSLv2 protocol was not also disabled via
5565 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5566 and Sebastian Schinzel.
5571 * Reject DH handshakes with parameters shorter than 1024 bits.
5575 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015] ###
5577 * Certificate verify crash with missing PSS parameter
5579 The signature verification routines will crash with a NULL pointer
5580 dereference if presented with an ASN.1 signature using the RSA PSS
5581 algorithm and absent mask generation function parameter. Since these
5582 routines are used to verify certificate signature algorithms this can be
5583 used to crash any certificate verification operation and exploited in a
5584 DoS attack. Any application which performs certificate verification is
5585 vulnerable including OpenSSL clients and servers which enable client
5588 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5593 * X509_ATTRIBUTE memory leak
5595 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5596 memory. This structure is used by the PKCS#7 and CMS routines so any
5597 application which reads PKCS#7 or CMS data from untrusted sources is
5598 affected. SSL/TLS is not affected.
5600 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5606 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5607 This changes the decoding behaviour for some invalid messages,
5608 though the change is mostly in the more lenient direction, and
5609 legacy behaviour is preserved as much as possible.
5613 * In DSA_generate_parameters_ex, if the provided seed is too short,
5614 use a random seed, as already documented.
5616 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5618 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015] ###
5620 * Alternate chains certificate forgery
5622 During certificate verfification, OpenSSL will attempt to find an
5623 alternative certificate chain if the first attempt to build such a chain
5624 fails. An error in the implementation of this logic can mean that an
5625 attacker could cause certain checks on untrusted certificates to be
5626 bypassed, such as the CA flag, enabling them to use a valid leaf
5627 certificate to act as a CA and "issue" an invalid certificate.
5629 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5635 * Race condition handling PSK identify hint
5637 If PSK identity hints are received by a multi-threaded client then
5638 the values are wrongly updated in the parent SSL_CTX structure. This can
5639 result in a race condition potentially leading to a double free of the
5645 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015] ###
5646 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5647 incompatibility in the handling of HMAC. The previous ABI has now been
5650 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015] ###
5652 * Malformed ECParameters causes infinite loop
5654 When processing an ECParameters structure OpenSSL enters an infinite loop
5655 if the curve specified is over a specially malformed binary polynomial
5658 This can be used to perform denial of service against any
5659 system which processes public keys, certificate requests or
5660 certificates. This includes TLS clients and TLS servers with
5661 client authentication enabled.
5663 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5668 * Exploitable out-of-bounds read in X509_cmp_time
5670 X509_cmp_time does not properly check the length of the ASN1_TIME
5671 string and can read a few bytes out of bounds. In addition,
5672 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5675 An attacker can use this to craft malformed certificates and CRLs of
5676 various sizes and potentially cause a segmentation fault, resulting in
5677 a DoS on applications that verify certificates or CRLs. TLS clients
5678 that verify CRLs are affected. TLS clients and servers with client
5679 authentication enabled may be affected if they use custom verification
5682 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5683 independently by Hanno Böck.
5688 * PKCS7 crash with missing EnvelopedContent
5690 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5691 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5692 with missing content and trigger a NULL pointer dereference on parsing.
5694 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5695 structures from untrusted sources are affected. OpenSSL clients and
5696 servers are not affected.
5698 This issue was reported to OpenSSL by Michal Zalewski (Google).
5703 * CMS verify infinite loop with unknown hash function
5705 When verifying a signedData message the CMS code can enter an infinite loop
5706 if presented with an unknown hash function OID. This can be used to perform
5707 denial of service against any system which verifies signedData messages using
5709 This issue was reported to OpenSSL by Johannes Bauer.
5714 * Race condition handling NewSessionTicket
5716 If a NewSessionTicket is received by a multi-threaded client when attempting to
5717 reuse a previous ticket then a race condition can occur potentially leading to
5718 a double free of the ticket data.
5723 * Reject DH handshakes with parameters shorter than 768 bits.
5725 *Kurt Roeckx and Emilia Kasper*
5727 * dhparam: generate 2048-bit parameters by default.
5729 *Kurt Roeckx and Emilia Kasper*
5731 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015] ###
5733 * Segmentation fault in ASN1_TYPE_cmp fix
5735 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5736 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5737 certificate signature algorithm consistency this can be used to crash any
5738 certificate verification operation and exploited in a DoS attack. Any
5739 application which performs certificate verification is vulnerable including
5740 OpenSSL clients and servers which enable client authentication.
5745 * ASN.1 structure reuse memory corruption fix
5747 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5748 memory corruption via an invalid write. Such reuse is and has been
5749 strongly discouraged and is believed to be rare.
5751 Applications that parse structures containing CHOICE or ANY DEFINED BY
5752 components may be affected. Certificate parsing (d2i_X509 and related
5753 functions) are however not affected. OpenSSL clients and servers are
5759 * PKCS7 NULL pointer dereferences fix
5761 The PKCS#7 parsing code does not handle missing outer ContentInfo
5762 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5763 missing content and trigger a NULL pointer dereference on parsing.
5765 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5766 otherwise parse PKCS#7 structures from untrusted sources are
5767 affected. OpenSSL clients and servers are not affected.
5769 This issue was reported to OpenSSL by Michal Zalewski (Google).
5774 * DoS via reachable assert in SSLv2 servers fix
5776 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5777 servers that both support SSLv2 and enable export cipher suites by sending
5778 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5780 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5781 (OpenSSL development team).
5786 * Use After Free following d2i_ECPrivatekey error fix
5788 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5789 could cause a use after free condition. This, in turn, could cause a double
5790 free in several private key parsing functions (such as d2i_PrivateKey
5791 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5792 for applications that receive EC private keys from untrusted
5793 sources. This scenario is considered rare.
5795 This issue was discovered by the BoringSSL project and fixed in their
5801 * X509_to_X509_REQ NULL pointer deref fix
5803 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5804 the certificate key is invalid. This function is rarely used in practice.
5806 This issue was discovered by Brian Carpenter.
5811 * Removed the export ciphers from the DEFAULT ciphers
5815 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015] ###
5817 * Build fixes for the Windows and OpenVMS platforms
5819 *Matt Caswell and Richard Levitte*
5821 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015] ###
5823 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
5824 message can cause a segmentation fault in OpenSSL due to a NULL pointer
5825 dereference. This could lead to a Denial Of Service attack. Thanks to
5826 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
5831 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
5832 dtls1_buffer_record function under certain conditions. In particular this
5833 could occur if an attacker sent repeated DTLS records with the same
5834 sequence number but for the next epoch. The memory leak could be exploited
5835 by an attacker in a Denial of Service attack through memory exhaustion.
5836 Thanks to Chris Mueller for reporting this issue.
5841 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
5842 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
5843 method would be set to NULL which could later result in a NULL pointer
5844 dereference. Thanks to Frank Schmirler for reporting this issue.
5849 * Abort handshake if server key exchange message is omitted for ephemeral
5852 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
5853 reporting this issue.
5858 * Remove non-export ephemeral RSA code on client and server. This code
5859 violated the TLS standard by allowing the use of temporary RSA keys in
5860 non-export ciphersuites and could be used by a server to effectively
5861 downgrade the RSA key length used to a value smaller than the server
5862 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
5863 INRIA or reporting this issue.
5868 * Fixed issue where DH client certificates are accepted without verification.
5869 An OpenSSL server will accept a DH certificate for client authentication
5870 without the certificate verify message. This effectively allows a client to
5871 authenticate without the use of a private key. This only affects servers
5872 which trust a client certificate authority which issues certificates
5873 containing DH keys: these are extremely rare and hardly ever encountered.
5874 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
5880 * Ensure that the session ID context of an SSL is updated when its
5881 SSL_CTX is updated via SSL_set_SSL_CTX.
5883 The session ID context is typically set from the parent SSL_CTX,
5884 and can vary with the CTX.
5888 * Fix various certificate fingerprint issues.
5890 By using non-DER or invalid encodings outside the signed portion of a
5891 certificate the fingerprint can be changed without breaking the signature.
5892 Although no details of the signed portion of the certificate can be changed
5893 this can cause problems with some applications: e.g. those using the
5894 certificate fingerprint for blacklists.
5896 1. Reject signatures with non zero unused bits.
5898 If the BIT STRING containing the signature has non zero unused bits reject
5899 the signature. All current signature algorithms require zero unused bits.
5901 2. Check certificate algorithm consistency.
5903 Check the AlgorithmIdentifier inside TBS matches the one in the
5904 certificate signature. NB: this will result in signature failure
5905 errors for some broken certificates.
5907 Thanks to Konrad Kraszewski from Google for reporting this issue.
5909 3. Check DSA/ECDSA signatures use DER.
5911 Re-encode DSA/ECDSA signatures and compare with the original received
5912 signature. Return an error if there is a mismatch.
5914 This will reject various cases including garbage after signature
5915 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
5916 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
5917 (negative or with leading zeroes).
5919 Further analysis was conducted and fixes were developed by Stephen Henson
5920 of the OpenSSL core team.
5926 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
5927 results on some platforms, including x86_64. This bug occurs at random
5928 with a very low probability, and is not known to be exploitable in any
5929 way, though its exact impact is difficult to determine. Thanks to Pieter
5930 Wuille (Blockstream) who reported this issue and also suggested an initial
5931 fix. Further analysis was conducted by the OpenSSL development team and
5932 Adam Langley of Google. The final fix was developed by Andy Polyakov of
5933 the OpenSSL core team.
5938 *) Do not resume sessions on the server if the negotiated protocol
5939 version does not match the session's version. Resuming with a different
5940 version, while not strictly forbidden by the RFC, is of questionable
5941 sanity and breaks all known clients.
5943 *David Benjamin, Emilia Käsper*
5945 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
5946 early CCS messages during renegotiation. (Note that because
5947 renegotiation is encrypted, this early CCS was not exploitable.)
5951 *) Tighten client-side session ticket handling during renegotiation:
5952 ensure that the client only accepts a session ticket if the server sends
5953 the extension anew in the ServerHello. Previously, a TLS client would
5954 reuse the old extension state and thus accept a session ticket if one was
5955 announced in the initial ServerHello.
5957 Similarly, ensure that the client requires a session ticket if one
5958 was advertised in the ServerHello. Previously, a TLS client would
5959 ignore a missing NewSessionTicket message.
5963 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014] ###
5967 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
5968 sends a carefully crafted handshake message, to cause OpenSSL to fail
5969 to free up to 64k of memory causing a memory leak. This could be
5970 exploited in a Denial Of Service attack. This issue affects OpenSSL
5971 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
5972 whether SRTP is used or configured. Implementations of OpenSSL that
5973 have been compiled with OPENSSL_NO_SRTP defined are not affected.
5975 The fix was developed by the OpenSSL team.
5980 * Session Ticket Memory Leak.
5982 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
5983 integrity of that ticket is first verified. In the event of a session
5984 ticket integrity check failing, OpenSSL will fail to free memory
5985 causing a memory leak. By sending a large number of invalid session
5986 tickets an attacker could exploit this issue in a Denial Of Service
5992 * Build option no-ssl3 is incomplete.
5994 When OpenSSL is configured with "no-ssl3" as a build option, servers
5995 could accept and complete a SSL 3.0 handshake, and clients could be
5996 configured to send them.
5999 *Akamai and the OpenSSL team*
6001 * Add support for TLS_FALLBACK_SCSV.
6002 Client applications doing fallback retries should call
6003 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6006 *Adam Langley, Bodo Moeller*
6008 * Add additional DigestInfo checks.
6010 Re-encode DigestInto in DER and check against the original when
6011 verifying RSA signature: this will reject any improperly encoded
6012 DigestInfo structures.
6014 Note: this is a precautionary measure and no attacks are currently known.
6019 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014] ###
6021 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6022 SRP code can be overrun an internal buffer. Add sanity check that
6023 g, A, B < N to SRP code.
6025 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6026 Group for discovering this issue.
6031 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6032 TLS 1.0 instead of higher protocol versions when the ClientHello message
6033 is badly fragmented. This allows a man-in-the-middle attacker to force a
6034 downgrade to TLS 1.0 even if both the server and the client support a
6035 higher protocol version, by modifying the client's TLS records.
6037 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6038 researching this issue.
6043 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6044 to a denial of service attack. A malicious server can crash the client
6045 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6046 ciphersuite and sending carefully crafted handshake messages.
6048 Thanks to Felix Gröbert (Google) for discovering and researching this
6054 * By sending carefully crafted DTLS packets an attacker could cause openssl
6055 to leak memory. This can be exploited through a Denial of Service attack.
6056 Thanks to Adam Langley for discovering and researching this issue.
6061 * An attacker can force openssl to consume large amounts of memory whilst
6062 processing DTLS handshake messages. This can be exploited through a
6063 Denial of Service attack.
6064 Thanks to Adam Langley for discovering and researching this issue.
6069 * An attacker can force an error condition which causes openssl to crash
6070 whilst processing DTLS packets due to memory being freed twice. This
6071 can be exploited through a Denial of Service attack.
6072 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6078 * If a multithreaded client connects to a malicious server using a resumed
6079 session and the server sends an ec point format extension it could write
6080 up to 255 bytes to freed memory.
6082 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6088 * A malicious server can crash an OpenSSL client with a null pointer
6089 dereference (read) by specifying an SRP ciphersuite even though it was not
6090 properly negotiated with the client. This can be exploited through a
6091 Denial of Service attack.
6093 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6094 discovering and researching this issue.
6099 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6100 X509_name_oneline, X509_name_print_ex et al. to leak some information
6101 from the stack. Applications may be affected if they echo pretty printing
6102 output to the attacker.
6104 Thanks to Ivan Fratric (Google) for discovering this issue.
6107 *Emilia Käsper, and Steve Henson*
6109 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6110 for corner cases. (Certain input points at infinity could lead to
6111 bogus results, with non-infinity inputs mapped to infinity too.)
6115 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014] ###
6117 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6118 handshake can force the use of weak keying material in OpenSSL
6119 SSL/TLS clients and servers.
6121 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6122 researching this issue. [CVE-2014-0224][]
6124 *KIKUCHI Masashi, Steve Henson*
6126 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6127 OpenSSL DTLS client the code can be made to recurse eventually crashing
6130 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6133 *Imre Rad, Steve Henson*
6135 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6136 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6137 client or server. This is potentially exploitable to run arbitrary
6138 code on a vulnerable client or server.
6140 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6142 *Jüri Aedla, Steve Henson*
6144 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6145 are subject to a denial of service attack.
6147 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6148 this issue. [CVE-2014-3470][]
6150 *Felix Gröbert, Ivan Fratric, Steve Henson*
6152 * Harmonize version and its documentation. -f flag is used to display
6155 *mancha <mancha1@zoho.com>*
6157 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6158 in i2d_ECPrivateKey.
6160 *mancha <mancha1@zoho.com>*
6162 * Fix some double frees. These are not thought to be exploitable.
6164 *mancha <mancha1@zoho.com>*
6166 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014] ###
6168 * A missing bounds check in the handling of the TLS heartbeat extension
6169 can be used to reveal up to 64k of memory to a connected client or
6172 Thanks for Neel Mehta of Google Security for discovering this bug and to
6173 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6174 preparing the fix [CVE-2014-0160][]
6176 *Adam Langley, Bodo Moeller*
6178 * Fix for the attack described in the paper "Recovering OpenSSL
6179 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6180 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6181 http://eprint.iacr.org/2014/140
6183 Thanks to Yuval Yarom and Naomi Benger for discovering this
6184 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6186 *Yuval Yarom and Naomi Benger*
6188 * TLS pad extension: draft-agl-tls-padding-03
6190 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6191 TLS client Hello record length value would otherwise be > 255 and
6192 less that 512 pad with a dummy extension containing zeroes so it
6193 is at least 512 bytes long.
6196 *Adam Langley, Steve Henson*
6198 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014] ###
6200 * Fix for TLS record tampering bug. A carefully crafted invalid
6201 handshake could crash OpenSSL with a NULL pointer exception.
6202 Thanks to Anton Johansson for reporting this issues.
6205 * Keep original DTLS digest and encryption contexts in retransmission
6206 structures so we can use the previous session parameters if they need
6207 to be resent. [CVE-2013-6450][]
6211 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6212 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6213 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6214 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6215 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6216 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6218 *Rob Stradling, Adam Langley*
6220 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013] ###
6222 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6223 supporting platforms or when small records were transferred.
6225 *Andy Polyakov, Steve Henson*
6227 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013] ###
6229 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6231 This addresses the flaw in CBC record processing discovered by
6232 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6233 at: http://www.isg.rhul.ac.uk/tls/
6235 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6236 Security Group at Royal Holloway, University of London
6237 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6238 Emilia Käsper for the initial patch.
6241 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6243 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6244 ciphersuites which can be exploited in a denial of service attack.
6245 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6246 and detecting this bug and to Wolfgang Ettlinger
6247 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6252 * Return an error when checking OCSP signatures when key is NULL.
6253 This fixes a DoS attack. [CVE-2013-0166][]
6257 * Make openssl verify return errors.
6259 *Chris Palmer <palmer@google.com> and Ben Laurie*
6261 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6262 the right response is stapled. Also change SSL_get_certificate()
6263 so it returns the certificate actually sent.
6264 See http://rt.openssl.org/Ticket/Display.html?id=2836.
6266 *Rob Stradling <rob.stradling@comodo.com>*
6268 * Fix possible deadlock when decoding public keys.
6272 * Don't use TLS 1.0 record version number in initial client hello
6277 ### Changes between 1.0.1b and 1.0.1c [10 May 2012] ###
6279 * Sanity check record length before skipping explicit IV in TLS
6280 1.2, 1.1 and DTLS to fix DoS attack.
6282 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6283 fuzzing as a service testing platform.
6288 * Initialise tkeylen properly when encrypting CMS messages.
6289 Thanks to Solar Designer of Openwall for reporting this issue.
6293 * In FIPS mode don't try to use composite ciphers as they are not
6298 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012] ###
6300 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6301 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6302 mean any application compiled against OpenSSL 1.0.0 headers setting
6303 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6304 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6305 0x10000000L Any application which was previously compiled against
6306 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6307 will need to be recompiled as a result. Letting be results in
6308 inability to disable specifically TLS 1.1 and in client context,
6309 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6313 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6314 disable just protocol X, but all protocols above X *if* there are
6315 protocols *below* X still enabled. In more practical terms it means
6316 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6317 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
6318 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
6323 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012] ###
6325 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6326 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6327 in CRYPTO_realloc_clean.
6329 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6330 issue and to Adam Langley <agl@chromium.org> for fixing it.
6333 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6335 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6339 * Workarounds for some broken servers that "hang" if a client hello
6340 record length exceeds 255 bytes.
6342 1. Do not use record version number > TLS 1.0 in initial client
6343 hello: some (but not all) hanging servers will now work.
6344 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6345 the number of ciphers sent in the client hello. This should be
6346 set to an even number, such as 50, for example by passing:
6347 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6348 Most broken servers should now work.
6349 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6350 TLS 1.2 client support entirely.
6354 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6358 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012] ###
6360 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6361 STRING form instead of a DigestInfo.
6365 * The format used for MDC2 RSA signatures is inconsistent between EVP
6366 and the RSA_sign/RSA_verify functions. This was made more apparent when
6367 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6368 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6369 the correct format in RSA_verify so both forms transparently work.
6373 * Some servers which support TLS 1.0 can choke if we initially indicate
6374 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6375 encrypted premaster secret. As a workaround use the maximum permitted
6376 client version in client hello, this should keep such servers happy
6377 and still work with previous versions of OpenSSL.
6381 * Add support for TLS/DTLS heartbeats.
6383 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6385 * Add support for SCTP.
6387 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6389 * Improved PRNG seeding for VOS.
6391 *Paul Green <Paul.Green@stratus.com>*
6393 * Extensive assembler packs updates, most notably:
6395 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6396 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6397 - x86_64: bit-sliced AES implementation;
6398 - ARM: NEON support, contemporary platforms optimizations;
6399 - s390x: z196 support;
6400 - *: GHASH and GF(2^m) multiplication implementations;
6404 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6405 (removal of unnecessary code)
6407 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6409 * Add TLS key material exporter from RFC 5705.
6413 * Add DTLS-SRTP negotiation from RFC 5764.
6417 * Add Next Protocol Negotiation,
6418 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
6419 disabled with a no-npn flag to config or Configure. Code donated
6422 *Adam Langley <agl@google.com> and Ben Laurie*
6424 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6425 NIST-P256, NIST-P521, with constant-time single point multiplication on
6426 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6427 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6428 Code made available under Apache License version 2.0.
6430 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6431 line to include this in your build of OpenSSL, and run "make depend" (or
6432 "make update"). This enables the following EC_METHODs:
6434 EC_GFp_nistp224_method()
6435 EC_GFp_nistp256_method()
6436 EC_GFp_nistp521_method()
6438 EC_GROUP_new_by_curve_name() will automatically use these (while
6439 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6442 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6444 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6445 all platforms. Move ssize_t definition from e_os.h to the public
6446 header file e_os2.h as it now appears in public header file cms.h
6450 * New -sigopt option to the ca, req and x509 utilities. Additional
6451 signature parameters can be passed using this option and in
6456 * Add RSA PSS signing function. This will generate and set the
6457 appropriate AlgorithmIdentifiers for PSS based on those in the
6458 corresponding EVP_MD_CTX structure. No application support yet.
6462 * Support for companion algorithm specific ASN1 signing routines.
6463 New function ASN1_item_sign_ctx() signs a pre-initialised
6464 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6465 the appropriate parameters.
6469 * Add new algorithm specific ASN1 verification initialisation function
6470 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6471 handling will be the same no matter what EVP_PKEY_METHOD is used.
6472 Add a PSS handler to support verification of PSS signatures: checked
6473 against a number of sample certificates.
6477 * Add signature printing for PSS. Add PSS OIDs.
6479 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6481 * Add algorithm specific signature printing. An individual ASN1 method
6482 can now print out signatures instead of the standard hex dump.
6484 More complex signatures (e.g. PSS) can print out more meaningful
6485 information. Include DSA version that prints out the signature
6490 * Password based recipient info support for CMS library: implementing
6495 * Split password based encryption into PBES2 and PBKDF2 functions. This
6496 neatly separates the code into cipher and PBE sections and is required
6497 for some algorithms that split PBES2 into separate pieces (such as
6498 password based CMS).
6502 * Session-handling fixes:
6503 - Fix handling of connections that are resuming with a session ID,
6504 but also support Session Tickets.
6505 - Fix a bug that suppressed issuing of a new ticket if the client
6506 presented a ticket with an expired session.
6507 - Try to set the ticket lifetime hint to something reasonable.
6508 - Make tickets shorter by excluding irrelevant information.
6509 - On the client side, don't ignore renewed tickets.
6511 *Adam Langley, Bodo Moeller (Google)*
6513 * Fix PSK session representation.
6517 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6519 This work was sponsored by Intel.
6523 * Add GCM support to TLS library. Some custom code is needed to split
6524 the IV between the fixed (from PRF) and explicit (from TLS record)
6525 portions. This adds all GCM ciphersuites supported by RFC5288 and
6526 RFC5289. Generalise some AES* cipherstrings to include GCM and
6527 add a special AESGCM string for GCM only.
6531 * Expand range of ctrls for AES GCM. Permit setting invocation
6532 field on decrypt and retrieval of invocation field only on encrypt.
6536 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6537 As required by RFC5289 these ciphersuites cannot be used if for
6538 versions of TLS earlier than 1.2.
6542 * For FIPS capable OpenSSL interpret a NULL default public key method
6543 as unset and return the appropriate default but do *not* set the default.
6544 This means we can return the appropriate method in applications that
6545 switch between FIPS and non-FIPS modes.
6549 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6550 ENGINE is used then we cannot handle that in the FIPS module so we
6551 keep original code iff non-FIPS operations are allowed.
6555 * Add -attime option to openssl utilities.
6557 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6559 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6563 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6564 FIPS EC methods unconditionally for now.
6568 * New build option no-ec2m to disable characteristic 2 code.
6572 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6573 all cases can be covered as some introduce binary incompatibilities.
6577 * Redirect RSA operations to FIPS module including keygen,
6578 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6582 * Add similar low level API blocking to ciphers.
6586 * Low level digest APIs are not approved in FIPS mode: any attempt
6587 to use these will cause a fatal error. Applications that *really* want
6588 to use them can use the private_* version instead.
6592 * Redirect cipher operations to FIPS module for FIPS builds.
6596 * Redirect digest operations to FIPS module for FIPS builds.
6600 * Update build system to add "fips" flag which will link in fipscanister.o
6601 for static and shared library builds embedding a signature if needed.
6605 * Output TLS supported curves in preference order instead of numerical
6606 order. This is currently hardcoded for the highest order curves first.
6607 This should be configurable so applications can judge speed vs strength.
6611 * Add TLS v1.2 server support for client authentication.
6615 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6620 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6621 FIPS modules versions.
6625 * Add TLS v1.2 client side support for client authentication. Keep cache
6626 of handshake records longer as we don't know the hash algorithm to use
6627 until after the certificate request message is received.
6631 * Initial TLS v1.2 client support. Add a default signature algorithms
6632 extension including all the algorithms we support. Parse new signature
6633 format in client key exchange. Relax some ECC signing restrictions for
6634 TLS v1.2 as indicated in RFC5246.
6638 * Add server support for TLS v1.2 signature algorithms extension. Switch
6639 to new signature format when needed using client digest preference.
6640 All server ciphersuites should now work correctly in TLS v1.2. No client
6641 support yet and no support for client certificates.
6645 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6646 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6647 ciphersuites. At present only RSA key exchange ciphersuites work with
6648 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6649 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6650 and version checking.
6654 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6655 with this defined it will not be affected by any changes to ssl internal
6656 structures. Add several utility functions to allow openssl application
6657 to work with OPENSSL_NO_SSL_INTERN defined.
6661 * A long standing patch to add support for SRP from EdelWeb (Peter
6662 Sylvester and Christophe Renou) was integrated.
6663 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6664 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6667 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6671 * Permit abbreviated handshakes when renegotiating using the function
6672 SSL_renegotiate_abbreviated().
6674 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6676 * Add call to ENGINE_register_all_complete() to
6677 ENGINE_load_builtin_engines(), so some implementations get used
6678 automatically instead of needing explicit application support.
6682 * Add support for TLS key exporter as described in RFC5705.
6684 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6686 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6687 a few changes are required:
6689 Add SSL_OP_NO_TLSv1_1 flag.
6690 Add TLSv1_1 methods.
6691 Update version checking logic to handle version 1.1.
6692 Add explicit IV handling (ported from DTLS code).
6693 Add command line options to s_client/s_server.
6700 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015] ###
6702 * X509_ATTRIBUTE memory leak
6704 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6705 memory. This structure is used by the PKCS#7 and CMS routines so any
6706 application which reads PKCS#7 or CMS data from untrusted sources is
6707 affected. SSL/TLS is not affected.
6709 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6715 * Race condition handling PSK identify hint
6717 If PSK identity hints are received by a multi-threaded client then
6718 the values are wrongly updated in the parent SSL_CTX structure. This can
6719 result in a race condition potentially leading to a double free of the
6725 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015] ###
6727 * Malformed ECParameters causes infinite loop
6729 When processing an ECParameters structure OpenSSL enters an infinite loop
6730 if the curve specified is over a specially malformed binary polynomial
6733 This can be used to perform denial of service against any
6734 system which processes public keys, certificate requests or
6735 certificates. This includes TLS clients and TLS servers with
6736 client authentication enabled.
6738 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6743 * Exploitable out-of-bounds read in X509_cmp_time
6745 X509_cmp_time does not properly check the length of the ASN1_TIME
6746 string and can read a few bytes out of bounds. In addition,
6747 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6750 An attacker can use this to craft malformed certificates and CRLs of
6751 various sizes and potentially cause a segmentation fault, resulting in
6752 a DoS on applications that verify certificates or CRLs. TLS clients
6753 that verify CRLs are affected. TLS clients and servers with client
6754 authentication enabled may be affected if they use custom verification
6757 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6758 independently by Hanno Böck.
6763 * PKCS7 crash with missing EnvelopedContent
6765 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6766 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6767 with missing content and trigger a NULL pointer dereference on parsing.
6769 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6770 structures from untrusted sources are affected. OpenSSL clients and
6771 servers are not affected.
6773 This issue was reported to OpenSSL by Michal Zalewski (Google).
6778 * CMS verify infinite loop with unknown hash function
6780 When verifying a signedData message the CMS code can enter an infinite loop
6781 if presented with an unknown hash function OID. This can be used to perform
6782 denial of service against any system which verifies signedData messages using
6784 This issue was reported to OpenSSL by Johannes Bauer.
6789 * Race condition handling NewSessionTicket
6791 If a NewSessionTicket is received by a multi-threaded client when attempting to
6792 reuse a previous ticket then a race condition can occur potentially leading to
6793 a double free of the ticket data.
6798 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015] ###
6800 * Segmentation fault in ASN1_TYPE_cmp fix
6802 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6803 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6804 certificate signature algorithm consistency this can be used to crash any
6805 certificate verification operation and exploited in a DoS attack. Any
6806 application which performs certificate verification is vulnerable including
6807 OpenSSL clients and servers which enable client authentication.
6812 * ASN.1 structure reuse memory corruption fix
6814 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6815 memory corruption via an invalid write. Such reuse is and has been
6816 strongly discouraged and is believed to be rare.
6818 Applications that parse structures containing CHOICE or ANY DEFINED BY
6819 components may be affected. Certificate parsing (d2i_X509 and related
6820 functions) are however not affected. OpenSSL clients and servers are
6826 * PKCS7 NULL pointer dereferences fix
6828 The PKCS#7 parsing code does not handle missing outer ContentInfo
6829 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6830 missing content and trigger a NULL pointer dereference on parsing.
6832 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6833 otherwise parse PKCS#7 structures from untrusted sources are
6834 affected. OpenSSL clients and servers are not affected.
6836 This issue was reported to OpenSSL by Michal Zalewski (Google).
6841 * DoS via reachable assert in SSLv2 servers fix
6843 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6844 servers that both support SSLv2 and enable export cipher suites by sending
6845 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6847 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6848 (OpenSSL development team).
6853 * Use After Free following d2i_ECPrivatekey error fix
6855 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6856 could cause a use after free condition. This, in turn, could cause a double
6857 free in several private key parsing functions (such as d2i_PrivateKey
6858 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6859 for applications that receive EC private keys from untrusted
6860 sources. This scenario is considered rare.
6862 This issue was discovered by the BoringSSL project and fixed in their
6868 * X509_to_X509_REQ NULL pointer deref fix
6870 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6871 the certificate key is invalid. This function is rarely used in practice.
6873 This issue was discovered by Brian Carpenter.
6878 * Removed the export ciphers from the DEFAULT ciphers
6882 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015] ###
6884 * Build fixes for the Windows and OpenVMS platforms
6886 *Matt Caswell and Richard Levitte*
6888 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015] ###
6890 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6891 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6892 dereference. This could lead to a Denial Of Service attack. Thanks to
6893 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6898 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6899 dtls1_buffer_record function under certain conditions. In particular this
6900 could occur if an attacker sent repeated DTLS records with the same
6901 sequence number but for the next epoch. The memory leak could be exploited
6902 by an attacker in a Denial of Service attack through memory exhaustion.
6903 Thanks to Chris Mueller for reporting this issue.
6908 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6909 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6910 method would be set to NULL which could later result in a NULL pointer
6911 dereference. Thanks to Frank Schmirler for reporting this issue.
6916 * Abort handshake if server key exchange message is omitted for ephemeral
6919 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6920 reporting this issue.
6925 * Remove non-export ephemeral RSA code on client and server. This code
6926 violated the TLS standard by allowing the use of temporary RSA keys in
6927 non-export ciphersuites and could be used by a server to effectively
6928 downgrade the RSA key length used to a value smaller than the server
6929 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6930 INRIA or reporting this issue.
6935 * Fixed issue where DH client certificates are accepted without verification.
6936 An OpenSSL server will accept a DH certificate for client authentication
6937 without the certificate verify message. This effectively allows a client to
6938 authenticate without the use of a private key. This only affects servers
6939 which trust a client certificate authority which issues certificates
6940 containing DH keys: these are extremely rare and hardly ever encountered.
6941 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6947 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6948 results on some platforms, including x86_64. This bug occurs at random
6949 with a very low probability, and is not known to be exploitable in any
6950 way, though its exact impact is difficult to determine. Thanks to Pieter
6951 Wuille (Blockstream) who reported this issue and also suggested an initial
6952 fix. Further analysis was conducted by the OpenSSL development team and
6953 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6954 the OpenSSL core team.
6959 * Fix various certificate fingerprint issues.
6961 By using non-DER or invalid encodings outside the signed portion of a
6962 certificate the fingerprint can be changed without breaking the signature.
6963 Although no details of the signed portion of the certificate can be changed
6964 this can cause problems with some applications: e.g. those using the
6965 certificate fingerprint for blacklists.
6967 1. Reject signatures with non zero unused bits.
6969 If the BIT STRING containing the signature has non zero unused bits reject
6970 the signature. All current signature algorithms require zero unused bits.
6972 2. Check certificate algorithm consistency.
6974 Check the AlgorithmIdentifier inside TBS matches the one in the
6975 certificate signature. NB: this will result in signature failure
6976 errors for some broken certificates.
6978 Thanks to Konrad Kraszewski from Google for reporting this issue.
6980 3. Check DSA/ECDSA signatures use DER.
6982 Reencode DSA/ECDSA signatures and compare with the original received
6983 signature. Return an error if there is a mismatch.
6985 This will reject various cases including garbage after signature
6986 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6987 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6988 (negative or with leading zeroes).
6990 Further analysis was conducted and fixes were developed by Stephen Henson
6991 of the OpenSSL core team.
6997 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014] ###
6999 * Session Ticket Memory Leak.
7001 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7002 integrity of that ticket is first verified. In the event of a session
7003 ticket integrity check failing, OpenSSL will fail to free memory
7004 causing a memory leak. By sending a large number of invalid session
7005 tickets an attacker could exploit this issue in a Denial Of Service
7011 * Build option no-ssl3 is incomplete.
7013 When OpenSSL is configured with "no-ssl3" as a build option, servers
7014 could accept and complete a SSL 3.0 handshake, and clients could be
7015 configured to send them.
7018 *Akamai and the OpenSSL team*
7020 * Add support for TLS_FALLBACK_SCSV.
7021 Client applications doing fallback retries should call
7022 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7025 *Adam Langley, Bodo Moeller*
7027 * Add additional DigestInfo checks.
7029 Reencode DigestInto in DER and check against the original when
7030 verifying RSA signature: this will reject any improperly encoded
7031 DigestInfo structures.
7033 Note: this is a precautionary measure and no attacks are currently known.
7038 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014] ###
7040 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7041 to a denial of service attack. A malicious server can crash the client
7042 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7043 ciphersuite and sending carefully crafted handshake messages.
7045 Thanks to Felix Gröbert (Google) for discovering and researching this
7051 * By sending carefully crafted DTLS packets an attacker could cause openssl
7052 to leak memory. This can be exploited through a Denial of Service attack.
7053 Thanks to Adam Langley for discovering and researching this issue.
7058 * An attacker can force openssl to consume large amounts of memory whilst
7059 processing DTLS handshake messages. This can be exploited through a
7060 Denial of Service attack.
7061 Thanks to Adam Langley for discovering and researching this issue.
7066 * An attacker can force an error condition which causes openssl to crash
7067 whilst processing DTLS packets due to memory being freed twice. This
7068 can be exploited through a Denial of Service attack.
7069 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7075 * If a multithreaded client connects to a malicious server using a resumed
7076 session and the server sends an ec point format extension it could write
7077 up to 255 bytes to freed memory.
7079 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7085 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7086 X509_name_oneline, X509_name_print_ex et al. to leak some information
7087 from the stack. Applications may be affected if they echo pretty printing
7088 output to the attacker.
7090 Thanks to Ivan Fratric (Google) for discovering this issue.
7093 *Emilia Käsper, and Steve Henson*
7095 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7096 for corner cases. (Certain input points at infinity could lead to
7097 bogus results, with non-infinity inputs mapped to infinity too.)
7101 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014] ###
7103 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7104 handshake can force the use of weak keying material in OpenSSL
7105 SSL/TLS clients and servers.
7107 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7108 researching this issue. [CVE-2014-0224][]
7110 *KIKUCHI Masashi, Steve Henson*
7112 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7113 OpenSSL DTLS client the code can be made to recurse eventually crashing
7116 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7119 *Imre Rad, Steve Henson*
7121 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7122 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7123 client or server. This is potentially exploitable to run arbitrary
7124 code on a vulnerable client or server.
7126 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7128 *Jüri Aedla, Steve Henson*
7130 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7131 are subject to a denial of service attack.
7133 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7134 this issue. [CVE-2014-3470][]
7136 *Felix Gröbert, Ivan Fratric, Steve Henson*
7138 * Harmonize version and its documentation. -f flag is used to display
7141 *mancha <mancha1@zoho.com>*
7143 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7144 in i2d_ECPrivateKey.
7146 *mancha <mancha1@zoho.com>*
7148 * Fix some double frees. These are not thought to be exploitable.
7150 *mancha <mancha1@zoho.com>*
7152 * Fix for the attack described in the paper "Recovering OpenSSL
7153 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7154 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7155 http://eprint.iacr.org/2014/140
7157 Thanks to Yuval Yarom and Naomi Benger for discovering this
7158 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7160 *Yuval Yarom and Naomi Benger*
7162 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014] ###
7164 * Keep original DTLS digest and encryption contexts in retransmission
7165 structures so we can use the previous session parameters if they need
7166 to be resent. [CVE-2013-6450][]
7170 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7171 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7172 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7173 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7174 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7175 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7177 *Rob Stradling, Adam Langley*
7179 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013] ###
7181 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7183 This addresses the flaw in CBC record processing discovered by
7184 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7185 at: http://www.isg.rhul.ac.uk/tls/
7187 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7188 Security Group at Royal Holloway, University of London
7189 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7190 Emilia Käsper for the initial patch.
7193 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7195 * Return an error when checking OCSP signatures when key is NULL.
7196 This fixes a DoS attack. [CVE-2013-0166][]
7200 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7201 the right response is stapled. Also change SSL_get_certificate()
7202 so it returns the certificate actually sent.
7203 See http://rt.openssl.org/Ticket/Display.html?id=2836.
7204 (This is a backport)
7206 *Rob Stradling <rob.stradling@comodo.com>*
7208 * Fix possible deadlock when decoding public keys.
7212 ### Changes between 1.0.0i and 1.0.0j [10 May 2012] ###
7214 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7217 * Sanity check record length before skipping explicit IV in DTLS
7220 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7221 fuzzing as a service testing platform.
7226 * Initialise tkeylen properly when encrypting CMS messages.
7227 Thanks to Solar Designer of Openwall for reporting this issue.
7231 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012] ###
7233 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7234 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7235 in CRYPTO_realloc_clean.
7237 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7238 issue and to Adam Langley <agl@chromium.org> for fixing it.
7241 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7243 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012] ###
7245 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7246 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7247 content decryption and always return the same error. Note: this attack
7248 needs on average 2^20 messages so it only affects automated senders. The
7249 old behaviour can be re-enabled in the CMS code by setting the
7250 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7251 an MMA defence is not necessary.
7252 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7253 this issue. [CVE-2012-0884][]
7257 * Fix CVE-2011-4619: make sure we really are receiving a
7258 client hello before rejecting multiple SGC restarts. Thanks to
7259 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7263 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012] ###
7265 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7266 Thanks to Antonio Martin, Enterprise Secure Access Research and
7267 Development, Cisco Systems, Inc. for discovering this bug and
7268 preparing a fix. [CVE-2012-0050][]
7272 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012] ###
7274 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7275 of the Vaudenay padding oracle attack on CBC mode encryption
7276 which enables an efficient plaintext recovery attack against
7277 the OpenSSL implementation of DTLS. Their attack exploits timing
7278 differences arising during decryption processing. A research
7279 paper describing this attack can be found at:
7280 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
7281 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7282 Security Group at Royal Holloway, University of London
7283 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7284 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7285 for preparing the fix. [CVE-2011-4108][]
7287 *Robin Seggelmann, Michael Tuexen*
7289 * Clear bytes used for block padding of SSL 3.0 records.
7292 *Adam Langley (Google)*
7294 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7295 Kadianakis <desnacked@gmail.com> for discovering this issue and
7296 Adam Langley for preparing the fix. [CVE-2011-4619][]
7298 *Adam Langley (Google)*
7300 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7302 *Andrey Kulikov <amdeich@gmail.com>*
7304 * Prevent malformed RFC3779 data triggering an assertion failure.
7305 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7306 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7308 *Rob Austein <sra@hactrn.net>*
7310 * Improved PRNG seeding for VOS.
7312 *Paul Green <Paul.Green@stratus.com>*
7314 * Fix ssl_ciph.c set-up race.
7316 *Adam Langley (Google)*
7318 * Fix spurious failures in ecdsatest.c.
7320 *Emilia Käsper (Google)*
7322 * Fix the BIO_f_buffer() implementation (which was mixing different
7323 interpretations of the '..._len' fields).
7325 *Adam Langley (Google)*
7327 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7328 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7329 threads won't reuse the same blinding coefficients.
7331 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7332 lock to call BN_BLINDING_invert_ex, and avoids one use of
7333 BN_BLINDING_update for each BN_BLINDING structure (previously,
7334 the last update always remained unused).
7336 *Emilia Käsper (Google)*
7338 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7340 *Bob Buckholz (Google)*
7342 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011] ###
7344 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7345 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7347 *Kaspar Brand <ossl@velox.ch>*
7349 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7350 for multi-threaded use of ECDH. [CVE-2011-3210][]
7352 *Adam Langley (Google)*
7354 * Fix x509_name_ex_d2i memory leak on bad inputs.
7358 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7359 signature public key algorithm by using OID xref utilities instead.
7360 Before this you could only use some ECC ciphersuites with SHA1 only.
7364 * Add protection against ECDSA timing attacks as mentioned in the paper
7365 by Billy Bob Brumley and Nicola Tuveri, see:
7367 http://eprint.iacr.org/2011/232.pdf
7370 *Billy Bob Brumley and Nicola Tuveri*
7372 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011] ###
7374 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7376 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7378 * Fix bug in string printing code: if *any* escaping is enabled we must
7379 escape the escape character (backslash) or the resulting string is
7384 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010] ###
7386 * Disable code workaround for ancient and obsolete Netscape browsers
7387 and servers: an attacker can use it in a ciphersuite downgrade attack.
7388 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7392 * Fixed J-PAKE implementation error, originally discovered by
7393 Sebastien Martini, further info and confirmation from Stefan
7394 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7398 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010] ###
7400 * Fix extension code to avoid race conditions which can result in a buffer
7401 overrun vulnerability: resumed sessions must not be modified as they can
7402 be shared by multiple threads. CVE-2010-3864
7406 * Fix WIN32 build system to correctly link an ENGINE directory into
7411 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010] ###
7413 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7416 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7418 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010] ###
7420 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7421 context. The operation can be customised via the ctrl mechanism in
7422 case ENGINEs want to include additional functionality.
7426 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7430 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7431 output hashes compatible with older versions of OpenSSL.
7433 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7435 * Fix compression algorithm handling: if resuming a session use the
7436 compression algorithm of the resumed session instead of determining
7437 it from client hello again. Don't allow server to change algorithm.
7441 * Add load_crls() function to apps tidying load_certs() too. Add option
7442 to verify utility to allow additional CRLs to be included.
7446 * Update OCSP request code to permit adding custom headers to the request:
7447 some responders need this.
7451 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7454 *Julia Lawall <julia@diku.dk>*
7456 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7457 needlessly dereferenced structures, used obsolete functions and
7458 didn't handle all updated verify codes correctly.
7462 * Disable MD2 in the default configuration.
7466 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7467 indicate the initial BIO being pushed or popped. This makes it possible
7468 to determine whether the BIO is the one explicitly called or as a result
7469 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7470 it handles reference counts correctly and doesn't zero out the I/O bio
7471 when it is not being explicitly popped. WARNING: applications which
7472 included workarounds for the old buggy behaviour will need to be modified
7473 or they could free up already freed BIOs.
7477 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7478 renaming to all platforms (within the 0.9.8 branch, this was
7479 done conditionally on Netware platforms to avoid a name clash).
7481 *Guenter <lists@gknw.net>*
7483 * Add ECDHE and PSK support to DTLS.
7485 *Michael Tuexen <tuexen@fh-muenster.de>*
7487 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7492 * Add "missing" function EVP_MD_flags() (without this the only way to
7493 retrieve a digest flags is by accessing the structure directly. Update
7494 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
7495 or cipher is registered as in the "from" argument. Print out all
7496 registered digests in the dgst usage message instead of manually
7497 attempting to work them out.
7501 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7502 this allows the use of compression and extensions. Change default cipher
7503 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7504 by default unless an application cipher string requests it.
7508 * Alter match criteria in PKCS12_parse(). It used to try to use local
7509 key ids to find matching certificates and keys but some PKCS#12 files
7510 don't follow the (somewhat unwritten) rules and this strategy fails.
7511 Now just gather all certificates together and the first private key
7512 then look for the first certificate that matches the key.
7516 * Support use of registered digest and cipher names for dgst and cipher
7517 commands instead of having to add each one as a special case. So now
7524 openssl dgst -sha256 foo
7526 and this works for ENGINE based algorithms too.
7531 * Update Gost ENGINE to support parameter files.
7533 *Victor B. Wagner <vitus@cryptocom.ru>*
7535 * Support GeneralizedTime in ca utility.
7537 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7539 * Enhance the hash format used for certificate directory links. The new
7540 form uses the canonical encoding (meaning equivalent names will work
7541 even if they aren't identical) and uses SHA1 instead of MD5. This form
7542 is incompatible with the older format and as a result c_rehash should
7543 be used to rebuild symbolic links.
7547 * Make PKCS#8 the default write format for private keys, replacing the
7548 traditional format. This form is standardised, more secure and doesn't
7549 include an implicit MD5 dependency.
7553 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7554 committed to OpenSSL should pass this lot as a minimum.
7558 * Add session ticket override functionality for use by EAP-FAST.
7560 *Jouni Malinen <j@w1.fi>*
7562 * Modify HMAC functions to return a value. Since these can be implemented
7563 in an ENGINE errors can occur.
7567 * Type-checked OBJ_bsearch_ex.
7571 * Type-checked OBJ_bsearch. Also some constification necessitated
7572 by type-checking. Still to come: TXT_DB, bsearch(?),
7573 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7578 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7579 seconds to a tm structure directly, instead of going through OS
7580 specific date routines. This avoids any issues with OS routines such
7581 as the year 2038 bug. New *_adj() functions for ASN1 time structures
7582 and X509_time_adj_ex() to cover the extended range. The existing
7583 X509_time_adj() is still usable and will no longer have any date issues.
7587 * Delta CRL support. New use deltas option which will attempt to locate
7588 and search any appropriate delta CRLs available.
7590 This work was sponsored by Google.
7594 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7595 code and add additional score elements. Validate alternate CRL paths
7596 as part of the CRL checking and indicate a new error "CRL path validation
7597 error" in this case. Applications wanting additional details can use
7598 the verify callback and check the new "parent" field. If this is not
7599 NULL CRL path validation is taking place. Existing applications won't
7600 see this because it requires extended CRL support which is off by
7603 This work was sponsored by Google.
7607 * Support for freshest CRL extension.
7609 This work was sponsored by Google.
7613 * Initial indirect CRL support. Currently only supported in the CRLs
7614 passed directly and not via lookup. Process certificate issuer
7615 CRL entry extension and lookup CRL entries by bother issuer name
7616 and serial number. Check and process CRL issuer entry in IDP extension.
7618 This work was sponsored by Google.
7622 * Add support for distinct certificate and CRL paths. The CRL issuer
7623 certificate is validated separately in this case. Only enabled if
7624 an extended CRL support flag is set: this flag will enable additional
7625 CRL functionality in future.
7627 This work was sponsored by Google.
7631 * Add support for policy mappings extension.
7633 This work was sponsored by Google.
7637 * Fixes to pathlength constraint, self issued certificate handling,
7638 policy processing to align with RFC3280 and PKITS tests.
7640 This work was sponsored by Google.
7644 * Support for name constraints certificate extension. DN, email, DNS
7645 and URI types are currently supported.
7647 This work was sponsored by Google.
7651 * To cater for systems that provide a pointer-based thread ID rather
7652 than numeric, deprecate the current numeric thread ID mechanism and
7653 replace it with a structure and associated callback type. This
7654 mechanism allows a numeric "hash" to be extracted from a thread ID in
7655 either case, and on platforms where pointers are larger than 'long',
7656 mixing is done to help ensure the numeric 'hash' is usable even if it
7657 can't be guaranteed unique. The default mechanism is to use "&errno"
7658 as a pointer-based thread ID to distinguish between threads.
7660 Applications that want to provide their own thread IDs should now use
7661 CRYPTO_THREADID_set_callback() to register a callback that will call
7662 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7664 Note that ERR_remove_state() is now deprecated, because it is tied
7665 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7666 to free the current thread's error state should be replaced by
7667 ERR_remove_thread_state(NULL).
7669 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7670 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7671 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7672 application was previously providing a numeric thread callback that
7673 was inappropriate for distinguishing threads, then uniqueness might
7674 have been obtained with &errno that happened immediately in the
7675 intermediate development versions of OpenSSL; this is no longer the
7676 case, the numeric thread callback will now override the automatic use
7679 *Geoff Thorpe, with help from Bodo Moeller*
7681 * Initial support for different CRL issuing certificates. This covers a
7682 simple case where the self issued certificates in the chain exist and
7683 the real CRL issuer is higher in the existing chain.
7685 This work was sponsored by Google.
7689 * Removed effectively defunct crypto/store from the build.
7693 * Revamp of STACK to provide stronger type-checking. Still to come:
7694 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7695 ASN1_STRING, CONF_VALUE.
7699 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7700 RAM on SSL connections. This option can save about 34k per idle SSL.
7704 * Revamp of LHASH to provide stronger type-checking. Still to come:
7705 STACK, TXT_DB, bsearch, qsort.
7709 * Initial support for Cryptographic Message Syntax (aka CMS) based
7710 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7711 support for data, signedData, compressedData, digestedData and
7712 encryptedData, envelopedData types included. Scripts to check against
7713 RFC4134 examples draft and interop and consistency checks of many
7714 content types and variants.
7718 * Add options to enc utility to support use of zlib compression BIO.
7722 * Extend mk1mf to support importing of options and assembly language
7723 files from Configure script, currently only included in VC-WIN32.
7724 The assembly language rules can now optionally generate the source
7725 files from the associated perl scripts.
7729 * Implement remaining functionality needed to support GOST ciphersuites.
7730 Interop testing has been performed using CryptoPro implementations.
7732 *Victor B. Wagner <vitus@cryptocom.ru>*
7734 * s390x assembler pack.
7738 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7743 * Implement Opaque PRF Input TLS extension as specified in
7744 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7745 official specification yet and no extension type assignment by
7746 IANA exists, this extension (for now) will have to be explicitly
7747 enabled when building OpenSSL by providing the extension number
7748 to use. For example, specify an option
7750 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7752 to the "config" or "Configure" script to enable the extension,
7753 assuming extension number 0x9527 (which is a completely arbitrary
7754 and unofficial assignment based on the MD5 hash of the Internet
7755 Draft). Note that by doing so, you potentially lose
7756 interoperability with other TLS implementations since these might
7757 be using the same extension number for other purposes.
7759 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7760 opaque PRF input value to use in the handshake. This will create
7761 an internal copy of the length-'len' string at 'src', and will
7762 return non-zero for success.
7764 To get more control and flexibility, provide a callback function
7767 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7768 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7772 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7775 Callback function 'cb' will be called in handshakes, and is
7776 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7777 Argument 'arg' is for application purposes (the value as given to
7778 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7779 be provided to the callback function). The callback function
7780 has to return non-zero to report success: usually 1 to use opaque
7781 PRF input just if possible, or 2 to enforce use of the opaque PRF
7782 input. In the latter case, the library will abort the handshake
7783 if opaque PRF input is not successfully negotiated.
7785 Arguments 'peerinput' and 'len' given to the callback function
7786 will always be NULL and 0 in the case of a client. A server will
7787 see the client's opaque PRF input through these variables if
7788 available (NULL and 0 otherwise). Note that if the server
7789 provides an opaque PRF input, the length must be the same as the
7790 length of the client's opaque PRF input.
7792 Note that the callback function will only be called when creating
7793 a new session (session resumption can resume whatever was
7794 previously negotiated), and will not be called in SSL 2.0
7795 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7796 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7797 for applications that need to enforce opaque PRF input.
7802 * Update ssl code to support digests other than SHA1+MD5 for handshake
7806 *Victor B. Wagner <vitus@cryptocom.ru>*
7808 * Add RFC4507 support to OpenSSL. This includes the corrections in
7809 RFC4507bis. The encrypted ticket format is an encrypted encoded
7810 SSL_SESSION structure, that way new session features are automatically
7813 If a client application caches session in an SSL_SESSION structure
7814 support is transparent because tickets are now stored in the encoded
7817 The SSL_CTX structure automatically generates keys for ticket
7818 protection in servers so again support should be possible
7819 with no application modification.
7821 If a client or server wishes to disable RFC4507 support then the option
7822 SSL_OP_NO_TICKET can be set.
7824 Add a TLS extension debugging callback to allow the contents of any client
7825 or server extensions to be examined.
7827 This work was sponsored by Google.
7831 * Final changes to avoid use of pointer pointer casts in OpenSSL.
7832 OpenSSL should now compile cleanly on gcc 4.2
7834 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
7836 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
7837 support including streaming MAC support: this is required for GOST
7838 ciphersuite support.
7840 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
7842 * Add option -stream to use PKCS#7 streaming in smime utility. New
7843 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
7844 to output in BER and PEM format.
7848 * Experimental support for use of HMAC via EVP_PKEY interface. This
7849 allows HMAC to be handled via the EVP_DigestSign*() interface. The
7850 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
7851 ENGINE support for HMAC keys which are unextractable. New -mac and
7852 -macopt options to dgst utility.
7856 * New option -sigopt to dgst utility. Update dgst to use
7857 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
7858 alternative signing parameters such as X9.31 or PSS in the dgst
7863 * Change ssl_cipher_apply_rule(), the internal function that does
7864 the work each time a ciphersuite string requests enabling
7865 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
7866 removing ("!foo+bar") a class of ciphersuites: Now it maintains
7867 the order of disabled ciphersuites such that those ciphersuites
7868 that most recently went from enabled to disabled not only stay
7869 in order with respect to each other, but also have higher priority
7870 than other disabled ciphersuites the next time ciphersuites are
7873 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
7874 the same ciphersuites as with "HIGH" alone, but in a specific
7875 order where the PSK ciphersuites come first (since they are the
7876 most recently disabled ciphersuites when "HIGH" is parsed).
7878 Also, change ssl_create_cipher_list() (using this new
7879 functionality) such that between otherwise identical
7880 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
7885 * Change ssl_create_cipher_list() so that it automatically
7886 arranges the ciphersuites in reasonable order before starting
7887 to process the rule string. Thus, the definition for "DEFAULT"
7888 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
7889 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
7890 This makes it much easier to arrive at a reasonable default order
7891 in applications for which anonymous ciphers are OK (meaning
7892 that you can't actually use DEFAULT).
7894 *Bodo Moeller; suggested by Victor Duchovni*
7896 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
7897 processing) into multiple integers instead of setting
7898 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
7899 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
7900 (These masks as well as the individual bit definitions are hidden
7901 away into the non-exported interface ssl/ssl_locl.h, so this
7902 change to the definition of the SSL_CIPHER structure shouldn't
7903 affect applications.) This give us more bits for each of these
7904 categories, so there is no longer a need to coagulate AES128 and
7905 AES256 into a single algorithm bit, and to coagulate Camellia128
7906 and Camellia256 into a single algorithm bit, which has led to all
7909 Thus, among other things, the kludge introduced in 0.9.7m and
7910 0.9.8e for masking out AES256 independently of AES128 or masking
7911 out Camellia256 independently of AES256 is not needed here in 0.9.9.
7913 With the change, we also introduce new ciphersuite aliases that
7914 so far were missing: "AES128", "AES256", "CAMELLIA128", and
7919 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
7920 Use the leftmost N bytes of the signature input if the input is
7921 larger than the prime q (with N being the size in bytes of q).
7925 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
7926 it yet and it is largely untested.
7930 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
7934 * Initial incomplete changes to avoid need for function casts in OpenSSL
7935 some compilers (gcc 4.2 and later) reject their use. Safestack is
7936 reimplemented. Update ASN1 to avoid use of legacy functions.
7940 * Win32/64 targets are linked with Winsock2.
7944 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
7945 to external functions. This can be used to increase CRL handling
7946 efficiency especially when CRLs are very large by (for example) storing
7947 the CRL revoked certificates in a database.
7951 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
7952 new CRLs added to a directory can be used. New command line option
7953 -verify_return_error to s_client and s_server. This causes real errors
7954 to be returned by the verify callback instead of carrying on no matter
7955 what. This reflects the way a "real world" verify callback would behave.
7959 * GOST engine, supporting several GOST algorithms and public key formats.
7960 Kindly donated by Cryptocom.
7964 * Partial support for Issuing Distribution Point CRL extension. CRLs
7965 partitioned by DP are handled but no indirect CRL or reason partitioning
7966 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
7967 selected via a scoring technique which handles IDP and AKID in CRLs.
7971 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
7972 will ultimately be used for all verify operations: this will remove the
7973 X509_STORE dependency on certificate verification and allow alternative
7974 lookup methods. X509_STORE based implementations of these two callbacks.
7978 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
7979 Modify get_crl() to find a valid (unexpired) CRL if possible.
7983 * New function X509_CRL_match() to check if two CRLs are identical. Normally
7984 this would be called X509_CRL_cmp() but that name is already used by
7985 a function that just compares CRL issuer names. Cache several CRL
7986 extensions in X509_CRL structure and cache CRLDP in X509.
7990 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
7991 this maps equivalent X509_NAME structures into a consistent structure.
7992 Name comparison can then be performed rapidly using memcmp().
7996 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8001 * Allow digests to supply their own micalg string for S/MIME type using
8002 the ctrl EVP_MD_CTRL_MICALG.
8006 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8007 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8008 ctrl. It can then customise the structure before and/or after signing
8013 * New function OBJ_add_sigid() to allow application defined signature OIDs
8014 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8015 to free up any added signature OIDs.
8019 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8020 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8021 digest and cipher tables. New options added to openssl utility:
8022 list-message-digest-algorithms and list-cipher-algorithms.
8026 * Change the array representation of binary polynomials: the list
8027 of degrees of non-zero coefficients is now terminated with -1.
8028 Previously it was terminated with 0, which was also part of the
8029 value; thus, the array representation was not applicable to
8030 polynomials where t^0 has coefficient zero. This change makes
8031 the array representation useful in a more general context.
8035 * Various modifications and fixes to SSL/TLS cipher string
8036 handling. For ECC, the code now distinguishes between fixed ECDH
8037 with RSA certificates on the one hand and with ECDSA certificates
8038 on the other hand, since these are separate ciphersuites. The
8039 unused code for Fortezza ciphersuites has been removed.
8041 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8042 (not "ECDHE"). For consistency with the code for DH
8043 certificates, use of ECDH certificates is now considered ECDH
8044 authentication, not RSA or ECDSA authentication (the latter is
8045 merely the CA's signing algorithm and not actively used in the
8048 The temporary ciphersuite alias "ECCdraft" is no longer
8049 available, and ECC ciphersuites are no longer excluded from "ALL"
8050 and "DEFAULT". The following aliases now exist for RFC 4492
8051 ciphersuites, most of these by analogy with the DH case:
8053 kECDHr - ECDH cert, signed with RSA
8054 kECDHe - ECDH cert, signed with ECDSA
8055 kECDH - ECDH cert (signed with either RSA or ECDSA)
8056 kEECDH - ephemeral ECDH
8057 ECDH - ECDH cert or ephemeral ECDH
8063 AECDH - anonymous ECDH
8064 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8069 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8070 Use correct micalg parameters depending on digest(s) in signed message.
8074 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8075 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8079 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8080 an engine to register a method. Add ENGINE lookups for methods and
8081 functional reference processing.
8085 * New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
8086 EVP_{Sign,Verify}* which allow an application to customise the signature
8091 * New -resign option to smime utility. This adds one or more signers
8092 to an existing PKCS#7 signedData structure. Also -md option to use an
8093 alternative message digest algorithm for signing.
8097 * Tidy up PKCS#7 routines and add new functions to make it easier to
8098 create PKCS7 structures containing multiple signers. Update smime
8099 application to support multiple signers.
8103 * New -macalg option to pkcs12 utility to allow setting of an alternative
8108 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8109 Reorganize PBE internals to lookup from a static table using NIDs,
8110 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8111 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8112 PRF which will be automatically used with PBES2.
8116 * Replace the algorithm specific calls to generate keys in "req" with the
8121 * Update PKCS#7 enveloped data routines to use new API. This is now
8122 supported by any public key method supporting the encrypt operation. A
8123 ctrl is added to allow the public key algorithm to examine or modify
8124 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8129 * Add a ctrl to asn1 method to allow a public key algorithm to express
8130 a default digest type to use. In most cases this will be SHA1 but some
8131 algorithms (such as GOST) need to specify an alternative digest. The
8132 return value indicates how strong the preference is 1 means optional and
8133 2 is mandatory (that is it is the only supported type). Modify
8134 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8135 use the default md. Update openssl utilities to use the default digest
8136 type for signing if it is not explicitly indicated.
8140 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8141 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8142 signing method from the key type. This effectively removes the link
8143 between digests and public key types.
8147 * Add an OID cross reference table and utility functions. Its purpose is to
8148 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8149 rsaEncryption. This will allow some of the algorithm specific hackery
8150 needed to use the correct OID to be removed.
8154 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8155 structures for PKCS7_sign(). They are now set up by the relevant public
8160 * Add provisional EC pkey method with support for ECDSA and ECDH.
8164 * Add support for key derivation (agreement) in the API, DH method and
8169 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8170 public and private key formats. As a side effect these add additional
8171 command line functionality not previously available: DSA signatures can be
8172 generated and verified using pkeyutl and DH key support and generation in
8179 *Oliver Tappe <zooey@hirschkaefer.de>*
8181 * New make target "install_html_docs" installs HTML renditions of the
8184 *Oliver Tappe <zooey@hirschkaefer.de>*
8186 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8187 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8188 support key and parameter generation and add initial key generation
8189 functionality for RSA.
8193 * Add functions for main EVP_PKEY_method operations. The undocumented
8194 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
8195 EVP_PKEY_{encrypt,decrypt}_old.
8199 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8200 key API, doesn't do much yet.
8204 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8205 public key algorithms. New option to openssl utility:
8206 "list-public-key-algorithms" to print out info.
8210 * Implement the Supported Elliptic Curves Extension for
8211 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8215 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8216 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8220 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8221 utilities such as rsa, dsa, dsaparam etc except they process any key
8226 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8227 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8228 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8233 * Initial support for pluggable public key ASN1.
8234 De-spaghettify the public key ASN1 handling. Move public and private
8235 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8236 algorithm specific handling to a single module within the relevant
8237 algorithm directory. Add functions to allow (near) opaque processing
8238 of public and private key structures.
8242 * Implement the Supported Point Formats Extension for
8243 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8247 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8248 for the psk identity [hint] and the psk callback functions to the
8249 SSL_SESSION, SSL and SSL_CTX structure.
8252 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8256 SSL_CTX_use_psk_identity_hint
8257 SSL_get_psk_identity_hint
8258 SSL_get_psk_identity
8259 SSL_use_psk_identity_hint
8262 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8264 * Add RFC 3161 compliant time stamp request creation, response generation
8265 and response verification functionality.
8267 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8269 * Add initial support for TLS extensions, specifically for the server_name
8270 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8271 have new members for a host name. The SSL data structure has an
8272 additional member SSL_CTX *initial_ctx so that new sessions can be
8273 stored in that context to allow for session resumption, even after the
8274 SSL has been switched to a new SSL_CTX in reaction to a client's
8275 server_name extension.
8277 New functions (subject to change):
8279 SSL_get_servername()
8280 SSL_get_servername_type()
8283 New CTRL codes and macros (subject to change):
8285 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8286 - SSL_CTX_set_tlsext_servername_callback()
8287 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8288 - SSL_CTX_set_tlsext_servername_arg()
8289 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8291 openssl s_client has a new '-servername ...' option.
8293 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8294 '-key2 ...', '-servername_fatal' (subject to change). This allows
8295 testing the HostName extension for a specific single host name ('-cert'
8296 and '-key' remain fallbacks for handshakes without HostName
8297 negotiation). If the unrecognized_name alert has to be sent, this by
8298 default is a warning; it becomes fatal with the '-servername_fatal'
8302 *Peter Sylvester, Remy Allais, Christophe Renou*
8304 * Whirlpool hash implementation is added.
8308 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8309 bn(64,32). Because of instruction set limitations it doesn't have
8310 any negative impact on performance. This was done mostly in order
8311 to make it possible to share assembler modules, such as bn_mul_mont
8312 implementations, between 32- and 64-bit builds without hassle.
8316 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8317 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8322 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8323 dedicated Montgomery multiplication procedure, is introduced.
8324 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8325 "64-bit" performance on certain 32-bit targets.
8329 * New option SSL_OP_NO_COMP to disable use of compression selectively
8330 in SSL structures. New SSL ctrl to set maximum send fragment size.
8331 Save memory by setting the I/O buffer sizes dynamically instead of
8332 using the maximum available value.
8336 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8337 in addition to the text details.
8341 * Very, very preliminary EXPERIMENTAL support for printing of general
8342 ASN1 structures. This currently produces rather ugly output and doesn't
8343 handle several customised structures at all.
8347 * Integrated support for PVK file format and some related formats such
8348 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8349 these in the 'rsa' and 'dsa' utilities.
8353 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8357 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8358 place for the (very old) "NETSCAPE" format certificates which are now
8359 handled using new ASN1 code equivalents.
8363 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8364 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8365 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8369 * Modify CRL distribution points extension code to print out previously
8370 unsupported fields. Enhance extension setting code to allow setting of
8375 * Add print and set support for Issuing Distribution Point CRL extension.
8379 * Change 'Configure' script to enable Camellia by default.
8386 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010] ###
8388 * When rejecting SSL/TLS records due to an incorrect version number, never
8389 update s->server with a new major version number. As of
8390 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8391 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8392 the previous behavior could result in a read attempt at NULL when
8393 receiving specific incorrect SSL/TLS records once record payload
8394 protection is active. [CVE-2010-0740][]
8396 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8398 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8399 could be crashed if the relevant tables were not present (e.g. chrooted).
8401 *Tomas Hoger <thoger@redhat.com>*
8403 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010] ###
8405 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8407 *Martin Olsson, Neel Mehta*
8409 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8410 accommodate for stack sorting, always a write lock!).
8414 * On some versions of WIN32 Heap32Next is very slow. This can cause
8415 excessive delays in the RAND_poll(): over a minute. As a workaround
8416 include a time check in the inner Heap32Next loop too.
8420 * The code that handled flushing of data in SSL/TLS originally used the
8421 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8422 the problem outlined in PR#1949. The fix suggested there however can
8423 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8424 of Apache). So instead simplify the code to flush unconditionally.
8425 This should be fine since flushing with no data to flush is a no op.
8429 * Handle TLS versions 2.0 and later properly and correctly use the
8430 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8431 off ancient servers have a habit of sticking around for a while...
8435 * Modify compression code so it frees up structures without using the
8436 ex_data callbacks. This works around a problem where some applications
8437 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8438 restarting) then use compression (e.g. SSL with compression) later.
8439 This results in significant per-connection memory leaks and
8440 has caused some security issues including CVE-2008-1678 and
8445 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8446 change when encrypting or decrypting.
8450 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8451 connect and renegotiate with servers which do not support RI.
8452 Until RI is more widely deployed this option is enabled by default.
8456 * Add "missing" ssl ctrls to clear options and mode.
8460 * If client attempts to renegotiate and doesn't support RI respond with
8461 a no_renegotiation alert as required by RFC5746. Some renegotiating
8462 TLS clients will continue a connection gracefully when they receive
8463 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8464 waiting for a server hello which it will never receive. Now we treat a
8465 received no_renegotiation alert as a fatal error. This is because
8466 applications requesting a renegotiation might well expect it to succeed
8467 and would have no code in place to handle the server denying it so the
8468 only safe thing to do is to terminate the connection.
8472 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8473 peer supports secure renegotiation and 0 otherwise. Print out peer
8474 renegotiation support in s_client/s_server.
8478 * Replace the highly broken and deprecated SPKAC certification method with
8479 the updated NID creation version. This should correctly handle UTF8.
8483 * Implement RFC5746. Re-enable renegotiation but require the extension
8484 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8485 turns out to be a bad idea. It has been replaced by
8486 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8487 SSL_CTX_set_options(). This is really not recommended unless you
8488 know what you are doing.
8490 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8492 * Fixes to stateless session resumption handling. Use initial_ctx when
8493 issuing and attempting to decrypt tickets in case it has changed during
8494 servername handling. Use a non-zero length session ID when attempting
8495 stateless session resumption: this makes it possible to determine if
8496 a resumption has occurred immediately after receiving server hello
8497 (several places in OpenSSL subtly assume this) instead of later in
8502 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8503 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8504 fixes for a few places where the return code is not checked
8507 *Julia Lawall <julia@diku.dk>*
8509 * Add --strict-warnings option to Configure script to include devteam
8510 warnings in other configurations.
8514 * Add support for --libdir option and LIBDIR variable in makefiles. This
8515 makes it possible to install openssl libraries in locations which
8516 have names other than "lib", for example "/usr/lib64" which some
8519 *Steve Henson, based on patch from Jeremy Utley*
8521 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8522 X690 8.9.12 and can produce some misleading textual output of OIDs.
8524 *Steve Henson, reported by Dan Kaminsky*
8526 * Delete MD2 from algorithm tables. This follows the recommendation in
8527 several standards that it is not used in new applications due to
8528 several cryptographic weaknesses. For binary compatibility reasons
8529 the MD2 API is still compiled in by default.
8533 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8538 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8539 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8542 *Guenter <lists@gknw.net>*
8544 * Fix the server certificate chain building code to use X509_verify_cert(),
8545 it used to have an ad-hoc builder which was unable to cope with anything
8546 other than a simple chain.
8548 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8550 * Don't check self signed certificate signatures in X509_verify_cert()
8551 by default (a flag can override this): it just wastes time without
8552 adding any security. As a useful side effect self signed root CAs
8553 with non-FIPS digests are now usable in FIPS mode.
8557 * In dtls1_process_out_of_seq_message() the check if the current message
8558 is already buffered was missing. For every new message was memory
8559 allocated, allowing an attacker to perform an denial of service attack
8560 with sending out of seq handshake messages until there is no memory
8561 left. Additionally every future message was buffered, even if the
8562 sequence number made no sense and would be part of another handshake.
8563 So only messages with sequence numbers less than 10 in advance will be
8564 buffered. [CVE-2009-1378][]
8566 *Robin Seggelmann, discovered by Daniel Mentz*
8568 * Records are buffered if they arrive with a future epoch to be
8569 processed after finishing the corresponding handshake. There is
8570 currently no limitation to this buffer allowing an attacker to perform
8571 a DOS attack with sending records with future epochs until there is no
8572 memory left. This patch adds the pqueue_size() function to determine
8573 the size of a buffer and limits the record buffer to 100 entries.
8576 *Robin Seggelmann, discovered by Daniel Mentz*
8578 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8579 parent structure is freed. [CVE-2009-1379][]
8583 * Handle non-blocking I/O properly in SSL_shutdown() call.
8585 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8589 *Ilya O. <vrghost@gmail.com>*
8591 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009] ###
8593 * Disable renegotiation completely - this fixes a severe security
8594 problem [CVE-2009-3555][] at the cost of breaking all
8595 renegotiation. Renegotiation can be re-enabled by setting
8596 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8597 run-time. This is really not recommended unless you know what
8602 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009] ###
8604 * Don't set val to NULL when freeing up structures, it is freed up by
8605 underlying code. If sizeof(void *) > sizeof(long) this can result in
8606 zeroing past the valid field. [CVE-2009-0789][]
8608 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8610 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8611 checked correctly. This would allow some invalid signed attributes to
8612 appear to verify correctly. [CVE-2009-0591][]
8614 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8616 * Reject UniversalString and BMPString types with invalid lengths. This
8617 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8618 a legal length. [CVE-2009-0590][]
8622 * Set S/MIME signing as the default purpose rather than setting it
8623 unconditionally. This allows applications to override it at the store
8628 * Permit restricted recursion of ASN1 strings. This is needed in practice
8629 to handle some structures.
8633 * Improve efficiency of mem_gets: don't search whole buffer each time
8636 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8638 * New -hex option for openssl rand.
8642 * Print out UTF8String and NumericString when parsing ASN1.
8646 * Support NumericString type for name components.
8650 * Allow CC in the environment to override the automatically chosen
8651 compiler. Note that nothing is done to ensure flags work with the
8656 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009] ###
8658 * Properly check EVP_VerifyFinal() and similar return values
8661 *Ben Laurie, Bodo Moeller, Google Security Team*
8663 * Enable TLS extensions by default.
8667 * Allow the CHIL engine to be loaded, whether the application is
8668 multithreaded or not. (This does not release the developer from the
8669 obligation to set up the dynamic locking callbacks.)
8671 *Sander Temme <sander@temme.net>*
8673 * Use correct exit code if there is an error in dgst command.
8675 *Steve Henson; problem pointed out by Roland Dirlewanger*
8677 * Tweak Configure so that you need to say "experimental-jpake" to enable
8678 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8682 * Add experimental JPAKE support, including demo authentication in
8683 s_client and s_server.
8687 * Set the comparison function in v3_addr_canonize().
8689 *Rob Austein <sra@hactrn.net>*
8691 * Add support for XMPP STARTTLS in s_client.
8693 *Philip Paeps <philip@freebsd.org>*
8695 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8696 to ensure that even with this option, only ciphersuites in the
8697 server's preference list will be accepted. (Note that the option
8698 applies only when resuming a session, so the earlier behavior was
8699 just about the algorithm choice for symmetric cryptography.)
8703 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008] ###
8705 * Fix NULL pointer dereference if a DTLS server received
8706 ChangeCipherSpec as first record [CVE-2009-1386][].
8710 * Fix a state transition in s3_srvr.c and d1_srvr.c
8711 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
8715 * The fix in 0.9.8c that supposedly got rid of unsafe
8716 double-checked locking was incomplete for RSA blinding,
8717 addressing just one layer of what turns out to have been
8718 doubly unsafe triple-checked locking.
8720 So now fix this for real by retiring the MONT_HELPER macro
8721 in crypto/rsa/rsa_eay.c.
8724 *Bodo Moeller; problem pointed out by Marius Schilder*
8726 * Various precautionary measures:
8728 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8730 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8731 (NB: This would require knowledge of the secret session ticket key
8732 to exploit, in which case you'd be SOL either way.)
8734 - Change bn_nist.c so that it will properly handle input BIGNUMs
8735 outside the expected range.
8737 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8741 *Neel Mehta, Bodo Moeller*
8743 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8744 the load fails. Useful for distros.
8746 *Ben Laurie and the FreeBSD team*
8748 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8752 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8756 * Expand ENGINE to support engine supplied SSL client certificate functions.
8758 This work was sponsored by Logica.
8762 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8763 keystores. Support for SSL/TLS client authentication too.
8764 Not compiled unless enable-capieng specified to Configure.
8766 This work was sponsored by Logica.
8770 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8771 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8772 attribute creation routines such as certificate requests and PKCS#12
8777 ### Changes between 0.9.8g and 0.9.8h [28 May 2008] ###
8779 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8780 handshake which could lead to a client crash as found using the
8781 Codenomicon TLS test suite [CVE-2008-1672][]
8783 *Steve Henson, Mark Cox*
8785 * Fix double free in TLS server name extensions which could lead to
8786 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8790 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8792 Clear the error queue to ensure that error entries left from
8793 older function calls do not interfere with the correct operation.
8795 *Lutz Jaenicke, Erik de Castro Lopo*
8797 * Remove root CA certificates of commercial CAs:
8799 The OpenSSL project does not recommend any specific CA and does not
8800 have any policy with respect to including or excluding any CA.
8801 Therefore it does not make any sense to ship an arbitrary selection
8802 of root CA certificates with the OpenSSL software.
8806 * RSA OAEP patches to fix two separate invalid memory reads.
8807 The first one involves inputs when 'lzero' is greater than
8808 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8809 before the beginning of from). The second one involves inputs where
8810 the 'db' section contains nothing but zeroes (there is a one-byte
8811 invalid read after the end of 'db').
8813 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8815 * Partial backport from 0.9.9-dev:
8817 Introduce bn_mul_mont (dedicated Montgomery multiplication
8818 procedure) as a candidate for BIGNUM assembler implementation.
8819 While 0.9.9-dev uses assembler for various architectures, only
8820 x86_64 is available by default here in the 0.9.8 branch, and
8821 32-bit x86 is available through a compile-time setting.
8823 To try the 32-bit x86 assembler implementation, use Configure
8824 option "enable-montasm" (which exists only for this backport).
8826 As "enable-montasm" for 32-bit x86 disclaims code stability
8827 anyway, in this constellation we activate additional code
8828 backported from 0.9.9-dev for further performance improvements,
8829 namely BN_from_montgomery_word. (To enable this otherwise,
8830 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
8833 *Andy Polyakov (backport partially by Bodo Moeller)*
8835 * Add TLS session ticket callback. This allows an application to set
8836 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
8837 values. This is useful for key rollover for example where several key
8838 sets may exist with different names.
8842 * Reverse ENGINE-internal logic for caching default ENGINE handles.
8843 This was broken until now in 0.9.8 releases, such that the only way
8844 a registered ENGINE could be used (assuming it initialises
8845 successfully on the host) was to explicitly set it as the default
8846 for the relevant algorithms. This is in contradiction with 0.9.7
8847 behaviour and the documentation. With this fix, when an ENGINE is
8848 registered into a given algorithm's table of implementations, the
8849 'uptodate' flag is reset so that auto-discovery will be used next
8850 time a new context for that algorithm attempts to select an
8853 *Ian Lister (tweaked by Geoff Thorpe)*
8855 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
8856 implementation in the following ways:
8858 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
8861 Lack of BER streaming support means one pass streaming processing is
8862 only supported if data is detached: setting the streaming flag is
8863 ignored for embedded content.
8865 CMS support is disabled by default and must be explicitly enabled
8866 with the enable-cms configuration option.
8870 * Update the GMP engine glue to do direct copies between BIGNUM and
8871 mpz_t when openssl and GMP use the same limb size. Otherwise the
8872 existing "conversion via a text string export" trick is still used.
8874 *Paul Sheer <paulsheer@gmail.com>*
8876 * Zlib compression BIO. This is a filter BIO which compressed and
8877 uncompresses any data passed through it.
8881 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
8882 RFC3394 compatible AES key wrapping.
8886 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
8887 sets string data without copying. X509_ALGOR_set0() and
8888 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
8889 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
8890 from an X509_ATTRIBUTE structure optionally checking it occurs only
8891 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
8896 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
8897 to get the expected BN_FLG_CONSTTIME behavior.
8899 *Bodo Moeller (Google)*
8903 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
8904 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
8905 - added some more tests to do_tests.pl
8906 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
8907 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
8908 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
8909 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
8910 - various changes to netware.pl to enable gcc-cross builds on Win32
8912 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
8913 - various changes to fix missing prototype warnings
8914 - fixed x86nasm.pl to create correct asm files for NASM COFF output
8915 - added AES, WHIRLPOOL and CPUID assembler code to build files
8916 - added missing AES assembler make rules to mk1mf.pl
8917 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
8919 *Guenter Knauf <eflash@gmx.net>*
8921 * Implement certificate status request TLS extension defined in RFC3546.
8922 A client can set the appropriate parameters and receive the encoded
8923 OCSP response via a callback. A server can query the supplied parameters
8924 and set the encoded OCSP response in the callback. Add simplified examples
8925 to s_client and s_server.
8929 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007] ###
8932 + Binary incompatibility of ssl_ctx_st structure
8933 + DTLS interoperation with non-compliant servers
8934 + Don't call get_session_cb() without proposed session
8935 + Fix ia64 assembler code
8937 *Andy Polyakov, Steve Henson*
8939 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007] ###
8941 * DTLS Handshake overhaul. There were longstanding issues with
8942 OpenSSL DTLS implementation, which were making it impossible for
8943 RFC 4347 compliant client to communicate with OpenSSL server.
8944 Unfortunately just fixing these incompatibilities would "cut off"
8945 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
8946 server keeps tolerating non RFC compliant syntax. The opposite is
8947 not true, 0.9.8f client can not communicate with earlier server.
8948 This update even addresses CVE-2007-4995.
8952 * Changes to avoid need for function casts in OpenSSL: some compilers
8953 (gcc 4.2 and later) reject their use.
8954 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
8957 * Add RFC4507 support to OpenSSL. This includes the corrections in
8958 RFC4507bis. The encrypted ticket format is an encrypted encoded
8959 SSL_SESSION structure, that way new session features are automatically
8962 If a client application caches session in an SSL_SESSION structure
8963 support is transparent because tickets are now stored in the encoded
8966 The SSL_CTX structure automatically generates keys for ticket
8967 protection in servers so again support should be possible
8968 with no application modification.
8970 If a client or server wishes to disable RFC4507 support then the option
8971 SSL_OP_NO_TICKET can be set.
8973 Add a TLS extension debugging callback to allow the contents of any client
8974 or server extensions to be examined.
8976 This work was sponsored by Google.
8980 * Add initial support for TLS extensions, specifically for the server_name
8981 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8982 have new members for a host name. The SSL data structure has an
8983 additional member SSL_CTX *initial_ctx so that new sessions can be
8984 stored in that context to allow for session resumption, even after the
8985 SSL has been switched to a new SSL_CTX in reaction to a client's
8986 server_name extension.
8988 New functions (subject to change):
8990 SSL_get_servername()
8991 SSL_get_servername_type()
8994 New CTRL codes and macros (subject to change):
8996 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8997 - SSL_CTX_set_tlsext_servername_callback()
8998 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8999 - SSL_CTX_set_tlsext_servername_arg()
9000 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9002 openssl s_client has a new '-servername ...' option.
9004 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9005 '-key2 ...', '-servername_fatal' (subject to change). This allows
9006 testing the HostName extension for a specific single host name ('-cert'
9007 and '-key' remain fallbacks for handshakes without HostName
9008 negotiation). If the unrecognized_name alert has to be sent, this by
9009 default is a warning; it becomes fatal with the '-servername_fatal'
9013 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9015 * Add AES and SSE2 assembly language support to VC++ build.
9019 * Mitigate attack on final subtraction in Montgomery reduction.
9023 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9024 (which previously caused an internal error).
9028 * Squeeze another 10% out of IGE mode when in != out.
9032 * AES IGE mode speedup.
9034 *Dean Gaudet (Google)*
9036 * Add the Korean symmetric 128-bit cipher SEED (see
9037 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
9038 add SEED ciphersuites from RFC 4162:
9040 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9041 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9042 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9043 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9045 To minimize changes between patchlevels in the OpenSSL 0.9.8
9046 series, SEED remains excluded from compilation unless OpenSSL
9047 is configured with 'enable-seed'.
9049 *KISA, Bodo Moeller*
9051 * Mitigate branch prediction attacks, which can be practical if a
9052 single processor is shared, allowing a spy process to extract
9053 information. For detailed background information, see
9054 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
9055 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9056 and Necessary Software Countermeasures"). The core of the change
9057 are new versions BN_div_no_branch() and
9058 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9059 respectively, which are slower, but avoid the security-relevant
9060 conditional branches. These are automatically called by BN_div()
9061 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9062 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9063 remove a conditional branch.
9065 BN_FLG_CONSTTIME is the new name for the previous
9066 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9067 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9068 in the exponent causes BN_mod_exp_mont() to use the alternative
9069 implementation in BN_mod_exp_mont_consttime().) The old name
9070 remains as a deprecated alias.
9072 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9073 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9074 constant-time implementations for more than just exponentiation.
9075 Here too the old name is kept as a deprecated alias.
9077 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9078 the BN_BLINDING structure gets an independent copy of the
9079 modulus. This means that the previous "BIGNUM *m" argument to
9080 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9081 essentially becomes "const BIGNUM *m", although we can't actually
9082 change this in the header file before 0.9.9. It allows
9083 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9084 enable BN_FLG_CONSTTIME.
9087 *Matthew D Wood (Intel Corp)*
9089 * In the SSL/TLS server implementation, be strict about session ID
9090 context matching (which matters if an application uses a single
9091 external cache for different purposes). Previously,
9092 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9093 set. This did ensure strict client verification, but meant that,
9094 with applications using a single external cache for quite
9095 different requirements, clients could circumvent ciphersuite
9096 restrictions for a given session ID context by starting a session
9097 in a different context.
9101 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9102 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9103 authentication-only ciphersuites.
9107 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9108 not complete and could lead to a possible single byte overflow
9109 [CVE-2007-5135][] [Ben Laurie]
9111 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007] ###
9113 * Since AES128 and AES256 (and similarly Camellia128 and
9114 Camellia256) share a single mask bit in the logic of
9115 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9116 kludge to work properly if AES128 is available and AES256 isn't
9117 (or if Camellia128 is available and Camellia256 isn't).
9121 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9122 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9123 When a point or a seed is encoded in a BIT STRING, we need to
9124 prevent the removal of trailing zero bits to get the proper DER
9125 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9126 of a NamedBitList, for which trailing 0 bits need to be removed.)
9130 * Have SSL/TLS server implementation tolerate "mismatched" record
9131 protocol version while receiving ClientHello even if the
9132 ClientHello is fragmented. (The server can't insist on the
9133 particular protocol version it has chosen before the ServerHello
9134 message has informed the client about his choice.)
9138 * Add RFC 3779 support.
9140 *Rob Austein for ARIN, Ben Laurie*
9142 * Load error codes if they are not already present instead of using a
9143 static variable. This allows them to be cleanly unloaded and reloaded.
9144 Improve header file function name parsing.
9148 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9149 or CAPABILITY handshake as required by RFCs.
9153 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006] ###
9155 * Introduce limits to prevent malicious keys being able to
9156 cause a denial of service. [CVE-2006-2940][]
9158 *Steve Henson, Bodo Moeller*
9160 * Fix ASN.1 parsing of certain invalid structures that can result
9161 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9163 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9164 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9166 * Fix SSL client code which could crash if connecting to a
9167 malicious SSLv2 server. [CVE-2006-4343][]
9169 *Tavis Ormandy and Will Drewry, Google Security Team*
9171 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9172 match only those. Before that, "AES256-SHA" would be interpreted
9173 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9174 the same strength classification in 0.9.7h) as we currently only
9175 have a single AES bit in the ciphersuite description bitmap.
9176 That change, however, also applied to ciphersuite strings such as
9177 "RC4-MD5" that intentionally matched multiple ciphersuites --
9178 namely, SSL 2.0 ciphersuites in addition to the more common ones
9179 from SSL 3.0/TLS 1.0.
9181 So we change the selection algorithm again: Naming an explicit
9182 ciphersuite selects this one ciphersuite, and any other similar
9183 ciphersuite (same bitmap) from *other* protocol versions.
9184 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9185 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9187 Since SSL 2.0 does not have any ciphersuites for which the
9188 128/256 bit distinction would be relevant, this works for now.
9189 The proper fix will be to use different bits for AES128 and
9190 AES256, which would have avoided the problems from the beginning;
9191 however, bits are scarce, so we can only do this in a new release
9192 (not just a patchlevel) when we can change the SSL_CIPHER
9193 definition to split the single 'unsigned long mask' bitmap into
9194 multiple values to extend the available space.
9199 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006] ###
9201 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9202 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9204 * Add AES IGE and biIGE modes.
9208 * Change the Unix randomness entropy gathering to use poll() when
9209 possible instead of select(), since the latter has some
9210 undesirable limitations.
9212 *Darryl Miles via Richard Levitte and Bodo Moeller*
9214 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9215 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9216 cannot be implicitly activated as part of, e.g., the "AES" alias.
9217 However, please upgrade to OpenSSL 0.9.9[-dev] for
9218 non-experimental use of the ECC ciphersuites to get TLS extension
9219 support, which is required for curve and point format negotiation
9220 to avoid potential handshake problems.
9224 * Disable rogue ciphersuites:
9226 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9227 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9228 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9230 The latter two were purportedly from
9231 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9234 Also deactivate the remaining ciphersuites from
9235 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9236 unofficial, and the ID has long expired.
9240 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9241 dual-core machines) and other potential thread-safety issues.
9245 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9246 versions), which is now available for royalty-free use
9247 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
9248 Also, add Camellia TLS ciphersuites from RFC 4132.
9250 To minimize changes between patchlevels in the OpenSSL 0.9.8
9251 series, Camellia remains excluded from compilation unless OpenSSL
9252 is configured with 'enable-camellia'.
9256 * Disable the padding bug check when compression is in use. The padding
9257 bug check assumes the first packet is of even length, this is not
9258 necessarily true if compression is enabled and can result in false
9259 positives causing handshake failure. The actual bug test is ancient
9260 code so it is hoped that implementations will either have fixed it by
9261 now or any which still have the bug do not support compression.
9265 ### Changes between 0.9.8a and 0.9.8b [04 May 2006] ###
9267 * When applying a cipher rule check to see if string match is an explicit
9268 cipher suite and only match that one cipher suite if it is.
9272 * Link in manifests for VC++ if needed.
9274 *Austin Ziegler <halostatue@gmail.com>*
9276 * Update support for ECC-based TLS ciphersuites according to
9277 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9278 TLS extensions, which are supported starting with the 0.9.9
9279 branch, not in the OpenSSL 0.9.8 branch).
9283 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9284 opaque EVP_CIPHER_CTX handling.
9288 * Fixes and enhancements to zlib compression code. We now only use
9289 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9290 to conform with the standards mentioned here:
9291 http://www.zlib.net/DLL_FAQ.txt
9292 Static zlib linking now works on Windows and the new --with-zlib-include
9293 --with-zlib-lib options to Configure can be used to supply the location
9294 of the headers and library. Gracefully handle case where zlib library
9299 * Several fixes and enhancements to the OID generation code. The old code
9300 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9301 handle numbers larger than ULONG_MAX, truncated printing and had a
9302 non standard OBJ_obj2txt() behaviour.
9306 * Add support for building of engines under engine/ as shared libraries
9307 under VC++ build system.
9311 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9312 Hopefully, we will not see any false combination of paths any more.
9316 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005] ###
9318 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9319 (part of SSL_OP_ALL). This option used to disable the
9320 countermeasure against man-in-the-middle protocol-version
9321 rollback in the SSL 2.0 server implementation, which is a bad
9322 idea. [CVE-2005-2969][]
9324 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9325 for Information Security, National Institute of Advanced Industrial
9326 Science and Technology [AIST*, Japan)]
9328 * Add two function to clear and return the verify parameter flags.
9332 * Keep cipherlists sorted in the source instead of sorting them at
9333 runtime, thus removing the need for a lock.
9337 * Avoid some small subgroup attacks in Diffie-Hellman.
9339 *Nick Mathewson and Ben Laurie*
9341 * Add functions for well-known primes.
9345 * Extended Windows CE support.
9347 *Satoshi Nakamura and Andy Polyakov*
9349 * Initialize SSL_METHOD structures at compile time instead of during
9350 runtime, thus removing the need for a lock.
9354 * Make PKCS7_decrypt() work even if no certificate is supplied by
9355 attempting to decrypt each encrypted key in turn. Add support to
9360 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005] ###
9362 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9365 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9369 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9370 key into the same file any more.
9374 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9378 * Add -utf8 command line and config file option to 'ca'.
9380 *Stefan <stf@udoma.org*
9382 * Removed the macro des_crypt(), as it seems to conflict with some
9383 libraries. Use DES_crypt().
9387 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9388 involves renaming the source and generated shared-libs for
9389 both. The engines will accept the corrected or legacy ids
9390 ('ncipher' and '4758_cca' respectively) when binding. NB,
9391 this only applies when building 'shared'.
9393 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9395 * Add attribute functions to EVP_PKEY structure. Modify
9396 PKCS12_create() to recognize a CSP name attribute and
9397 use it. Make -CSP option work again in pkcs12 utility.
9401 * Add new functionality to the bn blinding code:
9402 - automatic re-creation of the BN_BLINDING parameters after
9403 a fixed number of uses (currently 32)
9404 - add new function for parameter creation
9405 - introduce flags to control the update behaviour of the
9406 BN_BLINDING parameters
9407 - hide BN_BLINDING structure
9408 Add a second BN_BLINDING slot to the RSA structure to improve
9409 performance when a single RSA object is shared among several
9414 * Add support for DTLS.
9416 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9418 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9419 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9423 * Remove buggy and incomplete DH cert support from
9424 ssl/ssl_rsa.c and ssl/s3_both.c
9428 * Use SHA-1 instead of MD5 as the default digest algorithm for
9429 the apps/openssl applications.
9433 * Compile clean with "-Wall -Wmissing-prototypes
9434 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9435 DEBUG_SAFESTACK must also be set.
9439 * Change ./Configure so that certain algorithms can be disabled by default.
9440 The new counterpiece to "no-xxx" is "enable-xxx".
9442 The patented RC5 and MDC2 algorithms will now be disabled unless
9443 "enable-rc5" and "enable-mdc2", respectively, are specified.
9445 (IDEA remains enabled despite being patented. This is because IDEA
9446 is frequently required for interoperability, and there is no license
9447 fee for non-commercial use. As before, "no-idea" can be used to
9448 avoid this algorithm.)
9453 * Add processing of proxy certificates (see RFC 3820). This work was
9454 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9455 EGEE (Enabling Grids for E-science in Europe).
9459 * RC4 performance overhaul on modern architectures/implementations, such
9460 as Intel P4, IA-64 and AMD64.
9464 * New utility extract-section.pl. This can be used specify an alternative
9465 section number in a pod file instead of having to treat each file as
9466 a separate case in Makefile. This can be done by adding two lines to the
9469 =for comment openssl_section:XXX
9471 The blank line is mandatory.
9476 * New arguments -certform, -keyform and -pass for s_client and s_server
9477 to allow alternative format key and certificate files and passphrase
9482 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9483 update associated structures and add various utility functions.
9485 Add new policy related verify parameters, include policy checking in
9486 standard verify code. Enhance 'smime' application with extra parameters
9487 to support policy checking and print out.
9491 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9492 Nehemiah processors. These extensions support AES encryption in hardware
9493 as well as RNG (though RNG support is currently disabled).
9495 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9497 * Deprecate BN_[get|set]_params() functions (they were ignored internally).
9501 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9503 *Andy Polyakov and a number of other people*
9505 * Improved PowerPC platform support. Most notably BIGNUM assembler
9506 implementation contributed by IBM.
9508 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9510 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9511 exponent rather than 'unsigned long'. There is a corresponding change to
9512 the new 'rsa_keygen' element of the RSA_METHOD structure.
9514 *Jelte Jansen, Geoff Thorpe*
9516 * Functionality for creating the initial serial number file is now
9517 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9519 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9520 number file to 1, which is bound to cause problems. To avoid
9521 the problems while respecting compatibility between different 0.9.7
9522 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9523 CA.pl for serial number initialization. With the new release 0.9.8,
9524 we can fix the problem directly in the 'ca' utility.)
9528 * Reduced header interdependencies by declaring more opaque objects in
9529 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9530 give fewer recursive includes, which could break lazy source code - so
9531 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9532 developers should define this symbol when building and using openssl to
9533 ensure they track the recommended behaviour, interfaces, [etc], but
9534 backwards-compatible behaviour prevails when this isn't defined.
9538 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9542 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9543 This will generate a random key of the appropriate length based on the
9544 cipher context. The EVP_CIPHER can provide its own random key generation
9545 routine to support keys of a specific form. This is used in the des and
9546 3des routines to generate a key of the correct parity. Update S/MIME
9547 code to use new functions and hence generate correct parity DES keys.
9548 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9549 valid (weak or incorrect parity).
9553 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9554 as looking them up. This is useful when the verified structure may contain
9555 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9556 present unless the new PKCS7_NO_CRL flag is asserted.
9560 * Extend ASN1 oid configuration module. It now additionally accepts the
9563 shortName = some long name, 1.2.3.4
9567 * Reimplemented the BN_CTX implementation. There is now no more static
9568 limitation on the number of variables it can handle nor the depth of the
9569 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9570 information can now expand as required, and rather than having a single
9571 static array of bignums, BN_CTX now uses a linked-list of such arrays
9572 allowing it to expand on demand whilst maintaining the usefulness of
9573 BN_CTX's "bundling".
9577 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9578 to allow all RSA operations to function using a single BN_CTX.
9582 * Preliminary support for certificate policy evaluation and checking. This
9583 is initially intended to pass the tests outlined in "Conformance Testing
9584 of Relying Party Client Certificate Path Processing Logic" v1.07.
9588 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9589 remained unused and not that useful. A variety of other little bignum
9590 tweaks and fixes have also been made continuing on from the audit (see
9595 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9596 associated ASN1, EVP and SSL functions and old ASN1 macros.
9600 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9601 and this should never fail. So the return value from the use of
9602 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9603 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9607 * BN_CTX_get() should return zero-valued bignums, providing the same
9608 initialised value as BN_new().
9610 *Geoff Thorpe, suggested by Ulf Möller*
9612 * Support for inhibitAnyPolicy certificate extension.
9616 * An audit of the BIGNUM code is underway, for which debugging code is
9617 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9618 is considered valid when processing BIGNUMs, and causes execution to
9619 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9620 further steps are taken to deliberately pollute unused data in BIGNUM
9621 structures to try and expose faulty code further on. For now, openssl will
9622 (in its default mode of operation) continue to tolerate the inconsistent
9623 forms that it has tolerated in the past, but authors and packagers should
9624 consider trying openssl and their own applications when compiled with
9625 these debugging symbols defined. It will help highlight potential bugs in
9626 their own code, and will improve the test coverage for OpenSSL itself. At
9627 some point, these tighter rules will become openssl's default to improve
9628 maintainability, though the assert()s and other overheads will remain only
9629 in debugging configurations. See bn.h for more details.
9631 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9633 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9634 that can only be obtained through BN_CTX_new() (which implicitly
9635 initialises it). The presence of this function only made it possible
9636 to overwrite an existing structure (and cause memory leaks).
9640 * Because of the callback-based approach for implementing LHASH as a
9641 template type, lh_insert() adds opaque objects to hash-tables and
9642 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9643 to clean up those corresponding objects before destroying the hash table
9644 (and losing the object pointers). So some over-zealous constifications in
9645 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9646 objects as "const" and the lh_doall[_arg] callback wrappers are not
9647 prototyped to have "const" restrictions on the object pointers they are
9648 given (and so aren't required to cast them away any more).
9652 * The tmdiff.h API was so ugly and minimal that our own timing utility
9653 (speed) prefers to use its own implementation. The two implementations
9654 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9655 its object type properly exposed (MS_TM) instead of casting to/from "char
9656 *". This may still change yet if someone realises MS_TM and "ms_time_***"
9657 aren't necessarily the greatest nomenclatures - but this is what was used
9658 internally to the implementation so I've used that for now.
9662 * Ensure that deprecated functions do not get compiled when
9663 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9664 the self-tests were still using deprecated key-generation functions so
9665 these have been updated also.
9669 * Reorganise PKCS#7 code to separate the digest location functionality
9670 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9671 New function PKCS7_set_digest() to set the digest type for PKCS#7
9672 digestedData type. Add additional code to correctly generate the
9673 digestedData type and add support for this type in PKCS7 initialization
9678 * New function PKCS7_set0_type_other() this initializes a PKCS7
9679 structure of type "other".
9683 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9684 sure the loop does correctly stop and breaking ("division by zero")
9685 modulus operations are not performed. The (pre-generated) prime
9686 table crypto/bn/bn_prime.h was already correct, but it could not be
9687 re-generated on some platforms because of the "division by zero"
9688 situation in the script.
9690 *Ralf S. Engelschall*
9692 * Update support for ECC-based TLS ciphersuites according to
9693 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9694 SHA-1 now is only used for "small" curves (where the
9695 representation of a field element takes up to 24 bytes); for
9696 larger curves, the field element resulting from ECDH is directly
9697 used as premaster secret.
9699 *Douglas Stebila (Sun Microsystems Laboratories)*
9701 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9702 curve secp160r1 to the tests.
9704 *Douglas Stebila (Sun Microsystems Laboratories)*
9706 * Add the possibility to load symbols globally with DSO.
9708 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9710 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9711 control of the error stack.
9715 * Add support for STORE in ENGINE.
9719 * Add the STORE type. The intention is to provide a common interface
9720 to certificate and key stores, be they simple file-based stores, or
9721 HSM-type store, or LDAP stores, or...
9722 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9726 * Add a generic structure called OPENSSL_ITEM. This can be used to
9727 pass a list of arguments to any function as well as provide a way
9728 for a function to pass data back to the caller.
9732 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9733 works like BUF_strdup() but can be used to duplicate a portion of
9734 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9739 * Add the function sk_find_ex() which works like sk_find(), but will
9740 return an index to an element even if an exact match couldn't be
9741 found. The index is guaranteed to point at the element where the
9742 searched-for key would be inserted to preserve sorting order.
9746 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9747 takes an extra flags argument for optional functionality. Currently,
9748 the following flags are defined:
9750 OBJ_BSEARCH_VALUE_ON_NOMATCH
9751 This one gets OBJ_bsearch_ex() to return a pointer to the first
9752 element where the comparing function returns a negative or zero
9755 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9756 This one gets OBJ_bsearch_ex() to return a pointer to the first
9757 element where the comparing function returns zero. This is useful
9758 if there are more than one element where the comparing function
9763 * Make it possible to create self-signed certificates with 'openssl ca'
9764 in such a way that the self-signed certificate becomes part of the
9765 CA database and uses the same mechanisms for serial number generation
9766 as all other certificate signing. The new flag '-selfsign' enables
9767 this functionality. Adapt CA.sh and CA.pl.in.
9771 * Add functionality to check the public key of a certificate request
9772 against a given private. This is useful to check that a certificate
9773 request can be signed by that key (self-signing).
9777 * Make it possible to have multiple active certificates with the same
9778 subject in the CA index file. This is done only if the keyword
9779 'unique_subject' is set to 'no' in the main CA section (default
9780 if 'CA_default') of the configuration file. The value is saved
9781 with the database itself in a separate index attribute file,
9782 named like the index file with '.attr' appended to the name.
9786 * Generate multi-valued AVAs using '+' notation in config files for
9791 * Support for nameConstraints certificate extension.
9795 * Support for policyConstraints certificate extension.
9799 * Support for policyMappings certificate extension.
9803 * Make sure the default DSA_METHOD implementation only uses its
9804 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9805 and change its own handlers to be NULL so as to remove unnecessary
9806 indirection. This lets alternative implementations fallback to the
9807 default implementation more easily.
9811 * Support for directoryName in GeneralName related extensions
9816 * Make it possible to link applications using Makefile.shared.
9817 Make that possible even when linking against static libraries!
9821 * Support for single pass processing for S/MIME signing. This now
9822 means that S/MIME signing can be done from a pipe, in addition
9823 cleartext signing (multipart/signed type) is effectively streaming
9824 and the signed data does not need to be all held in memory.
9826 This is done with a new flag PKCS7_STREAM. When this flag is set
9827 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
9828 is done after the data is output (and digests calculated) in
9829 SMIME_write_PKCS7().
9833 * Add full support for -rpath/-R, both in shared libraries and
9834 applications, at least on the platforms where it's known how
9839 * In crypto/ec/ec_mult.c, implement fast point multiplication with
9840 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
9841 will now compute a table of multiples of the generator that
9842 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
9843 faster (notably in the case of a single point multiplication,
9844 scalar * generator).
9846 *Nils Larsch, Bodo Moeller*
9848 * IPv6 support for certificate extensions. The various extensions
9849 which use the IP:a.b.c.d can now take IPv6 addresses using the
9850 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
9855 * Added an ENGINE that implements RSA by performing private key
9856 exponentiations with the GMP library. The conversions to and from
9857 GMP's mpz_t format aren't optimised nor are any montgomery forms
9858 cached, and on x86 it appears OpenSSL's own performance has caught up.
9859 However there are likely to be other architectures where GMP could
9860 provide a boost. This ENGINE is not built in by default, but it can be
9861 specified at Configure time and should be accompanied by the necessary
9862 linker additions, eg;
9863 ./config -DOPENSSL_USE_GMP -lgmp
9867 * "openssl engine" will not display ENGINE/DSO load failure errors when
9868 testing availability of engines with "-t" - the old behaviour is
9869 produced by increasing the feature's verbosity with "-tt".
9873 * ECDSA routines: under certain error conditions uninitialized BN objects
9874 could be freed. Solution: make sure initialization is performed early
9875 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
9880 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
9881 and DH_METHOD (eg. by ENGINE implementations) to override the normal
9882 software implementations. For DSA and DH, parameter generation can
9883 also be overridden by providing the appropriate method callbacks.
9887 * Change the "progress" mechanism used in key-generation and
9888 primality testing to functions that take a new BN_GENCB pointer in
9889 place of callback/argument pairs. The new API functions have "_ex"
9890 postfixes and the older functions are reimplemented as wrappers for
9891 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
9892 declarations of the old functions to help (graceful) attempts to
9893 migrate to the new functions. Also, the new key-generation API
9894 functions operate on a caller-supplied key-structure and return
9895 success/failure rather than returning a key or NULL - this is to
9896 help make "keygen" another member function of RSA_METHOD etc.
9898 Example for using the new callback interface:
9900 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
9904 BN_GENCB_set(&my_cb, my_callback, my_arg);
9906 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
9907 /* For the meaning of a, b in calls to my_callback(), see the
9908 * documentation of the function that calls the callback.
9909 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
9910 * my_callback should return 1 if it wants BN_is_prime_ex()
9911 * to continue, or 0 to stop.
9916 * Change the ZLIB compression method to be stateful, and make it
9917 available to TLS with the number defined in
9918 draft-ietf-tls-compression-04.txt.
9922 * Add the ASN.1 structures and functions for CertificatePair, which
9923 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
9925 CertificatePair ::= SEQUENCE {
9926 forward [0] Certificate OPTIONAL,
9927 reverse [1] Certificate OPTIONAL,
9928 -- at least one of the pair shall be present -- }
9930 Also implement the PEM functions to read and write certificate
9931 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
9933 This needed to be defined, mostly for the sake of the LDAP
9934 attribute crossCertificatePair, but may prove useful elsewhere as
9939 * Make it possible to inhibit symlinking of shared libraries in
9940 Makefile.shared, for Cygwin's sake.
9944 * Extend the BIGNUM API by creating a function
9945 void BN_set_negative(BIGNUM *a, int neg);
9946 and a macro that behave like
9947 int BN_is_negative(const BIGNUM *a);
9949 to avoid the need to access 'a->neg' directly in applications.
9953 * Implement fast modular reduction for pseudo-Mersenne primes
9954 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
9955 EC_GROUP_new_curve_GFp() will now automatically use this
9958 *Nils Larsch <nla@trustcenter.de>*
9960 * Add new lock type (CRYPTO_LOCK_BN).
9964 * Change the ENGINE framework to automatically load engines
9965 dynamically from specific directories unless they could be
9966 found to already be built in or loaded. Move all the
9967 current engines except for the cryptodev one to a new
9969 The engines in engines/ are built as shared libraries if
9970 the "shared" options was given to ./Configure or ./config.
9971 Otherwise, they are inserted in libcrypto.a.
9972 /usr/local/ssl/engines is the default directory for dynamic
9973 engines, but that can be overridden at configure time through
9974 the usual use of --prefix and/or --openssldir, and at run
9975 time with the environment variable OPENSSL_ENGINES.
9977 *Geoff Thorpe and Richard Levitte*
9979 * Add Makefile.shared, a helper makefile to build shared
9980 libraries. Adapt Makefile.org.
9984 * Add version info to Win32 DLLs.
9986 *Peter 'Luna' Runestig" <peter@runestig.com>*
9988 * Add new 'medium level' PKCS#12 API. Certificates and keys
9989 can be added using this API to created arbitrary PKCS#12
9990 files while avoiding the low level API.
9992 New options to PKCS12_create(), key or cert can be NULL and
9993 will then be omitted from the output file. The encryption
9994 algorithm NIDs can be set to -1 for no encryption, the mac
9995 iteration count can be set to 0 to omit the mac.
9997 Enhance pkcs12 utility by making the -nokeys and -nocerts
9998 options work when creating a PKCS#12 file. New option -nomac
9999 to omit the mac, NONE can be set for an encryption algorithm.
10000 New code is modified to use the enhanced PKCS12_create()
10001 instead of the low level API.
10005 * Extend ASN1 encoder to support indefinite length constructed
10006 encoding. This can output sequences tags and octet strings in
10007 this form. Modify pk7_asn1.c to support indefinite length
10008 encoding. This is experimental and needs additional code to
10009 be useful, such as an ASN1 bio and some enhanced streaming
10012 Extend template encode functionality so that tagging is passed
10013 down to the template encoder.
10017 * Let 'openssl req' fail if an argument to '-newkey' is not
10018 recognized instead of using RSA as a default.
10022 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10023 As these are not official, they are not included in "ALL";
10024 the "ECCdraft" ciphersuite group alias can be used to select them.
10026 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10028 * Add ECDH engine support.
10030 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10032 * Add ECDH in new directory crypto/ecdh/.
10034 *Douglas Stebila (Sun Microsystems Laboratories)*
10036 * Let BN_rand_range() abort with an error after 100 iterations
10037 without success (which indicates a broken PRNG).
10041 * Change BN_mod_sqrt() so that it verifies that the input value
10042 is really the square of the return value. (Previously,
10043 BN_mod_sqrt would show GIGO behaviour.)
10047 * Add named elliptic curves over binary fields from X9.62, SECG,
10048 and WAP/WTLS; add OIDs that were still missing.
10050 *Sheueling Chang Shantz and Douglas Stebila
10051 (Sun Microsystems Laboratories)*
10053 * Extend the EC library for elliptic curves over binary fields
10054 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10057 EC_GF2m_simple_method
10061 EC_GROUP_new_curve_GF2m
10062 EC_GROUP_set_curve_GF2m
10063 EC_GROUP_get_curve_GF2m
10064 EC_POINT_set_affine_coordinates_GF2m
10065 EC_POINT_get_affine_coordinates_GF2m
10066 EC_POINT_set_compressed_coordinates_GF2m
10068 Point compression for binary fields is disabled by default for
10069 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10072 As binary polynomials are represented as BIGNUMs, various members
10073 of the EC_GROUP and EC_POINT data structures can be shared
10074 between the implementations for prime fields and binary fields;
10075 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
10076 are essentially identical to their ..._GFp counterparts.
10077 (For simplicity, the '..._GFp' prefix has been dropped from
10078 various internal method names.)
10080 An internal 'field_div' method (similar to 'field_mul' and
10081 'field_sqr') has been added; this is used only for binary fields.
10083 *Sheueling Chang Shantz and Douglas Stebila
10084 (Sun Microsystems Laboratories)*
10086 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10087 through methods ('mul', 'precompute_mult').
10089 The generic implementations (now internally called 'ec_wNAF_mul'
10090 and 'ec_wNAF_precomputed_mult') remain the default if these
10091 methods are undefined.
10093 *Sheueling Chang Shantz and Douglas Stebila
10094 (Sun Microsystems Laboratories)*
10096 * New function EC_GROUP_get_degree, which is defined through
10097 EC_METHOD. For curves over prime fields, this returns the bit
10098 length of the modulus.
10100 *Sheueling Chang Shantz and Douglas Stebila
10101 (Sun Microsystems Laboratories)*
10103 * New functions EC_GROUP_dup, EC_POINT_dup.
10104 (These simply call ..._new and ..._copy).
10106 *Sheueling Chang Shantz and Douglas Stebila
10107 (Sun Microsystems Laboratories)*
10109 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10110 Polynomials are represented as BIGNUMs (where the sign bit is not
10111 used) in the following functions [macros]:
10114 BN_GF2m_sub [= BN_GF2m_add]
10115 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10116 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10117 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10119 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10120 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10121 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10122 BN_GF2m_cmp [= BN_ucmp]
10124 (Note that only the 'mod' functions are actually for fields GF(2^m).
10125 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10127 For some functions, an the irreducible polynomial defining a
10128 field can be given as an 'unsigned int[]' with strictly
10129 decreasing elements giving the indices of those bits that are set;
10130 i.e., p[] represents the polynomial
10131 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10133 p[0] > p[1] > ... > p[k] = 0.
10134 This applies to the following functions:
10137 BN_GF2m_mod_mul_arr
10138 BN_GF2m_mod_sqr_arr
10139 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10140 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10141 BN_GF2m_mod_exp_arr
10142 BN_GF2m_mod_sqrt_arr
10143 BN_GF2m_mod_solve_quad_arr
10147 Conversion can be performed by the following functions:
10152 bntest.c has additional tests for binary polynomial arithmetic.
10154 Two implementations for BN_GF2m_mod_div() are available.
10155 The default algorithm simply uses BN_GF2m_mod_inv() and
10156 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10157 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10158 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10160 *Sheueling Chang Shantz and Douglas Stebila
10161 (Sun Microsystems Laboratories)*
10163 * Add new error code 'ERR_R_DISABLED' that can be used when some
10164 functionality is disabled at compile-time.
10166 *Douglas Stebila <douglas.stebila@sun.com>*
10168 * Change default behaviour of 'openssl asn1parse' so that more
10169 information is visible when viewing, e.g., a certificate:
10171 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10172 mode the content of non-printable OCTET STRINGs is output in a
10173 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10174 avoid the appearance of a printable string.
10176 *Nils Larsch <nla@trustcenter.de>*
10178 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10180 EC_GROUP_set_asn1_flag()
10181 EC_GROUP_get_asn1_flag()
10182 EC_GROUP_set_point_conversion_form()
10183 EC_GROUP_get_point_conversion_form()
10184 These control ASN1 encoding details:
10185 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10186 has been set to OPENSSL_EC_NAMED_CURVE.
10187 - Points are encoded in uncompressed form by default; options for
10188 asn1_for are as for point2oct, namely
10189 POINT_CONVERSION_COMPRESSED
10190 POINT_CONVERSION_UNCOMPRESSED
10191 POINT_CONVERSION_HYBRID
10193 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10195 EC_GROUP_set_seed()
10196 EC_GROUP_get0_seed()
10197 EC_GROUP_get_seed_len()
10198 This is used only for ASN1 purposes (so far).
10200 *Nils Larsch <nla@trustcenter.de>*
10202 * Add 'field_type' member to EC_METHOD, which holds the NID
10203 of the appropriate field type OID. The new function
10204 EC_METHOD_get_field_type() returns this value.
10206 *Nils Larsch <nla@trustcenter.de>*
10209 EC_POINT_point2bn()
10210 EC_POINT_bn2point()
10211 EC_POINT_point2hex()
10212 EC_POINT_hex2point()
10213 providing useful interfaces to EC_POINT_point2oct() and
10214 EC_POINT_oct2point().
10216 *Nils Larsch <nla@trustcenter.de>*
10218 * Change internals of the EC library so that the functions
10219 EC_GROUP_set_generator()
10220 EC_GROUP_get_generator()
10221 EC_GROUP_get_order()
10222 EC_GROUP_get_cofactor()
10223 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10224 to methods, which would lead to unnecessary code duplication when
10225 adding different types of curves.
10227 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10229 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10230 arithmetic, and such that modified wNAFs are generated
10231 (which avoid length expansion in many cases).
10235 * Add a function EC_GROUP_check_discriminant() (defined via
10236 EC_METHOD) that verifies that the curve discriminant is non-zero.
10238 Add a function EC_GROUP_check() that makes some sanity tests
10239 on a EC_GROUP, its generator and order. This includes
10240 EC_GROUP_check_discriminant().
10242 *Nils Larsch <nla@trustcenter.de>*
10244 * Add ECDSA in new directory crypto/ecdsa/.
10246 Add applications 'openssl ecparam' and 'openssl ecdsa'
10247 (these are based on 'openssl dsaparam' and 'openssl dsa').
10249 ECDSA support is also included in various other files across the
10250 library. Most notably,
10251 - 'openssl req' now has a '-newkey ecdsa:file' option;
10252 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10253 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10254 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10255 them suitable for ECDSA where domain parameters must be
10256 extracted before the specific public key;
10257 - ECDSA engine support has been added.
10259 *Nils Larsch <nla@trustcenter.de>*
10261 * Include some named elliptic curves, and add OIDs from X9.62,
10262 SECG, and WAP/WTLS. Each curve can be obtained from the new
10264 EC_GROUP_new_by_curve_name(),
10265 and the list of available named curves can be obtained with
10266 EC_get_builtin_curves().
10267 Also add a 'curve_name' member to EC_GROUP objects, which can be
10269 EC_GROUP_set_curve_name()
10270 EC_GROUP_get_curve_name()
10272 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10274 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10275 was actually never needed) and in BN_mul(). The removal in BN_mul()
10276 required a small change in bn_mul_part_recursive() and the addition
10277 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10278 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10279 bn_sub_words() and bn_add_words() except they take arrays with
10284 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007] ###
10286 * Cleanse PEM buffers before freeing them since they may contain
10289 *Benjamin Bennett <ben@psc.edu>*
10291 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10292 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10293 authentication-only ciphersuites.
10297 * Since AES128 and AES256 share a single mask bit in the logic of
10298 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10299 kludge to work properly if AES128 is available and AES256 isn't.
10303 * Expand security boundary to match 1.1.1 module.
10307 * Remove redundant features: hash file source, editing of test vectors
10308 modify fipsld to use external fips_premain.c signature.
10312 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10313 run algorithm test programs.
10317 * Make algorithm test programs more tolerant of whitespace.
10321 * Have SSL/TLS server implementation tolerate "mismatched" record
10322 protocol version while receiving ClientHello even if the
10323 ClientHello is fragmented. (The server can't insist on the
10324 particular protocol version it has chosen before the ServerHello
10325 message has informed the client about his choice.)
10329 * Load error codes if they are not already present instead of using a
10330 static variable. This allows them to be cleanly unloaded and reloaded.
10334 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006] ###
10336 * Introduce limits to prevent malicious keys being able to
10337 cause a denial of service. [CVE-2006-2940][]
10339 *Steve Henson, Bodo Moeller*
10341 * Fix ASN.1 parsing of certain invalid structures that can result
10342 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10344 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10345 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10347 * Fix SSL client code which could crash if connecting to a
10348 malicious SSLv2 server. [CVE-2006-4343][]
10350 *Tavis Ormandy and Will Drewry, Google Security Team*
10352 * Change ciphersuite string processing so that an explicit
10353 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10354 will no longer include "AES128-SHA"), and any other similar
10355 ciphersuite (same bitmap) from *other* protocol versions (so that
10356 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10357 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10358 changes from 0.9.8b and 0.9.8d.
10362 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006] ###
10364 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10365 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10367 * Change the Unix randomness entropy gathering to use poll() when
10368 possible instead of select(), since the latter has some
10369 undesirable limitations.
10371 *Darryl Miles via Richard Levitte and Bodo Moeller*
10373 * Disable rogue ciphersuites:
10375 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10376 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10377 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10379 The latter two were purportedly from
10380 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10383 Also deactivate the remaining ciphersuites from
10384 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10385 unofficial, and the ID has long expired.
10389 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10390 dual-core machines) and other potential thread-safety issues.
10394 ### Changes between 0.9.7i and 0.9.7j [04 May 2006] ###
10396 * Adapt fipsld and the build system to link against the validated FIPS
10397 module in FIPS mode.
10401 * Fixes for VC++ 2005 build under Windows.
10405 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10406 from a Windows bash shell such as MSYS. It is autodetected from the
10407 "config" script when run from a VC++ environment. Modify standard VC++
10408 build to use fipscanister.o from the GNU make build.
10412 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005] ###
10414 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10415 The value now differs depending on if you build for FIPS or not.
10416 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10417 safely run with a non-FIPSed libcrypto, as it may crash because of
10418 the difference induced by this change.
10422 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005] ###
10424 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10425 (part of SSL_OP_ALL). This option used to disable the
10426 countermeasure against man-in-the-middle protocol-version
10427 rollback in the SSL 2.0 server implementation, which is a bad
10428 idea. [CVE-2005-2969][]
10430 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10431 for Information Security, National Institute of Advanced Industrial
10432 Science and Technology [AIST*, Japan)]
10434 * Minimal support for X9.31 signatures and PSS padding modes. This is
10435 mainly for FIPS compliance and not fully integrated at this stage.
10439 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10440 the exponentiation using a fixed-length exponent. (Otherwise,
10441 the information leaked through timing could expose the secret key
10442 after many signatures; cf. Bleichenbacher's attack on DSA with
10447 * Make a new fixed-window mod_exp implementation the default for
10448 RSA, DSA, and DH private-key operations so that the sequence of
10449 squares and multiplies and the memory access pattern are
10450 independent of the particular secret key. This will mitigate
10451 cache-timing and potential related attacks.
10453 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10454 and this is automatically used by BN_mod_exp_mont() if the new flag
10455 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10456 will use this BN flag for private exponents unless the flag
10457 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10458 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10461 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10463 * Change the client implementation for SSLv23_method() and
10464 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10465 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10466 (Previously, the SSL 2.0 backwards compatible Client Hello
10467 message format would be used even with SSL_OP_NO_SSLv2.)
10471 * Add support for smime-type MIME parameter in S/MIME messages which some
10476 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10477 a threadsafe manner. Modify rsa code to use new function and add calls
10478 to dsa and dh code (which had race conditions before).
10482 * Include the fixed error library code in the C error file definitions
10483 instead of fixing them up at runtime. This keeps the error code
10484 structures constant.
10488 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005] ###
10490 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10493 * Fixes for newer kerberos headers. NB: the casts are needed because
10494 the 'length' field is signed on one version and unsigned on another
10495 with no (?) obvious way to tell the difference, without these VC++
10496 complains. Also the "definition" of FAR (blank) is no longer included
10497 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10498 some needed definitions.
10502 * Undo Cygwin change.
10506 * Added support for proxy certificates according to RFC 3820.
10507 Because they may be a security thread to unaware applications,
10508 they must be explicitly allowed in run-time. See
10509 docs/HOWTO/proxy_certificates.txt for further information.
10513 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005] ###
10515 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10516 server and client random values. Previously
10517 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10518 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10520 This change has negligible security impact because:
10522 1. Server and client random values still have 24 bytes of pseudo random
10525 2. Server and client random values are sent in the clear in the initial
10528 3. The master secret is derived using the premaster secret (48 bytes in
10529 size for static RSA ciphersuites) as well as client server and random
10532 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10535 *Stephen Henson, reported by UK NISCC*
10537 * Use Windows randomness collection on Cygwin.
10541 * Fix hang in EGD/PRNGD query when communication socket is closed
10542 prematurely by EGD/PRNGD.
10544 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10546 * Prompt for pass phrases when appropriate for PKCS12 input format.
10550 * Back-port of selected performance improvements from development
10551 branch, as well as improved support for PowerPC platforms.
10555 * Add lots of checks for memory allocation failure, error codes to indicate
10556 failure and freeing up memory if a failure occurs.
10558 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10560 * Add new -passin argument to dgst.
10564 * Perform some character comparisons of different types in X509_NAME_cmp:
10565 this is needed for some certificates that re-encode DNs into UTF8Strings
10566 (in violation of RFC3280) and can't or won't issue name rollover
10571 * Make an explicit check during certificate validation to see that
10572 the CA setting in each certificate on the chain is correct. As a
10573 side effect always do the following basic checks on extensions,
10574 not just when there's an associated purpose to the check:
10576 - if there is an unhandled critical extension (unless the user
10577 has chosen to ignore this fault)
10578 - if the path length has been exceeded (if one is set at all)
10579 - that certain extensions fit the associated purpose (if one has
10584 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004] ###
10586 * Avoid a race condition when CRLs are checked in a multi threaded
10587 environment. This would happen due to the reordering of the revoked
10588 entries during signature checking and serial number lookup. Now the
10589 encoding is cached and the serial number sort performed under a lock.
10590 Add new STACK function sk_is_sorted().
10594 * Add Delta CRL to the extension code.
10598 * Various fixes to s3_pkt.c so alerts are sent properly.
10600 *David Holmes <d.holmes@f5.com>*
10602 * Reduce the chances of duplicate issuer name and serial numbers (in
10603 violation of RFC3280) using the OpenSSL certificate creation utilities.
10604 This is done by creating a random 64 bit value for the initial serial
10605 number when a serial number file is created or when a self signed
10606 certificate is created using 'openssl req -x509'. The initial serial
10607 number file is created using 'openssl x509 -next_serial' in CA.pl
10608 rather than being initialized to 1.
10612 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004] ###
10614 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10615 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10617 *Joe Orton, Steve Henson*
10619 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10622 *Joe Orton, Steve Henson*
10624 * Make it possible to have multiple active certificates with the same
10625 subject in the CA index file. This is done only if the keyword
10626 'unique_subject' is set to 'no' in the main CA section (default
10627 if 'CA_default') of the configuration file. The value is saved
10628 with the database itself in a separate index attribute file,
10629 named like the index file with '.attr' appended to the name.
10633 * X509 verify fixes. Disable broken certificate workarounds when
10634 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10635 keyUsage extension present. Don't accept CRLs with unhandled critical
10636 extensions: since verify currently doesn't process CRL extensions this
10637 rejects a CRL with *any* critical extensions. Add new verify error codes
10642 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10643 A clarification of RFC2560 will require the use of OCTET STRINGs and
10644 some implementations cannot handle the current raw format. Since OpenSSL
10645 copies and compares OCSP nonces as opaque blobs without any attempt at
10646 parsing them this should not create any compatibility issues.
10650 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10651 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10652 this HMAC (and other) operations are several times slower than OpenSSL
10657 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10659 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10661 * Use the correct content when signing type "other".
10665 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003] ###
10667 * Fix various bugs revealed by running the NISCC test suite:
10669 Stop out of bounds reads in the ASN1 code when presented with
10670 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10672 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10674 If verify callback ignores invalid public key errors don't try to check
10675 certificate signature with the NULL public key.
10680 * New -ignore_err option in ocsp application to stop the server
10681 exiting on the first error in a request.
10685 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10686 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10691 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10692 extra data after the compression methods not only for TLS 1.0
10693 but also for SSL 3.0 (as required by the specification).
10695 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10697 * Change X509_certificate_type() to mark the key as exported/exportable
10698 when it's 512 *bits* long, not 512 bytes.
10702 * Change AES_cbc_encrypt() so it outputs exact multiple of
10703 blocks during encryption.
10707 * Various fixes to base64 BIO and non blocking I/O. On write
10708 flushes were not handled properly if the BIO retried. On read
10709 data was not being buffered properly and had various logic bugs.
10710 This also affects blocking I/O when the data being decoded is a
10715 * Various S/MIME bugfixes and compatibility changes:
10716 output correct application/pkcs7 MIME type if
10717 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10718 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10719 of files as .eml work). Correctly handle very long lines in MIME
10724 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003] ###
10726 * Countermeasure against the Klima-Pokorny-Rosa extension of
10727 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10728 a protocol version number mismatch like a decryption error
10729 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10733 * Turn on RSA blinding by default in the default implementation
10734 to avoid a timing attack. Applications that don't want it can call
10735 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10736 They would be ill-advised to do so in most cases.
10738 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10740 * Change RSA blinding code so that it works when the PRNG is not
10741 seeded (in this case, the secret RSA exponent is abused as
10742 an unpredictable seed -- if it is not unpredictable, there
10743 is no point in blinding anyway). Make RSA blinding thread-safe
10744 by remembering the creator's thread ID in rsa->blinding and
10745 having all other threads use local one-time blinding factors
10746 (this requires more computation than sharing rsa->blinding, but
10747 avoids excessive locking; and if an RSA object is not shared
10748 between threads, blinding will still be very fast).
10752 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10753 ENGINE as defaults for all supported algorithms irrespective of
10754 the 'flags' parameter. 'flags' is now honoured, so applications
10755 should make sure they are passing it correctly.
10759 * Target "mingw" now allows native Windows code to be generated in
10760 the Cygwin environment as well as with the MinGW compiler.
10764 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003] ###
10766 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10767 via timing by performing a MAC computation even if incorrect
10768 block cipher padding has been found. This is a countermeasure
10769 against active attacks where the attacker has to distinguish
10770 between bad padding and a MAC verification error. [CVE-2003-0078][]
10772 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10773 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10774 Martin Vuagnoux (EPFL, Ilion)*
10776 * Make the no-err option work as intended. The intention with no-err
10777 is not to have the whole error stack handling routines removed from
10778 libcrypto, it's only intended to remove all the function name and
10779 reason texts, thereby removing some of the footprint that may not
10780 be interesting if those errors aren't displayed anyway.
10782 NOTE: it's still possible for any application or module to have its
10783 own set of error texts inserted. The routines are there, just not
10784 used by default when no-err is given.
10788 * Add support for FreeBSD on IA64.
10790 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10792 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10793 Kerberos function mit_des_cbc_cksum(). Before this change,
10794 the value returned by DES_cbc_cksum() was like the one from
10795 mit_des_cbc_cksum(), except the bytes were swapped.
10797 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10799 * Allow an application to disable the automatic SSL chain building.
10800 Before this a rather primitive chain build was always performed in
10801 ssl3_output_cert_chain(): an application had no way to send the
10802 correct chain if the automatic operation produced an incorrect result.
10804 Now the chain builder is disabled if either:
10806 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10808 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10810 The reasoning behind this is that an application would not want the
10811 auto chain building to take place if extra chain certificates are
10812 present and it might also want a means of sending no additional
10813 certificates (for example the chain has two certificates and the
10818 * Add the possibility to build without the ENGINE framework.
10820 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10822 * Under Win32 gmtime() can return NULL: check return value in
10823 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
10827 * DSA routines: under certain error conditions uninitialized BN objects
10828 could be freed. Solution: make sure initialization is performed early
10829 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
10830 Nils Larsch <nla@trustcenter.de> via PR#459)
10834 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
10835 checked on reconnect on the client side, therefore session resumption
10836 could still fail with a "ssl session id is different" error. This
10837 behaviour is masked when SSL_OP_ALL is used due to
10838 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
10839 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10840 followup to PR #377.
10844 * IA-32 assembler support enhancements: unified ELF targets, support
10845 for SCO/Caldera platforms, fix for Cygwin shared build.
10849 * Add support for FreeBSD on sparc64. As a consequence, support for
10850 FreeBSD on non-x86 processors is separate from x86 processors on
10851 the config script, much like the NetBSD support.
10853 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
10855 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002] ###
10857 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
10860 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
10861 code (06) was taken as the first octet of the session ID and the last
10862 octet was ignored consequently. As a result SSLv2 client side session
10863 caching could not have worked due to the session ID mismatch between
10865 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
10870 * Change the declaration of needed Kerberos libraries to use EX_LIBS
10871 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
10876 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
10877 seems that in spite of existing for more than a year, many application
10878 author have done nothing to provide the necessary callbacks, which
10879 means that this particular engine will not work properly anywhere.
10880 This is a very unfortunate situation which forces us, in the name
10881 of usability, to give the hw_ncipher.c a static lock, which is part
10883 NOTE: This is for the 0.9.7 series ONLY. This hack will never
10884 appear in 0.9.8 or later. We EXPECT application authors to have
10885 dealt properly with this when 0.9.8 is released (unless we actually
10886 make such changes in the libcrypto locking code that changes will
10887 have to be made anyway).
10891 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
10892 octets have been read, EOF or an error occurs. Without this change
10893 some truncated ASN1 structures will not produce an error.
10897 * Disable Heimdal support, since it hasn't been fully implemented.
10898 Still give the possibility to force the use of Heimdal, but with
10899 warnings and a request that patches get sent to openssl-dev.
10903 * Add the VC-CE target, introduce the WINCE sysname, and add
10904 INSTALL.WCE and appropriate conditionals to make it build.
10906 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10908 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
10909 cygssl-x.y.z.dll, where x, y and z are the major, minor and
10910 edit numbers of the version.
10912 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
10914 * Introduce safe string copy and catenation functions
10915 (BUF_strlcpy() and BUF_strlcat()).
10917 *Ben Laurie (CHATS) and Richard Levitte*
10919 * Avoid using fixed-size buffers for one-line DNs.
10921 *Ben Laurie (CHATS)*
10923 * Add BUF_MEM_grow_clean() to avoid information leakage when
10924 resizing buffers containing secrets, and use where appropriate.
10926 *Ben Laurie (CHATS)*
10928 * Avoid using fixed size buffers for configuration file location.
10930 *Ben Laurie (CHATS)*
10932 * Avoid filename truncation for various CA files.
10934 *Ben Laurie (CHATS)*
10936 * Use sizeof in preference to magic numbers.
10938 *Ben Laurie (CHATS)*
10940 * Avoid filename truncation in cert requests.
10942 *Ben Laurie (CHATS)*
10944 * Add assertions to check for (supposedly impossible) buffer
10947 *Ben Laurie (CHATS)*
10949 * Don't cache truncated DNS entries in the local cache (this could
10950 potentially lead to a spoofing attack).
10952 *Ben Laurie (CHATS)*
10954 * Fix various buffers to be large enough for hex/decimal
10955 representations in a platform independent manner.
10957 *Ben Laurie (CHATS)*
10959 * Add CRYPTO_realloc_clean() to avoid information leakage when
10960 resizing buffers containing secrets, and use where appropriate.
10962 *Ben Laurie (CHATS)*
10964 * Add BIO_indent() to avoid much slightly worrying code to do
10967 *Ben Laurie (CHATS)*
10969 * Convert sprintf()/BIO_puts() to BIO_printf().
10971 *Ben Laurie (CHATS)*
10973 * buffer_gets() could terminate with the buffer only half
10976 *Ben Laurie (CHATS)*
10978 * Add assertions to prevent user-supplied crypto functions from
10979 overflowing internal buffers by having large block sizes, etc.
10981 *Ben Laurie (CHATS)*
10983 * New OPENSSL_assert() macro (similar to assert(), but enabled
10986 *Ben Laurie (CHATS)*
10988 * Eliminate unused copy of key in RC4.
10990 *Ben Laurie (CHATS)*
10992 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
10994 *Ben Laurie (CHATS)*
10996 * Fix off-by-one error in EGD path.
10998 *Ben Laurie (CHATS)*
11000 * If RANDFILE path is too long, ignore instead of truncating.
11002 *Ben Laurie (CHATS)*
11004 * Eliminate unused and incorrectly sized X.509 structure
11007 *Ben Laurie (CHATS)*
11009 * Eliminate unused and dangerous function knumber().
11011 *Ben Laurie (CHATS)*
11013 * Eliminate unused and dangerous structure, KSSL_ERR.
11015 *Ben Laurie (CHATS)*
11017 * Protect against overlong session ID context length in an encoded
11018 session object. Since these are local, this does not appear to be
11021 *Ben Laurie (CHATS)*
11023 * Change from security patch (see 0.9.6e below) that did not affect
11024 the 0.9.6 release series:
11026 Remote buffer overflow in SSL3 protocol - an attacker could
11027 supply an oversized master key in Kerberos-enabled versions.
11030 *Ben Laurie (CHATS)*
11032 * Change the SSL kerb5 codes to match RFC 2712.
11036 * Make -nameopt work fully for req and add -reqopt switch.
11038 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11040 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11042 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11044 * Make sure tests can be performed even if the corresponding algorithms
11045 have been removed entirely. This was also the last step to make
11046 OpenSSL compilable with DJGPP under all reasonable conditions.
11048 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11050 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11051 to allow version independent disabling of normally unselected ciphers,
11052 which may be activated as a side-effect of selecting a single cipher.
11054 (E.g., cipher list string "RSA" enables ciphersuites that are left
11055 out of "ALL" because they do not provide symmetric encryption.
11056 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11058 *Lutz Jaenicke, Bodo Moeller*
11060 * Add appropriate support for separate platform-dependent build
11061 directories. The recommended way to make a platform-dependent
11062 build directory is the following (tested on Linux), maybe with
11065 # Place yourself outside of the OpenSSL source tree. In
11066 # this example, the environment variable OPENSSL_SOURCE
11067 # is assumed to contain the absolute OpenSSL source directory.
11068 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11069 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11070 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11071 mkdir -p `dirname $F`
11072 ln -s $OPENSSL_SOURCE/$F $F
11075 To be absolutely sure not to disturb the source tree, a "make clean"
11076 is a good thing. If it isn't successful, don't worry about it,
11077 it probably means the source directory is very clean.
11081 * Make sure any ENGINE control commands make local copies of string
11082 pointers passed to them whenever necessary. Otherwise it is possible
11083 the caller may have overwritten (or deallocated) the original string
11084 data when a later ENGINE operation tries to use the stored values.
11086 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11088 * Improve diagnostics in file reading and command-line digests.
11090 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11092 * Add AES modes CFB and OFB to the object database. Correct an
11093 error in AES-CFB decryption.
11097 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11098 allows existing EVP_CIPHER_CTX structures to be reused after
11099 calling EVP_*Final(). This behaviour is used by encryption
11100 BIOs and some applications. This has the side effect that
11101 applications must explicitly clean up cipher contexts with
11102 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11106 * Check the values of dna and dnb in bn_mul_recursive before calling
11107 bn_mul_comba (a non zero value means the a or b arrays do not contain
11108 n2 elements) and fallback to bn_mul_normal if either is not zero.
11112 * Fix escaping of non-ASCII characters when using the -subj option
11113 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11117 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11118 form for "surname", serialNumber has no short form.
11119 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11120 therefore remove "mail" short name for "internet 7".
11121 The OID for unique identifiers in X509 certificates is
11122 x500UniqueIdentifier, not uniqueIdentifier.
11123 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11127 * Add an "init" command to the ENGINE config module and auto initialize
11128 ENGINEs. Without any "init" command the ENGINE will be initialized
11129 after all ctrl commands have been executed on it. If init=1 the
11130 ENGINE is initialized at that point (ctrls before that point are run
11131 on the uninitialized ENGINE and after on the initialized one). If
11132 init=0 then the ENGINE will not be initialized at all.
11136 * Fix the 'app_verify_callback' interface so that the user-defined
11137 argument is actually passed to the callback: In the
11138 SSL_CTX_set_cert_verify_callback() prototype, the callback
11139 declaration has been changed from
11142 int (*cb)(X509_STORE_CTX *,void *);
11143 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11144 i=s->ctx->app_verify_callback(&ctx)
11145 has been changed into
11146 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11148 To update applications using SSL_CTX_set_cert_verify_callback(),
11149 a dummy argument can be added to their callback functions.
11151 *D. K. Smetters <smetters@parc.xerox.com>*
11153 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11155 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11157 * Add and OPENSSL_LOAD_CONF define which will cause
11158 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11159 This allows older applications to transparently support certain
11160 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11161 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11162 load the config file and OPENSSL_add_all_algorithms_conf() which will
11163 always load it have also been added.
11167 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11168 Adjust NIDs and EVP layer.
11170 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11172 * Config modules support in openssl utility.
11174 Most commands now load modules from the config file,
11175 though in a few (such as version) this isn't done
11176 because it couldn't be used for anything.
11178 In the case of ca and req the config file used is
11179 the same as the utility itself: that is the -config
11180 command line option can be used to specify an
11185 * Move default behaviour from OPENSSL_config(). If appname is NULL
11186 use "openssl_conf" if filename is NULL use default openssl config file.
11190 * Add an argument to OPENSSL_config() to allow the use of an alternative
11191 config section name. Add a new flag to tolerate a missing config file
11192 and move code to CONF_modules_load_file().
11196 * Support for crypto accelerator cards from Accelerated Encryption
11197 Processing, www.aep.ie. (Use engine 'aep')
11198 The support was copied from 0.9.6c [engine] and adapted/corrected
11199 to work with the new engine framework.
11201 *AEP Inc. and Richard Levitte*
11203 * Support for SureWare crypto accelerator cards from Baltimore
11204 Technologies. (Use engine 'sureware')
11205 The support was copied from 0.9.6c [engine] and adapted
11206 to work with the new engine framework.
11210 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11211 make the newer ENGINE framework commands for the CHIL engine work.
11213 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11215 * Make it possible to produce shared libraries on ReliantUNIX.
11217 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11219 * Add the configuration target debug-linux-ppro.
11220 Make 'openssl rsa' use the general key loading routines
11221 implemented in apps.c, and make those routines able to
11222 handle the key format FORMAT_NETSCAPE and the variant
11225 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11227 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11229 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11231 * Add -keyform to rsautl, and document -engine.
11233 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11235 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11236 BIO_R_NO_SUCH_FILE error code rather than the generic
11237 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11241 * Add new functions
11242 ERR_peek_last_error
11243 ERR_peek_last_error_line
11244 ERR_peek_last_error_line_data.
11245 These are similar to
11247 ERR_peek_error_line
11248 ERR_peek_error_line_data,
11249 but report on the latest error recorded rather than the first one
11250 still in the error queue.
11252 *Ben Laurie, Bodo Moeller*
11254 * default_algorithms option in ENGINE config module. This allows things
11256 default_algorithms = ALL
11257 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11261 * Preliminary ENGINE config module.
11265 * New experimental application configuration code.
11269 * Change the AES code to follow the same name structure as all other
11270 symmetric ciphers, and behave the same way. Move everything to
11271 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11273 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11275 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11277 *Ben Laurie and Theo de Raadt*
11279 * Add option to output public keys in req command.
11281 *Massimiliano Pala madwolf@openca.org*
11283 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11284 (up to about 10% better than before for P-192 and P-224).
11288 * New functions/macros
11290 SSL_CTX_set_msg_callback(ctx, cb)
11291 SSL_CTX_set_msg_callback_arg(ctx, arg)
11292 SSL_set_msg_callback(ssl, cb)
11293 SSL_set_msg_callback_arg(ssl, arg)
11295 to request calling a callback function
11297 void cb(int write_p, int version, int content_type,
11298 const void *buf, size_t len, SSL *ssl, void *arg)
11300 whenever a protocol message has been completely received
11301 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11302 protocol version according to which the SSL library interprets
11303 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11304 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11305 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11306 specification (change_cipher_spec(20), alert(21), handshake(22)).
11307 'buf' and 'len' point to the actual message, 'ssl' to the
11308 SSL object, and 'arg' is the application-defined value set by
11309 SSL[_CTX]_set_msg_callback_arg().
11311 'openssl s_client' and 'openssl s_server' have new '-msg' options
11312 to enable a callback that displays all protocol messages.
11316 * Change the shared library support so shared libraries are built as
11317 soon as the corresponding static library is finished, and thereby get
11318 openssl and the test programs linked against the shared library.
11319 This still only happens when the keyword "shard" has been given to
11320 the configuration scripts.
11322 NOTE: shared library support is still an experimental thing, and
11323 backward binary compatibility is still not guaranteed.
11325 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11327 * Add support for Subject Information Access extension.
11329 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11331 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11332 additional bytes when new memory had to be allocated, not just
11333 when reusing an existing buffer.
11337 * New command line and configuration option 'utf8' for the req command.
11338 This allows field values to be specified as UTF8 strings.
11342 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11343 runs for the former and machine-readable output for the latter.
11347 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11348 of the e-mail address in the DN (i.e., it will go into a certificate
11349 extension only). The new configuration file option 'email_in_dn = no'
11350 has the same effect.
11352 *Massimiliano Pala madwolf@openca.org*
11354 * Change all functions with names starting with des_ to be starting
11355 with DES_ instead. Add wrappers that are compatible with libdes,
11356 but are named _ossl_old_des_*. Finally, add macros that map the
11357 des_* symbols to the corresponding _ossl_old_des_* if libdes
11358 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11359 desired, the des_* symbols will be mapped to DES_*, with one
11362 Since we provide two compatibility mappings, the user needs to
11363 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11364 compatibility is desired. The default (i.e., when that macro
11365 isn't defined) is OpenSSL 0.9.6c compatibility.
11367 There are also macros that enable and disable the support of old
11368 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11369 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11370 are defined, the default will apply: to support the old des routines.
11372 In either case, one must include openssl/des.h to get the correct
11373 definitions. Do not try to just include openssl/des_old.h, that
11376 NOTE: This is a major break of an old API into a new one. Software
11377 authors are encouraged to switch to the DES_ style functions. Some
11378 time in the future, des_old.h and the libdes compatibility functions
11379 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11380 default), and then completely removed.
11384 * Test for certificates which contain unsupported critical extensions.
11385 If such a certificate is found during a verify operation it is
11386 rejected by default: this behaviour can be overridden by either
11387 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11388 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11389 X509_supported_extension() has also been added which returns 1 if a
11390 particular extension is supported.
11394 * Modify the behaviour of EVP cipher functions in similar way to digests
11395 to retain compatibility with existing code.
11399 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11400 compatibility with existing code. In particular the 'ctx' parameter does
11401 not have to be to be initialized before the call to EVP_DigestInit() and
11402 it is tidied up after a call to EVP_DigestFinal(). New function
11403 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11404 EVP_MD_CTX_copy() changed to not require the destination to be
11405 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11406 requires the destination to be valid.
11408 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11409 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11413 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11414 so that complete 'Handshake' protocol structures are kept in memory
11415 instead of overwriting 'msg_type' and 'length' with 'body' data.
11419 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11421 *Massimo Santin via Richard Levitte*
11423 * Major restructuring to the underlying ENGINE code. This includes
11424 reduction of linker bloat, separation of pure "ENGINE" manipulation
11425 (initialisation, etc) from functionality dealing with implementations
11426 of specific crypto interfaces. This change also introduces integrated
11427 support for symmetric ciphers and digest implementations - so ENGINEs
11428 can now accelerate these by providing EVP_CIPHER and EVP_MD
11429 implementations of their own. This is detailed in crypto/engine/README
11430 as it couldn't be adequately described here. However, there are a few
11431 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11432 were changed in the original introduction of ENGINE code have now
11433 reverted back - the hooking from this code to ENGINE is now a good
11434 deal more passive and at run-time, operations deal directly with
11435 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11436 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11437 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
11438 they were not being used by the framework as there is no concept of a
11439 BIGNUM_METHOD and they could not be generalised to the new
11440 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11441 ENGINE_cpy() has been removed as it cannot be consistently defined in
11446 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11450 * Change mkdef.pl to sort symbols that get the same entry number,
11451 and make sure the automatically generated functions ERR_load_*
11452 become part of libeay.num as well.
11456 * New function SSL_renegotiate_pending(). This returns true once
11457 renegotiation has been requested (either SSL_renegotiate() call
11458 or HelloRequest/ClientHello received from the peer) and becomes
11459 false once a handshake has been completed.
11460 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11461 sends a HelloRequest, but does not ensure that a handshake takes
11462 place. SSL_renegotiate_pending() is useful for checking if the
11463 client has followed the request.)
11467 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11468 By default, clients may request session resumption even during
11469 renegotiation (if session ID contexts permit); with this option,
11470 session resumption is possible only in the first handshake.
11472 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11473 more bits available for options that should not be part of
11474 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11478 * Add some demos for certificate and certificate request creation.
11482 * Make maximum certificate chain size accepted from the peer application
11483 settable (SSL*_get/set_max_cert_list()), as proposed by
11484 "Douglas E. Engert" <deengert@anl.gov>.
11488 * Add support for shared libraries for Unixware-7
11489 (Boyd Lynn Gerber <gerberb@zenez.com>).
11493 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11494 be done prior to destruction. Use this to unload error strings from
11495 ENGINEs that load their own error strings. NB: This adds two new API
11496 functions to "get" and "set" this destroy handler in an ENGINE.
11500 * Alter all existing ENGINE implementations (except "openssl" and
11501 "openbsd") to dynamically instantiate their own error strings. This
11502 makes them more flexible to be built both as statically-linked ENGINEs
11503 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11504 Also, add stub code to each that makes building them as self-contained
11505 shared-libraries easier (see README.ENGINE).
11509 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11510 implementations into applications that are completely implemented in
11511 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11512 commands that can be used to configure what shared-library to load and
11513 to control aspects of the way it is handled. Also, made an update to
11514 the README.ENGINE file that brings its information up-to-date and
11515 provides some information and instructions on the "dynamic" ENGINE
11516 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11520 * Make it possible to unload ranges of ERR strings with a new
11521 "ERR_unload_strings" function.
11525 * Add a copy() function to EVP_MD.
11529 * Make EVP_MD routines take a context pointer instead of just the
11530 md_data void pointer.
11534 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11535 that the digest can only process a single chunk of data
11536 (typically because it is provided by a piece of
11537 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11538 is only going to provide a single chunk of data, and hence the
11539 framework needn't accumulate the data for oneshot drivers.
11543 * As with "ERR", make it possible to replace the underlying "ex_data"
11544 functions. This change also alters the storage and management of global
11545 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11546 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11547 index counters. The API functions that use this state have been changed
11548 to take a "class_index" rather than pointers to the class's local STACK
11549 and counter, and there is now an API function to dynamically create new
11550 classes. This centralisation allows us to (a) plug a lot of the
11551 thread-safety problems that existed, and (b) makes it possible to clean
11552 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11553 such data would previously have always leaked in application code and
11554 workarounds were in place to make the memory debugging turn a blind eye
11555 to it. Application code that doesn't use this new function will still
11556 leak as before, but their memory debugging output will announce it now
11557 rather than letting it slide.
11559 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11560 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11561 has a return value to indicate success or failure.
11565 * Make it possible to replace the underlying "ERR" functions such that the
11566 global state (2 LHASH tables and 2 locks) is only used by the "default"
11567 implementation. This change also adds two functions to "get" and "set"
11568 the implementation prior to it being automatically set the first time
11569 any other ERR function takes place. Ie. an application can call "get",
11570 pass the return value to a module it has just loaded, and that module
11571 can call its own "set" function using that value. This means the
11572 module's "ERR" operations will use (and modify) the error state in the
11573 application and not in its own statically linked copy of OpenSSL code.
11577 * Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
11578 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11579 the operation, and provides a more encapsulated way for external code
11580 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11581 to use these functions rather than manually incrementing the counts.
11583 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11587 * Add EVP test program.
11591 * Add symmetric cipher support to ENGINE. Expect the API to change!
11595 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11596 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11597 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11598 These allow a CRL to be built without having to access X509_CRL fields
11599 directly. Modify 'ca' application to use new functions.
11603 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11604 bug workarounds. Rollback attack detection is a security feature.
11605 The problem will only arise on OpenSSL servers when TLSv1 is not
11606 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11607 Software authors not wanting to support TLSv1 will have special reasons
11608 for their choice and can explicitly enable this option.
11610 *Bodo Moeller, Lutz Jaenicke*
11612 * Rationalise EVP so it can be extended: don't include a union of
11613 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11614 (similar to those existing for EVP_CIPHER_CTX).
11619 EVP_MD_CTX_init(&md); /* new function call */
11620 EVP_DigestInit(&md, EVP_sha1());
11621 EVP_DigestUpdate(&md, in, len);
11622 EVP_DigestFinal(&md, out, NULL);
11623 EVP_MD_CTX_cleanup(&md); /* new function call */
11628 * Make DES key schedule conform to the usual scheme, as well as
11629 correcting its structure. This means that calls to DES functions
11630 now have to pass a pointer to a des_key_schedule instead of a
11631 plain des_key_schedule (which was actually always a pointer
11634 des_key_schedule ks;
11636 des_set_key_checked(..., &ks);
11637 des_ncbc_encrypt(..., &ks, ...);
11639 (Note that a later change renames 'des_...' into 'DES_...'.)
11643 * Initial reduction of linker bloat: the use of some functions, such as
11644 PEM causes large amounts of unused functions to be linked in due to
11645 poor organisation. For example pem_all.c contains every PEM function
11646 which has a knock on effect of linking in large amounts of (unused)
11647 ASN1 code. Grouping together similar functions and splitting unrelated
11648 functions prevents this.
11652 * Cleanup of EVP macros.
11656 * Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
11657 correct _ecb suffix.
11661 * Add initial OCSP responder support to ocsp application. The
11662 revocation information is handled using the text based index
11663 use by the ca application. The responder can either handle
11664 requests generated internally, supplied in files (for example
11665 via a CGI script) or using an internal minimal server.
11669 * Add configuration choices to get zlib compression for TLS.
11673 * Changes to Kerberos SSL for RFC 2712 compliance:
11674 1. Implemented real KerberosWrapper, instead of just using
11675 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11676 2. Implemented optional authenticator field of KerberosWrapper.
11678 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11679 and authenticator structs; see crypto/krb5/.
11681 Generalized Kerberos calls to support multiple Kerberos libraries.
11682 *Vern Staats <staatsvr@asc.hpc.mil>,
11683 Jeffrey Altman <jaltman@columbia.edu>
11684 via Richard Levitte*
11686 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11687 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11688 values for each of the key sizes rather than having just
11689 parameters (and 'speed' generating keys each time).
11693 * Speed up EVP routines.
11696 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11697 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11698 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11699 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11701 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11702 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11703 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11706 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11708 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11712 * Added the OS2-EMX target.
11714 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11716 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11717 to support NCONF routines in extension code. New function CONF_set_nconf()
11718 to allow functions which take an NCONF to also handle the old LHASH
11719 structure: this means that the old CONF compatible routines can be
11720 retained (in particular wrt extensions) without having to duplicate the
11721 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11725 * Enhance the general user interface with mechanisms for inner control
11726 and with possibilities to have yes/no kind of prompts.
11730 * Change all calls to low level digest routines in the library and
11731 applications to use EVP. Add missing calls to HMAC_cleanup() and
11732 don't assume HMAC_CTX can be copied using memcpy().
11734 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11736 * Add the possibility to control engines through control names but with
11737 arbitrary arguments instead of just a string.
11738 Change the key loaders to take a UI_METHOD instead of a callback
11739 function pointer. NOTE: this breaks binary compatibility with earlier
11740 versions of OpenSSL [engine].
11741 Adapt the nCipher code for these new conditions and add a card insertion
11746 * Enhance the general user interface with mechanisms to better support
11747 dialog box interfaces, application-defined prompts, the possibility
11748 to use defaults (for example default passwords from somewhere else)
11749 and interrupts/cancellations.
11753 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11754 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11758 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11759 tidy up some unnecessarily weird code in 'sk_new()').
11761 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11763 * Change the key loading routines for ENGINEs to use the same kind
11764 callback (pem_password_cb) as all other routines that need this
11769 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11770 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11771 than this minimum value is recommended.
11775 * New random seeder for OpenVMS, using the system process statistics
11776 that are easily reachable.
11780 * Windows apparently can't transparently handle global
11781 variables defined in DLLs. Initialisations such as:
11783 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11785 won't compile. This is used by the any applications that need to
11786 declare their own ASN1 modules. This was fixed by adding the option
11787 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11788 needed for static libraries under Win32.
11792 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11793 setting of purpose and trust fields. New X509_STORE trust and
11794 purpose functions and tidy up setting in other SSL functions.
11798 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11799 structure. These are inherited by X509_STORE_CTX when it is
11800 initialised. This allows various defaults to be set in the
11801 X509_STORE structure (such as flags for CRL checking and custom
11802 purpose or trust settings) for functions which only use X509_STORE_CTX
11803 internally such as S/MIME.
11805 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11806 trust settings if they are not set in X509_STORE. This allows X509_STORE
11807 purposes and trust (in S/MIME for example) to override any set by default.
11809 Add command line options for CRL checking to smime, s_client and s_server
11814 * Initial CRL based revocation checking. If the CRL checking flag(s)
11815 are set then the CRL is looked up in the X509_STORE structure and
11816 its validity and signature checked, then if the certificate is found
11817 in the CRL the verify fails with a revoked error.
11819 Various new CRL related callbacks added to X509_STORE_CTX structure.
11821 Command line options added to 'verify' application to support this.
11823 This needs some additional work, such as being able to handle multiple
11824 CRLs with different times, extension based lookup (rather than just
11825 by subject name) and ultimately more complete V2 CRL extension
11830 * Add a general user interface API (crypto/ui/). This is designed
11831 to replace things like des_read_password and friends (backward
11832 compatibility functions using this new API are provided).
11833 The purpose is to remove prompting functions from the DES code
11834 section as well as provide for prompting through dialog boxes in
11835 a window system and the like.
11839 * Add "ex_data" support to ENGINE so implementations can add state at a
11840 per-structure level rather than having to store it globally.
11844 * Make it possible for ENGINE structures to be copied when retrieved by
11845 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
11846 This causes the "original" ENGINE structure to act like a template,
11847 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
11848 operational state can be localised to each ENGINE structure, despite the
11849 fact they all share the same "methods". New ENGINE structures returned in
11850 this case have no functional references and the return value is the single
11851 structural reference. This matches the single structural reference returned
11852 by ENGINE_by_id() normally, when it is incremented on the pre-existing
11857 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
11858 needs to match any other type at all we need to manually clear the
11863 * Changes to the "openssl engine" utility to include;
11864 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
11865 about an ENGINE's available control commands.
11866 - executing control commands from command line arguments using the
11867 '-pre' and '-post' switches. '-post' is only used if '-t' is
11868 specified and the ENGINE is successfully initialised. The syntax for
11869 the individual commands are colon-separated, for example;
11870 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
11874 * New dynamic control command support for ENGINEs. ENGINEs can now
11875 declare their own commands (numbers), names (strings), descriptions,
11876 and input types for run-time discovery by calling applications. A
11877 subset of these commands are implicitly classed as "executable"
11878 depending on their input type, and only these can be invoked through
11879 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
11880 can be based on user input, config files, etc). The distinction is
11881 that "executable" commands cannot return anything other than a boolean
11882 result and can only support numeric or string input, whereas some
11883 discoverable commands may only be for direct use through
11884 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
11885 pointers, or other custom uses. The "executable" commands are to
11886 support parameterisations of ENGINE behaviour that can be
11887 unambiguously defined by ENGINEs and used consistently across any
11888 OpenSSL-based application. Commands have been added to all the
11889 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
11890 control over shared-library paths without source code alterations.
11894 * Changed all ENGINE implementations to dynamically allocate their
11895 ENGINEs rather than declaring them statically. Apart from this being
11896 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
11897 this also allows the implementations to compile without using the
11898 internal engine_int.h header.
11902 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
11903 'const' value. Any code that should be able to modify a RAND_METHOD
11904 should already have non-const pointers to it (ie. they should only
11905 modify their own ones).
11909 * Made a variety of little tweaks to the ENGINE code.
11910 - "atalla" and "ubsec" string definitions were moved from header files
11911 to C code. "nuron" string definitions were placed in variables
11912 rather than hard-coded - allowing parameterisation of these values
11913 later on via ctrl() commands.
11914 - Removed unused "#if 0"'d code.
11915 - Fixed engine list iteration code so it uses ENGINE_free() to release
11916 structural references.
11917 - Constified the RAND_METHOD element of ENGINE structures.
11918 - Constified various get/set functions as appropriate and added
11919 missing functions (including a catch-all ENGINE_cpy that duplicates
11920 all ENGINE values onto a new ENGINE except reference counts/state).
11921 - Removed NULL parameter checks in get/set functions. Setting a method
11922 or function to NULL is a way of cancelling out a previously set
11923 value. Passing a NULL ENGINE parameter is just plain stupid anyway
11924 and doesn't justify the extra error symbols and code.
11925 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
11926 flags from engine_int.h to engine.h.
11927 - Changed prototypes for ENGINE handler functions (init(), finish(),
11928 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
11932 * Implement binary inversion algorithm for BN_mod_inverse in addition
11933 to the algorithm using long division. The binary algorithm can be
11934 used only if the modulus is odd. On 32-bit systems, it is faster
11935 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
11936 roughly 5-15% for 256-bit moduli), so we use it only for moduli
11937 up to 450 bits. In 64-bit environments, the binary algorithm
11938 appears to be advantageous for much longer moduli; here we use it
11939 for moduli up to 2048 bits.
11943 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
11944 could not support the combine flag in choice fields.
11948 * Add a 'copy_extensions' option to the 'ca' utility. This copies
11949 extensions from a certificate request to the certificate.
11953 * Allow multiple 'certopt' and 'nameopt' options to be separated
11954 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
11955 file: this allows the display of the certificate about to be
11956 signed to be customised, to allow certain fields to be included
11957 or excluded and extension details. The old system didn't display
11958 multicharacter strings properly, omitted fields not in the policy
11959 and couldn't display additional details such as extensions.
11963 * Function EC_POINTs_mul for multiple scalar multiplication
11964 of an arbitrary number of elliptic curve points
11965 \sum scalars[i]*points[i],
11966 optionally including the generator defined for the EC_GROUP:
11967 scalar*generator + \sum scalars[i]*points[i].
11969 EC_POINT_mul is a simple wrapper function for the typical case
11970 that the point list has just one item (besides the optional
11975 * First EC_METHODs for curves over GF(p):
11977 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
11978 operations and provides various method functions that can also
11979 operate with faster implementations of modular arithmetic.
11981 EC_GFp_mont_method() reuses most functions that are part of
11982 EC_GFp_simple_method, but uses Montgomery arithmetic.
11984 *Bodo Moeller; point addition and point doubling
11985 implementation directly derived from source code provided by
11986 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
11988 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
11989 crypto/ec/ec_lib.c):
11991 Curves are EC_GROUP objects (with an optional group generator)
11992 based on EC_METHODs that are built into the library.
11994 Points are EC_POINT objects based on EC_GROUP objects.
11996 Most of the framework would be able to handle curves over arbitrary
11997 finite fields, but as there are no obvious types for fields other
11998 than GF(p), some functions are limited to that for now.
12002 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12003 that the file contains a complete HTTP response.
12007 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12008 change the def and num file printf format specifier from "%-40sXXX"
12009 to "%-39s XXX". The latter will always guarantee a space after the
12010 field while the former will cause them to run together if the field
12011 is 40 of more characters long.
12015 * Constify the cipher and digest 'method' functions and structures
12016 and modify related functions to take constant EVP_MD and EVP_CIPHER
12021 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12022 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12026 * Modify EVP_Digest*() routines so they now return values. Although the
12027 internal software routines can never fail additional hardware versions
12032 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12034 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12035 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12038 ERR_R_NESTED_ASN1_ERROR
12040 ERR_R_MISSING_ASN1_EOS
12041 were 4 .. 9, conflicting with
12042 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12044 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12045 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12047 Add new error code 'ERR_R_INTERNAL_ERROR'.
12051 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12056 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12057 sets the subject name for a new request or supersedes the
12058 subject name in a given request. Formats that can be parsed are
12059 'CN=Some Name, OU=myOU, C=IT'
12061 'CN=Some Name/OU=myOU/C=IT'.
12063 Add options '-batch' and '-verbose' to 'openssl req'.
12065 *Massimiliano Pala <madwolf@hackmasters.net>*
12067 * Introduce the possibility to access global variables through
12068 functions on platform were that's the best way to handle exporting
12069 global variables in shared libraries. To enable this functionality,
12070 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12071 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12072 is normally done by Configure or something similar).
12074 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12075 in the source file (foo.c) like this:
12077 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12078 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12080 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12081 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12083 OPENSSL_DECLARE_GLOBAL(int,foo);
12084 #define foo OPENSSL_GLOBAL_REF(foo)
12085 OPENSSL_DECLARE_GLOBAL(double,bar);
12086 #define bar OPENSSL_GLOBAL_REF(bar)
12088 The #defines are very important, and therefore so is including the
12089 header file everywhere where the defined globals are used.
12091 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12092 of ASN.1 items, but that structure is a bit different.
12094 The largest change is in util/mkdef.pl which has been enhanced with
12095 better and easier to understand logic to choose which symbols should
12096 go into the Windows .def files as well as a number of fixes and code
12097 cleanup (among others, algorithm keywords are now sorted
12098 lexicographically to avoid constant rewrites).
12102 * In BN_div() keep a copy of the sign of 'num' before writing the
12103 result to 'rm' because if rm==num the value will be overwritten
12104 and produce the wrong result if 'num' is negative: this caused
12105 problems with BN_mod() and BN_nnmod().
12109 * Function OCSP_request_verify(). This checks the signature on an
12110 OCSP request and verifies the signer certificate. The signer
12111 certificate is just checked for a generic purpose and OCSP request
12116 * Add OCSP_check_validity() function to check the validity of OCSP
12117 responses. OCSP responses are prepared in real time and may only
12118 be a few seconds old. Simply checking that the current time lies
12119 between thisUpdate and nextUpdate max reject otherwise valid responses
12120 caused by either OCSP responder or client clock inaccuracy. Instead
12121 we allow thisUpdate and nextUpdate to fall within a certain period of
12122 the current time. The age of the response can also optionally be
12123 checked. Two new options -validity_period and -status_age added to
12128 * If signature or public key algorithm is unrecognized print out its
12129 OID rather that just UNKNOWN.
12133 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12134 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12135 ID to be generated from the issuer certificate alone which can then be
12136 passed to OCSP_id_issuer_cmp().
12140 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12141 ASN1 modules to export functions returning ASN1_ITEM pointers
12142 instead of the ASN1_ITEM structures themselves. This adds several
12143 new macros which allow the underlying ASN1 function/structure to
12144 be accessed transparently. As a result code should not use ASN1_ITEM
12145 references directly (such as &X509_it) but instead use the relevant
12146 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12147 use of the new ASN1 code on platforms where exporting structures
12148 is problematical (for example in shared libraries) but exporting
12149 functions returning pointers to structures is not.
12153 * Add support for overriding the generation of SSL/TLS session IDs.
12154 These callbacks can be registered either in an SSL_CTX or per SSL.
12155 The purpose of this is to allow applications to control, if they wish,
12156 the arbitrary values chosen for use as session IDs, particularly as it
12157 can be useful for session caching in multiple-server environments. A
12158 command-line switch for testing this (and any client code that wishes
12159 to use such a feature) has been added to "s_server".
12161 *Geoff Thorpe, Lutz Jaenicke*
12163 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12164 of the form '#if defined(...) || defined(...) || ...' and
12165 '#if !defined(...) && !defined(...) && ...'. This also avoids
12166 the growing number of special cases it was previously handling.
12170 * Make all configuration macros available for application by making
12171 sure they are available in opensslconf.h, by giving them names starting
12172 with "OPENSSL_" to avoid conflicts with other packages and by making
12173 sure e_os2.h will cover all platform-specific cases together with
12175 Additionally, it is now possible to define configuration/platform-
12176 specific names (called "system identities"). In the C code, these
12177 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
12178 macro with the name beginning with "OPENSSL_SYS_", which is determined
12179 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
12184 * New option -set_serial to 'req' and 'x509' this allows the serial
12185 number to use to be specified on the command line. Previously self
12186 signed certificates were hard coded with serial number 0 and the
12187 CA options of 'x509' had to use a serial number in a file which was
12192 * New options to 'ca' utility to support V2 CRL entry extensions.
12193 Currently CRL reason, invalidity date and hold instruction are
12194 supported. Add new CRL extensions to V3 code and some new objects.
12198 * New function EVP_CIPHER_CTX_set_padding() this is used to
12199 disable standard block padding (aka PKCS#5 padding) in the EVP
12200 API, which was previously mandatory. This means that the data is
12201 not padded in any way and so the total length much be a multiple
12202 of the block size, otherwise an error occurs.
12206 * Initial (incomplete) OCSP SSL support.
12210 * New function OCSP_parse_url(). This splits up a URL into its host,
12211 port and path components: primarily to parse OCSP URLs. New -url
12212 option to ocsp utility.
12216 * New nonce behavior. The return value of OCSP_check_nonce() now
12217 reflects the various checks performed. Applications can decide
12218 whether to tolerate certain situations such as an absent nonce
12219 in a response when one was present in a request: the ocsp application
12220 just prints out a warning. New function OCSP_add1_basic_nonce()
12221 this is to allow responders to include a nonce in a response even if
12222 the request is nonce-less.
12226 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12227 skipped when using openssl x509 multiple times on a single input file,
12228 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
12232 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12233 set string type: to handle setting ASN1_TIME structures. Fix ca
12234 utility to correctly initialize revocation date of CRLs.
12238 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12239 the clients preferred ciphersuites and rather use its own preferences.
12240 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12241 Internet Explorer by ensuring unchanged hash method during stepup.
12242 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12246 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12247 to aes and add a new 'exist' option to print out symbols that don't
12252 * Additional options to ocsp utility to allow flags to be set and
12253 additional certificates supplied.
12257 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12258 OCSP client a number of certificate to only verify the response
12263 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12264 handle the new API. Currently only ECB, CBC modes supported. Add new
12267 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12268 Encryption Standard (AES) Ciphersuites for Transport Layer
12269 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12270 not enabled by default and were not part of the "ALL" ciphersuite
12271 alias because they were not yet official; they could be
12272 explicitly requested by specifying the "AESdraft" ciphersuite
12273 group alias. In the final release of OpenSSL 0.9.7, the group
12274 alias is called "AES" and is part of "ALL".)
12276 *Ben Laurie, Steve Henson, Bodo Moeller*
12278 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12279 request to response.
12283 * Functions for OCSP responders. OCSP_request_onereq_count(),
12284 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12285 extract information from a certificate request. OCSP_response_create()
12286 creates a response and optionally adds a basic response structure.
12287 OCSP_basic_add1_status() adds a complete single response to a basic
12288 response and returns the OCSP_SINGLERESP structure just added (to allow
12289 extensions to be included for example). OCSP_basic_add1_cert() adds a
12290 certificate to a basic response and OCSP_basic_sign() signs a basic
12291 response with various flags. New helper functions ASN1_TIME_check()
12292 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12293 (converts ASN1_TIME to GeneralizedTime).
12297 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12298 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12299 structure from a certificate. X509_pubkey_digest() digests the public_key
12300 contents: this is used in various key identifiers.
12304 * Make sk_sort() tolerate a NULL argument.
12306 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12308 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12309 passed by the function are trusted implicitly. If any of them signed the
12310 response then it is assumed to be valid and is not verified.
12314 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12315 to data. This was previously part of the PKCS7 ASN1 code. This
12316 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12317 *Steve Henson, reported by Kenneth R. Robinette
12318 <support@securenetterm.com>*
12320 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12321 routines: without these tracing memory leaks is very painful.
12322 Fix leaks in PKCS12 and PKCS7 routines.
12326 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12327 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12328 effectively meant GeneralizedTime would never be used. Now it
12329 is initialised to -1 but X509_time_adj() now has to check the value
12330 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12331 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12332 *Steve Henson, reported by Kenneth R. Robinette
12333 <support@securenetterm.com>*
12335 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12336 result in a zero length in the ASN1_INTEGER structure which was
12337 not consistent with the structure when d2i_ASN1_INTEGER() was used
12338 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12339 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12340 where it did not print out a minus for negative ASN1_INTEGER.
12344 * Add summary printout to ocsp utility. The various functions which
12345 convert status values to strings have been renamed to:
12346 OCSP_response_status_str(), OCSP_cert_status_str() and
12347 OCSP_crl_reason_str() and are no longer static. New options
12348 to verify nonce values and to disable verification. OCSP response
12349 printout format cleaned up.
12353 * Add additional OCSP certificate checks. These are those specified
12354 in RFC2560. This consists of two separate checks: the CA of the
12355 certificate being checked must either be the OCSP signer certificate
12356 or the issuer of the OCSP signer certificate. In the latter case the
12357 OCSP signer certificate must contain the OCSP signing extended key
12358 usage. This check is performed by attempting to match the OCSP
12359 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12360 in the OCSP_CERTID structures of the response.
12364 * Initial OCSP certificate verification added to OCSP_basic_verify()
12365 and related routines. This uses the standard OpenSSL certificate
12366 verify routines to perform initial checks (just CA validity) and
12367 to obtain the certificate chain. Then additional checks will be
12368 performed on the chain. Currently the root CA is checked to see
12369 if it is explicitly trusted for OCSP signing. This is used to set
12370 a root CA as a global signing root: that is any certificate that
12371 chains to that CA is an acceptable OCSP signing certificate.
12375 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12376 extensions from a separate configuration file.
12377 As when reading extensions from the main configuration file,
12378 the '-extensions ...' option may be used for specifying the
12381 *Massimiliano Pala <madwolf@comune.modena.it>*
12383 * New OCSP utility. Allows OCSP requests to be generated or
12384 read. The request can be sent to a responder and the output
12385 parsed, outputted or printed in text form. Not complete yet:
12386 still needs to check the OCSP response validity.
12390 * New subcommands for 'openssl ca':
12391 'openssl ca -status <serial>' prints the status of the cert with
12392 the given serial number (according to the index file).
12393 'openssl ca -updatedb' updates the expiry status of certificates
12396 *Massimiliano Pala <madwolf@comune.modena.it>*
12398 * New '-newreq-nodes' command option to CA.pl. This is like
12399 '-newreq', but calls 'openssl req' with the '-nodes' option
12400 so that the resulting key is not encrypted.
12402 *Damien Miller <djm@mindrot.org>*
12404 * New configuration for the GNU Hurd.
12406 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12408 * Initial code to implement OCSP basic response verify. This
12409 is currently incomplete. Currently just finds the signer's
12410 certificate and verifies the signature on the response.
12414 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12415 value of OPENSSLDIR. This is available via the new '-d' option
12416 to 'openssl version', and is also included in 'openssl version -a'.
12420 * Allowing defining memory allocation callbacks that will be given
12421 file name and line number information in additional arguments
12422 (a const char* and an int). The basic functionality remains, as
12423 well as the original possibility to just replace malloc(),
12424 realloc() and free() by functions that do not know about these
12425 additional arguments. To register and find out the current
12426 settings for extended allocation functions, the following
12427 functions are provided:
12429 CRYPTO_set_mem_ex_functions
12430 CRYPTO_set_locked_mem_ex_functions
12431 CRYPTO_get_mem_ex_functions
12432 CRYPTO_get_locked_mem_ex_functions
12434 These work the same way as CRYPTO_set_mem_functions and friends.
12435 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
12436 extended allocation function is enabled.
12437 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
12438 a conventional allocation function is enabled.
12440 *Richard Levitte, Bodo Moeller*
12442 * Finish off removing the remaining LHASH function pointer casts.
12443 There should no longer be any prototype-casting required when using
12444 the LHASH abstraction, and any casts that remain are "bugs". See
12445 the callback types and macros at the head of lhash.h for details
12446 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12450 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12451 If /dev/[u]random devices are not available or do not return enough
12452 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12454 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12455 /etc/entropy will be queried once each in this sequence, querying stops
12456 when enough entropy was collected without querying more sockets.
12460 * Change the Unix RAND_poll() variant to be able to poll several
12461 random devices, as specified by DEVRANDOM, until a sufficient amount
12462 of data has been collected. We spend at most 10 ms on each file
12463 (select timeout) and read in non-blocking mode. DEVRANDOM now
12464 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12465 (previously it was just the string "/dev/urandom"), so on typical
12466 platforms the 10 ms delay will never occur.
12467 Also separate out the Unix variant to its own file, rand_unix.c.
12468 For VMS, there's a currently-empty rand_vms.c.
12472 * Move OCSP client related routines to ocsp_cl.c. These
12473 provide utility functions which an application needing
12474 to issue a request to an OCSP responder and analyse the
12475 response will typically need: as opposed to those which an
12476 OCSP responder itself would need which will be added later.
12478 OCSP_request_sign() signs an OCSP request with an API similar
12479 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12480 response. OCSP_response_get1_basic() extracts basic response
12481 from response. OCSP_resp_find_status(): finds and extracts status
12482 information from an OCSP_CERTID structure (which will be created
12483 when the request structure is built). These are built from lower
12484 level functions which work on OCSP_SINGLERESP structures but
12485 won't normally be used unless the application wishes to examine
12486 extensions in the OCSP response for example.
12488 Replace nonce routines with a pair of functions.
12489 OCSP_request_add1_nonce() adds a nonce value and optionally
12490 generates a random value. OCSP_check_nonce() checks the
12491 validity of the nonce in an OCSP response.
12495 * Change function OCSP_request_add() to OCSP_request_add0_id().
12496 This doesn't copy the supplied OCSP_CERTID and avoids the
12497 need to free up the newly created id. Change return type
12498 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12499 This can then be used to add extensions to the request.
12500 Deleted OCSP_request_new(), since most of its functionality
12501 is now in OCSP_REQUEST_new() (and the case insensitive name
12502 clash) apart from the ability to set the request name which
12503 will be added elsewhere.
12507 * Update OCSP API. Remove obsolete extensions argument from
12508 various functions. Extensions are now handled using the new
12509 OCSP extension code. New simple OCSP HTTP function which
12510 can be used to send requests and parse the response.
12514 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12515 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12516 uses the special reorder version of SET OF to sort the attributes
12517 and reorder them to match the encoded order. This resolves a long
12518 standing problem: a verify on a PKCS7 structure just after signing
12519 it used to fail because the attribute order did not match the
12520 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12521 it uses the received order. This is necessary to tolerate some broken
12522 software that does not order SET OF. This is handled by encoding
12523 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12524 to produce the required SET OF.
12528 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12529 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12530 files to get correct declarations of the ASN.1 item variables.
12534 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12535 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12536 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12537 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12538 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12539 ASN1_ITEM and no wrapper functions.
12543 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12544 replace the old function pointer based I/O routines. Change most of
12545 the *_d2i_bio() and *_d2i_fp() functions to use these.
12549 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12550 lines, recognize more "algorithms" that can be deselected, and make
12551 it complain about algorithm deselection that isn't recognised.
12555 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12556 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12557 to use new functions. Add NO_ASN1_OLD which can be set to remove
12558 some old style ASN1 functions: this can be used to determine if old
12559 code will still work when these eventually go away.
12563 * New extension functions for OCSP structures, these follow the
12564 same conventions as certificates and CRLs.
12568 * New function X509V3_add1_i2d(). This automatically encodes and
12569 adds an extension. Its behaviour can be customised with various
12570 flags to append, replace or delete. Various wrappers added for
12571 certificates and CRLs.
12575 * Fix to avoid calling the underlying ASN1 print routine when
12576 an extension cannot be parsed. Correct a typo in the
12577 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12581 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12582 entries for variables.
12586 * Add functionality to apps/openssl.c for detecting locking
12587 problems: As the program is single-threaded, all we have
12588 to do is register a locking callback using an array for
12589 storing which locks are currently held by the program.
12593 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12594 SSL_get_ex_data_X509_STORE_idx(), which is used in
12595 ssl_verify_cert_chain() and thus can be called at any time
12596 during TLS/SSL handshakes so that thread-safety is essential.
12597 Unfortunately, the ex_data design is not at all suited
12598 for multi-threaded use, so it probably should be abolished.
12602 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12604 *Broadcom, tweaked and integrated by Geoff Thorpe*
12606 * Move common extension printing code to new function
12607 X509V3_print_extensions(). Reorganise OCSP print routines and
12608 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12612 * New function X509_signature_print() to remove duplication in some
12617 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12618 set (this was treated exactly the same as SET OF previously). This
12619 is used to reorder the STACK representing the structure to match the
12620 encoding. This will be used to get round a problem where a PKCS7
12621 structure which was signed could not be verified because the STACK
12622 order did not reflect the encoded order.
12626 * Reimplement the OCSP ASN1 module using the new code.
12630 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12631 for its ASN1 operations. The old style function pointers still exist
12632 for now but they will eventually go away.
12636 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12637 completely replaces the old ASN1 functionality with a table driven
12638 encoder and decoder which interprets an ASN1_ITEM structure describing
12639 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12640 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12641 has also been converted to the new form.
12645 * Change BN_mod_exp_recp so that negative moduli are tolerated
12646 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12647 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12648 for negative moduli.
12652 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12653 of not touching the result's sign bit.
12657 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12662 * Changed the LHASH code to use prototypes for callbacks, and created
12663 macros to declare and implement thin (optionally static) functions
12664 that provide type-safety and avoid function pointer casting for the
12665 type-specific callbacks.
12669 * Added Kerberos Cipher Suites to be used with TLS, as written in
12671 *Veers Staats <staatsvr@asc.hpc.mil>,
12672 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12674 * Reformat the FAQ so the different questions and answers can be divided
12675 in sections depending on the subject.
12679 * Have the zlib compression code load ZLIB.DLL dynamically under
12684 * New function BN_mod_sqrt for computing square roots modulo a prime
12685 (using the probabilistic Tonelli-Shanks algorithm unless
12686 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12687 be handled deterministically).
12689 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12691 * Make BN_mod_inverse faster by explicitly handling small quotients
12692 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12693 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12697 * New function BN_kronecker.
12701 * Fix BN_gcd so that it works on negative inputs; the result is
12702 positive unless both parameters are zero.
12703 Previously something reasonably close to an infinite loop was
12704 possible because numbers could be growing instead of shrinking
12705 in the implementation of Euclid's algorithm.
12709 * Fix BN_is_word() and BN_is_one() macros to take into account the
12710 sign of the number in question.
12712 Fix BN_is_word(a,w) to work correctly for w == 0.
12714 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12715 because its test if the absolute value of 'a' equals 'w'.
12716 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12717 it exists mostly for use in the implementations of BN_is_zero(),
12718 BN_is_one(), and BN_is_word().
12722 * New function BN_swap.
12726 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12727 the exponentiation functions are more likely to produce reasonable
12728 results on negative inputs.
12732 * Change BN_mod_mul so that the result is always non-negative.
12733 Previously, it could be negative if one of the factors was negative;
12734 I don't think anyone really wanted that behaviour.
12738 * Move BN_mod_... functions into new file crypto/bn/bn_mod.c
12739 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
12740 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
12741 and add new functions:
12750 BN_mod_lshift1_quick
12752 BN_mod_lshift_quick
12754 These functions always generate non-negative results.
12756 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
12757 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
12759 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
12760 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
12761 be reduced modulo m.
12763 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12766 The following entry accidentally appeared in the CHANGES file
12767 distributed with OpenSSL 0.9.7. The modifications described in
12768 it do *not* apply to OpenSSL 0.9.7.
12770 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12771 was actually never needed) and in BN_mul(). The removal in BN_mul()
12772 required a small change in bn_mul_part_recursive() and the addition
12773 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12774 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12775 bn_sub_words() and bn_add_words() except they take arrays with
12781 * In 'openssl passwd', verify passwords read from the terminal
12782 unless the '-salt' option is used (which usually means that
12783 verification would just waste user's time since the resulting
12784 hash is going to be compared with some given password hash)
12785 or the new '-noverify' option is used.
12787 This is an incompatible change, but it does not affect
12788 non-interactive use of 'openssl passwd' (passwords on the command
12789 line, '-stdin' option, '-in ...' option) and thus should not
12790 cause any problems.
12794 * Remove all references to RSAref, since there's no more need for it.
12798 * Make DSO load along a path given through an environment variable
12799 (SHLIB_PATH) with shl_load().
12803 * Constify the ENGINE code as a result of BIGNUM constification.
12804 Also constify the RSA code and most things related to it. In a
12805 few places, most notable in the depth of the ASN.1 code, ugly
12806 casts back to non-const were required (to be solved at a later
12811 * Make it so the openssl application has all engines loaded by default.
12815 * Constify the BIGNUM routines a little more.
12819 * Add the following functions:
12821 ENGINE_load_cswift()
12823 ENGINE_load_atalla()
12824 ENGINE_load_nuron()
12825 ENGINE_load_builtin_engines()
12827 That way, an application can itself choose if external engines that
12828 are built-in in OpenSSL shall ever be used or not. The benefit is
12829 that applications won't have to be linked with libdl or other dso
12830 libraries unless it's really needed.
12832 Changed 'openssl engine' to load all engines on demand.
12833 Changed the engine header files to avoid the duplication of some
12834 declarations (they differed!).
12838 * 'openssl engine' can now list capabilities.
12842 * Better error reporting in 'openssl engine'.
12846 * Never call load_dh_param(NULL) in s_server.
12850 * Add engine application. It can currently list engines by name and
12851 identity, and test if they are actually available.
12855 * Improve RPM specification file by forcing symbolic linking and making
12856 sure the installed documentation is also owned by root.root.
12858 *Damien Miller <djm@mindrot.org>*
12860 * Give the OpenSSL applications more possibilities to make use of
12861 keys (public as well as private) handled by engines.
12865 * Add OCSP code that comes from CertCo.
12869 * Add VMS support for the Rijndael code.
12873 * Added untested support for Nuron crypto accelerator.
12877 * Add support for external cryptographic devices. This code was
12878 previously distributed separately as the "engine" branch.
12880 *Geoff Thorpe, Richard Levitte*
12882 * Rework the filename-translation in the DSO code. It is now possible to
12883 have far greater control over how a "name" is turned into a filename
12884 depending on the operating environment and any oddities about the
12885 different shared library filenames on each system.
12889 * Support threads on FreeBSD-elf in Configure.
12893 * Fix for SHA1 assembly problem with MASM: it produces
12894 warnings about corrupt line number information when assembling
12895 with debugging information. This is caused by the overlapping
12898 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
12901 NCONF_get_number() has no error checking at all. As a replacement,
12902 NCONF_get_number_e() is defined (_e for "error checking") and is
12903 promoted strongly. The old NCONF_get_number is kept around for
12904 binary backward compatibility.
12905 Make it possible for methods to load from something other than a BIO,
12906 by providing a function pointer that is given a name instead of a BIO.
12907 For example, this could be used to load configuration data from an
12912 * Fix for non blocking accept BIOs. Added new I/O special reason
12913 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
12914 with non blocking I/O was not possible because no retry code was
12915 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
12920 * Added the beginnings of Rijndael support.
12924 * Fix for bug in DirectoryString mask setting. Add support for
12925 X509_NAME_print_ex() in 'req' and X509_print_ex() function
12926 to allow certificate printing to more controllable, additional
12927 'certopt' option to 'x509' to allow new printing options to be
12932 * Clean old EAY MD5 hack from e_os.h.
12936 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004] ###
12938 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12939 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
12941 *Joe Orton, Steve Henson*
12943 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003] ###
12945 * Fix additional bug revealed by the NISCC test suite:
12947 Stop bug triggering large recursion when presented with
12948 certain ASN.1 tags [CVE-2003-0851][]
12952 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003] ###
12954 * Fix various bugs revealed by running the NISCC test suite:
12956 Stop out of bounds reads in the ASN1 code when presented with
12957 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12959 If verify callback ignores invalid public key errors don't try to check
12960 certificate signature with the NULL public key.
12965 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12966 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12971 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12972 extra data after the compression methods not only for TLS 1.0
12973 but also for SSL 3.0 (as required by the specification).
12975 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12977 * Change X509_certificate_type() to mark the key as exported/exportable
12978 when it's 512 *bits* long, not 512 bytes.
12982 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003] ###
12984 * Countermeasure against the Klima-Pokorny-Rosa extension of
12985 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12986 a protocol version number mismatch like a decryption error
12987 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12991 * Turn on RSA blinding by default in the default implementation
12992 to avoid a timing attack. Applications that don't want it can call
12993 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12994 They would be ill-advised to do so in most cases.
12996 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12998 * Change RSA blinding code so that it works when the PRNG is not
12999 seeded (in this case, the secret RSA exponent is abused as
13000 an unpredictable seed -- if it is not unpredictable, there
13001 is no point in blinding anyway). Make RSA blinding thread-safe
13002 by remembering the creator's thread ID in rsa->blinding and
13003 having all other threads use local one-time blinding factors
13004 (this requires more computation than sharing rsa->blinding, but
13005 avoids excessive locking; and if an RSA object is not shared
13006 between threads, blinding will still be very fast).
13010 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003] ###
13012 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13013 via timing by performing a MAC computation even if incorrect
13014 block cipher padding has been found. This is a countermeasure
13015 against active attacks where the attacker has to distinguish
13016 between bad padding and a MAC verification error. [CVE-2003-0078][]
13018 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13019 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13020 Martin Vuagnoux (EPFL, Ilion)*
13022 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002] ###
13024 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13025 memory from its contents. This is done with a counter that will
13026 place alternating values in each byte. This can be used to solve
13027 two issues: 1) the removal of calls to memset() by highly optimizing
13028 compilers, and 2) cleansing with other values than 0, since those can
13029 be read through on certain media, for example a swap space on disk.
13033 * Bugfix: client side session caching did not work with external caching,
13034 because the session->cipher setting was not restored when reloading
13035 from the external cache. This problem was masked, when
13036 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13037 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13041 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13042 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13044 *Zeev Lieber <zeev-l@yahoo.com>*
13046 * Undo an undocumented change introduced in 0.9.6e which caused
13047 repeated calls to OpenSSL_add_all_ciphers() and
13048 OpenSSL_add_all_digests() to be ignored, even after calling
13053 * Change the default configuration reader to deal with last line not
13054 being properly terminated.
13058 * Change X509_NAME_cmp() so it applies the special rules on handling
13059 DN values that are of type PrintableString, as well as RDNs of type
13060 emailAddress where the value has the type ia5String.
13062 *stefank@valicert.com via Richard Levitte*
13064 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13065 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13066 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13067 the bitwise-OR of the two for use by the majority of applications
13068 wanting this behaviour, and update the docs. The documented
13069 behaviour and actual behaviour were inconsistent and had been
13070 changing anyway, so this is more a bug-fix than a behavioural
13073 *Geoff Thorpe, diagnosed by Nadav Har'El*
13075 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13076 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13080 * Fix initialization code race conditions in
13081 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13082 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13083 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13084 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13085 ssl2_get_cipher_by_char(),
13086 ssl3_get_cipher_by_char().
13088 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13090 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13091 the cached sessions are flushed, as the remove_cb() might use ex_data
13092 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13093 (see [openssl.org #212]).
13095 *Geoff Thorpe, Lutz Jaenicke*
13097 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13098 length, instead of the encoding length to d2i_ASN1_OBJECT.
13102 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002] ###
13104 * [In 0.9.6g-engine release:]
13105 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
13107 *Lynn Gazis <lgazis@rainbow.com>*
13109 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002] ###
13111 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13112 and get fix the header length calculation.
13113 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13114 Alon Kantor <alonk@checkpoint.com> (and others),
13117 * Use proper error handling instead of 'assertions' in buffer
13118 overflow checks added in 0.9.6e. This prevents DoS (the
13119 assertions could call abort()).
13121 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13123 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002] ###
13125 * Add various sanity checks to asn1_get_length() to reject
13126 the ASN1 length bytes if they exceed sizeof(long), will appear
13127 negative or the content length exceeds the length of the
13130 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13132 * Fix cipher selection routines: ciphers without encryption had no flags
13133 for the cipher strength set and where therefore not handled correctly
13134 by the selection routines (PR #130).
13138 * Fix EVP_dsa_sha macro.
13143 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13144 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13145 that was added in OpenSSL 0.9.6d.
13147 As the countermeasure turned out to be incompatible with some
13148 broken SSL implementations, the new option is part of SSL_OP_ALL.
13149 SSL_OP_ALL is usually employed when compatibility with weird SSL
13150 implementations is desired (e.g. '-bugs' option to 's_client' and
13151 's_server'), so the new option is automatically set in many
13156 * Changes in security patch:
13158 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13159 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13160 Air Force Materiel Command, USAF, under agreement number
13163 * Add various sanity checks to asn1_get_length() to reject
13164 the ASN1 length bytes if they exceed sizeof(long), will appear
13165 negative or the content length exceeds the length of the
13166 supplied buffer. [CVE-2002-0659][]
13168 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13170 * Assertions for various potential buffer overflows, not known to
13171 happen in practice.
13173 *Ben Laurie (CHATS)*
13175 * Various temporary buffers to hold ASCII versions of integers were
13176 too small for 64 bit platforms. [CVE-2002-0655][]
13177 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
13179 * Remote buffer overflow in SSL3 protocol - an attacker could
13180 supply an oversized session ID to a client. [CVE-2002-0656][]
13182 *Ben Laurie (CHATS)*
13184 * Remote buffer overflow in SSL2 protocol - an attacker could
13185 supply an oversized client master key. [CVE-2002-0656][]
13187 *Ben Laurie (CHATS)*
13189 ### Changes between 0.9.6c and 0.9.6d [9 May 2002] ###
13191 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13192 encoded as NULL) with id-dsa-with-sha1.
13194 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13196 * Check various X509_...() return values in apps/req.c.
13198 *Nils Larsch <nla@trustcenter.de>*
13200 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13201 an end-of-file condition would erroneously be flagged, when the CRLF
13202 was just at the end of a processed block. The bug was discovered when
13203 processing data through a buffering memory BIO handing the data to a
13204 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13205 <ptsekov@syntrex.com> and Nedelcho Stanev.
13209 * Implement a countermeasure against a vulnerability recently found
13210 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13211 before application data chunks to avoid the use of known IVs
13212 with data potentially chosen by the attacker.
13216 * Fix length checks in ssl3_get_client_hello().
13220 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13221 to prevent ssl3_read_internal() from incorrectly assuming that
13222 ssl3_read_bytes() found application data while handshake
13223 processing was enabled when in fact s->s3->in_read_app_data was
13224 merely automatically cleared during the initial handshake.
13226 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13228 * Fix object definitions for Private and Enterprise: they were not
13229 recognized in their shortname (=lowercase) representation. Extend
13230 obj_dat.pl to issue an error when using undefined keywords instead
13231 of silently ignoring the problem (Svenning Sorensen
13232 <sss@sss.dnsalias.net>).
13236 * Fix DH_generate_parameters() so that it works for 'non-standard'
13237 generators, i.e. generators other than 2 and 5. (Previously, the
13238 code did not properly initialise the 'add' and 'rem' values to
13239 BN_generate_prime().)
13241 In the new general case, we do not insist that 'generator' is
13242 actually a primitive root: This requirement is rather pointless;
13243 a generator of the order-q subgroup is just as good, if not
13248 * Map new X509 verification errors to alerts. Discovered and submitted by
13249 Tom Wu <tom@arcot.com>.
13253 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13254 returning non-zero before the data has been completely received
13255 when using non-blocking I/O.
13257 *Bodo Moeller; problem pointed out by John Hughes*
13259 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13261 *Ben Laurie, Lutz Jaenicke*
13263 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13264 Yoram Zahavi <YoramZ@gilian.com>).
13268 * Add information about CygWin 1.3 and on, and preserve proper
13269 configuration for the versions before that.
13271 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13273 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13274 check whether we deal with a copy of a session and do not delete from
13275 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13276 <izhar@checkpoint.com>.
13280 * Do not store session data into the internal session cache, if it
13281 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13282 flag is set). Proposed by Aslam <aslam@funk.com>.
13286 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13291 * [In 0.9.6d-engine release:]
13292 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13294 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13296 * Add the configuration target linux-s390x.
13298 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13300 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13301 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13302 variable as an indication that a ClientHello message has been
13303 received. As the flag value will be lost between multiple
13304 invocations of ssl3_accept when using non-blocking I/O, the
13305 function may not be aware that a handshake has actually taken
13306 place, thus preventing a new session from being added to the
13309 To avoid this problem, we now set s->new_session to 2 instead of
13310 using a local variable.
13312 *Lutz Jaenicke, Bodo Moeller*
13314 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13315 if the SSL_R_LENGTH_MISMATCH error is detected.
13317 *Geoff Thorpe, Bodo Moeller*
13319 * New 'shared_ldflag' column in Configure platform table.
13323 * Fix EVP_CIPHER_mode macro.
13325 *"Dan S. Camper" <dan@bti.net>*
13327 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13328 type, we must throw them away by setting rr->length to 0.
13330 *D P Chang <dpc@qualys.com>*
13332 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001] ###
13334 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13335 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13336 worked incorrectly for those cases where range = 10..._2 and
13337 3*range is two bits longer than range.)
13341 * Only add signing time to PKCS7 structures if it is not already
13346 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13347 OBJ_ld_ce should be OBJ_id_ce.
13348 Also some ip-pda OIDs in crypto/objects/objects.txt were
13349 incorrect (cf. RFC 3039).
13351 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13353 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13354 returns early because it has nothing to do.
13356 *Andy Schneider <andy.schneider@bjss.co.uk>*
13358 * [In 0.9.6c-engine release:]
13359 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13361 *Andy Schneider <andy.schneider@bjss.co.uk>*
13363 * [In 0.9.6c-engine release:]
13364 Add support for Cryptographic Appliance's keyserver technology.
13365 (Use engine 'keyclient')
13367 *Cryptographic Appliances and Geoff Thorpe*
13369 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13370 is called via tools/c89.sh because arguments have to be
13371 rearranged (all '-L' options must appear before the first object
13374 *Richard Shapiro <rshapiro@abinitio.com>*
13376 * [In 0.9.6c-engine release:]
13377 Add support for Broadcom crypto accelerator cards, backported
13380 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13382 * [In 0.9.6c-engine release:]
13383 Add support for SureWare crypto accelerator cards from
13384 Baltimore Technologies. (Use engine 'sureware')
13386 *Baltimore Technologies and Mark Cox*
13388 * [In 0.9.6c-engine release:]
13389 Add support for crypto accelerator cards from Accelerated
13390 Encryption Processing, www.aep.ie. (Use engine 'aep')
13392 *AEP Inc. and Mark Cox*
13394 * Add a configuration entry for gcc on UnixWare.
13396 *Gary Benson <gbenson@redhat.com>*
13398 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13399 messages are stored in a single piece (fixed-length part and
13400 variable-length part combined) and fix various bugs found on the way.
13404 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13405 instead. BIO_gethostbyname() does not know what timeouts are
13406 appropriate, so entries would stay in cache even when they have
13408 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
13410 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13411 faced with a pathologically small ClientHello fragment that does
13412 not contain client_version: Instead of aborting with an error,
13413 simply choose the highest available protocol version (i.e.,
13414 TLS 1.0 unless it is disabled). In practice, ClientHello
13415 messages are never sent like this, but this change gives us
13416 strictly correct behaviour at least for TLS.
13420 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13421 never resets s->method to s->ctx->method when called from within
13422 one of the SSL handshake functions.
13424 *Bodo Moeller; problem pointed out by Niko Baric*
13426 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13427 (sent using the client's version number) if client_version is
13428 smaller than the protocol version in use. Also change
13429 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13430 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13431 the client will at least see that alert.
13435 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13440 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13441 client receives HelloRequest while in a handshake.
13443 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13445 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13446 should end in 'break', not 'goto end' which circumvents various
13447 cleanups done in state SSL_ST_OK. But session related stuff
13448 must be disabled for SSL_ST_OK in the case that we just sent a
13451 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13452 before just sending a HelloRequest.
13454 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13456 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13457 reveal whether illegal block cipher padding was found or a MAC
13458 verification error occurred. (Neither SSLerr() codes nor alerts
13459 are directly visible to potential attackers, but the information
13460 may leak via logfiles.)
13462 Similar changes are not required for the SSL 2.0 implementation
13463 because the number of padding bytes is sent in clear for SSL 2.0,
13464 and the extra bytes are just ignored. However ssl/s2_pkt.c
13465 failed to verify that the purported number of padding bytes is in
13470 * Add OpenUNIX-8 support including shared libraries
13471 (Boyd Lynn Gerber <gerberb@zenez.com>).
13475 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13476 'wristwatch attack' using huge encoding parameters (cf.
13477 James H. Manger's CRYPTO 2001 paper). Note that the
13478 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13479 encoding parameters and hence was not vulnerable.
13483 * BN_sqr() bug fix.
13485 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13487 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13488 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13489 followed by modular reduction.
13491 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13493 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13494 equivalent based on BN_pseudo_rand() instead of BN_rand().
13498 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13499 This function was broken, as the check for a new client hello message
13500 to handle SGC did not allow these large messages.
13501 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13505 * Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
13509 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13510 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13514 * Rework the configuration and shared library support for Tru64 Unix.
13515 The configuration part makes use of modern compiler features and
13516 still retains old compiler behavior for those that run older versions
13517 of the OS. The shared library support part includes a variant that
13518 uses the RPATH feature, and is available through the special
13519 configuration target "alpha-cc-rpath", which will never be selected
13522 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13524 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13525 with the same message size as in ssl3_get_certificate_request().
13526 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13527 messages might inadvertently be reject as too long.
13529 *Petr Lampa <lampa@fee.vutbr.cz>*
13531 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13535 * Modified SSL library such that the verify_callback that has been set
13536 specifically for an SSL object with SSL_set_verify() is actually being
13537 used. Before the change, a verify_callback set with this function was
13538 ignored and the verify_callback() set in the SSL_CTX at the time of
13539 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13540 to allow the necessary settings.
13544 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13545 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13546 done automatically (in contradiction to the requirements of the C
13547 standard). This made problems when used from OpenSSH.
13551 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13552 dh->length and always used
13554 BN_rand_range(priv_key, dh->p).
13556 BN_rand_range() is not necessary for Diffie-Hellman, and this
13557 specific range makes Diffie-Hellman unnecessarily inefficient if
13558 dh->length (recommended exponent length) is much smaller than the
13559 length of dh->p. We could use BN_rand_range() if the order of
13560 the subgroup was stored in the DH structure, but we only have
13565 BN_rand(priv_key, l, ...)
13567 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13574 RSA_eay_public_encrypt
13575 RSA_eay_private_decrypt
13576 RSA_eay_private_encrypt (signing)
13577 RSA_eay_public_decrypt (signature verification)
13579 (default implementations for RSA_public_encrypt,
13580 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13581 always reject numbers >= n.
13585 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13586 to synchronize access to 'locking_thread'. This is necessary on
13587 systems where access to 'locking_thread' (an 'unsigned long'
13588 variable) is not atomic.
13592 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13593 *before* setting the 'crypto_lock_rand' flag. The previous code had
13594 a race condition if 0 is a valid thread ID.
13596 *Travis Vitek <vitek@roguewave.com>*
13598 * Add support for shared libraries under Irix.
13600 *Albert Chin-A-Young <china@thewrittenword.com>*
13602 * Add configuration option to build on Linux on both big-endian and
13603 little-endian MIPS.
13605 *Ralf Baechle <ralf@uni-koblenz.de>*
13607 * Add the possibility to create shared libraries on HP-UX.
13611 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001] ###
13613 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13614 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13615 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13616 PRNG state recovery was possible based on the output of
13617 one PRNG request appropriately sized to gain knowledge on
13618 'md' followed by enough consecutive 1-byte PRNG requests
13619 to traverse all of 'state'.
13621 1. When updating 'md_local' (the current thread's copy of 'md')
13622 during PRNG output generation, hash all of the previous
13623 'md_local' value, not just the half used for PRNG output.
13625 2. Make the number of bytes from 'state' included into the hash
13626 independent from the number of PRNG bytes requested.
13628 The first measure alone would be sufficient to avoid
13629 Markku-Juhani's attack. (Actually it had never occurred
13630 to me that the half of 'md_local' used for chaining was the
13631 half from which PRNG output bytes were taken -- I had always
13632 assumed that the secret half would be used.) The second
13633 measure makes sure that additional data from 'state' is never
13634 mixed into 'md_local' in small portions; this heuristically
13635 further strengthens the PRNG.
13639 * Fix crypto/bn/asm/mips3.s.
13643 * When only the key is given to "enc", the IV is undefined. Print out
13644 an error message in this case.
13648 * Handle special case when X509_NAME is empty in X509 printing routines.
13652 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13653 positive and less than q.
13657 * Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
13658 used: it isn't thread safe and the add_lock_callback should handle
13661 *Paul Rose <Paul.Rose@bridge.com>*
13663 * Verify that incoming data obeys the block size in
13664 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13670 *Ulf Möller, Bodo Möller*
13672 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13673 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13674 when fixing the server behaviour for backwards-compatible 'client
13675 hello' messages. (Note that the attack is impractical against
13676 SSL 3.0 and TLS 1.0 anyway because length and version checking
13677 means that the probability of guessing a valid ciphertext is
13678 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13681 Before 0.9.5, the countermeasure (hide the error by generating a
13682 random 'decryption result') did not work properly because
13683 ERR_clear_error() was missing, meaning that SSL_get_error() would
13684 detect the supposedly ignored error.
13686 Both problems are now fixed.
13690 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13691 (previously it was 1024).
13695 * Fix for compatibility mode trust settings: ignore trust settings
13696 unless some valid trust or reject settings are present.
13700 * Fix for blowfish EVP: its a variable length cipher.
13704 * Fix various bugs related to DSA S/MIME verification. Handle missing
13705 parameters in DSA public key structures and return an error in the
13706 DSA routines if parameters are absent.
13710 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13711 in the current directory if neither $RANDFILE nor $HOME was set.
13712 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13713 caused some confusion to Windows users who haven't defined $HOME.
13714 Thus RAND_file_name() is changed again: e_os.h can define a
13715 DEFAULT_HOME, which will be used if $HOME is not set.
13716 For Windows, we use "C:"; on other platforms, we still require
13717 environment variables.
13719 * Move 'if (!initialized) RAND_poll()' into regions protected by
13720 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13721 having multiple threads call RAND_poll() concurrently.
13725 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13726 combination of a flag and a thread ID variable.
13727 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13728 flag), *other* threads can enter ssleay_add_bytes without obeying
13729 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13730 that they do not hold after the first thread unsets add_do_not_lock).
13734 * Change bctest again: '-x' expressions are not available in all
13735 versions of 'test'.
13739 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001] ###
13741 * Fix a couple of memory leaks in PKCS7_dataDecode()
13743 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13745 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13746 the default extension for executables, if any. Also, make the perl
13747 scripts that use symlink() to test if it really exists and use "cp"
13748 if it doesn't. All this made OpenSSL compilable and installable in
13753 * Fix for asn1_GetSequence() for indefinite length constructed data.
13754 If SEQUENCE is length is indefinite just set c->slen to the total
13755 amount of data available.
13757 *Steve Henson, reported by shige@FreeBSD.org*
13759 *This change does not apply to 0.9.7.*
13761 * Change bctest to avoid here-documents inside command substitution
13762 (workaround for FreeBSD /bin/sh bug).
13763 For compatibility with Ultrix, avoid shell functions (introduced
13764 in the bctest version that searches along $PATH).
13768 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13769 with des_encrypt() defined on some operating systems, like Solaris
13774 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13775 On the Importance of Eliminating Errors in Cryptographic
13776 Computations, J. Cryptology 14 (2001) 2, 101-119,
13777 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
13781 * MIPS assembler BIGNUM division bug fix.
13785 * Disabled incorrect Alpha assembler code.
13789 * Fix PKCS#7 decode routines so they correctly update the length
13790 after reading an EOC for the EXPLICIT tag.
13794 *This change does not apply to 0.9.7.*
13796 * Fix bug in PKCS#12 key generation routines. This was triggered
13797 if a 3DES key was generated with a 0 initial byte. Include
13798 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13799 (but broken) behaviour.
13803 * Enhance bctest to search for a working bc along $PATH and print
13806 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13808 * Fix memory leaks in err.c: free err_data string if necessary;
13809 don't write to the wrong index in ERR_set_error_data.
13813 * Implement ssl23_peek (analogous to ssl23_read), which previously
13818 * Replace rdtsc with _emit statements for VC++ version 5.
13820 *Jeremy Cooper <jeremy@baymoo.org>*
13822 * Make it possible to reuse SSLv2 sessions.
13826 * In copy_email() check for >= 0 as a return value for
13827 X509_NAME_get_index_by_NID() since 0 is a valid index.
13829 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
13831 * Avoid coredump with unsupported or invalid public keys by checking if
13832 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
13833 PKCS7_verify() fails with non detached data.
13837 * Don't use getenv in library functions when run as setuid/setgid.
13838 New function OPENSSL_issetugid().
13842 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
13843 due to incorrect handling of multi-threading:
13845 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
13847 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
13849 3. Count how many times MemCheck_off() has been called so that
13850 nested use can be treated correctly. This also avoids
13851 inband-signalling in the previous code (which relied on the
13852 assumption that thread ID 0 is impossible).
13856 * Add "-rand" option also to s_client and s_server.
13860 * Fix CPU detection on Irix 6.x.
13861 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
13862 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13864 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
13869 *This change does not apply to 0.9.7.*
13871 * Use the cached encoding of an X509_NAME structure rather than
13872 copying it. This is apparently the reason for the libsafe "errors"
13873 but the code is actually correct.
13877 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
13878 Bleichenbacher's DSA attack.
13879 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
13880 to be set and top=0 forces the highest bit to be set; top=-1 is new
13881 and leaves the highest bit random.
13883 *Ulf Moeller, Bodo Moeller*
13885 * In the NCONF_...-based implementations for CONF_... queries
13886 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
13887 a temporary CONF structure with the data component set to NULL
13888 (which gives segmentation faults in lh_retrieve).
13889 Instead, use NULL for the CONF pointer in CONF_get_string and
13890 CONF_get_number (which may use environment variables) and directly
13891 return NULL from CONF_get_section.
13895 * Fix potential buffer overrun for EBCDIC.
13899 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
13900 keyUsage if basicConstraints absent for a CA.
13904 * Make SMIME_write_PKCS7() write mail header values with a format that
13905 is more generally accepted (no spaces before the semicolon), since
13906 some programs can't parse those values properly otherwise. Also make
13907 sure BIO's that break lines after each write do not create invalid
13912 * Make the CRL encoding routines work with empty SEQUENCE OF. The
13913 macros previously used would not encode an empty SEQUENCE OF
13914 and break the signature.
13918 *This change does not apply to 0.9.7.*
13920 * Zero the premaster secret after deriving the master secret in
13925 * Add some EVP_add_digest_alias registrations (as found in
13926 OpenSSL_add_all_digests()) to SSL_library_init()
13927 aka OpenSSL_add_ssl_algorithms(). This provides improved
13928 compatibility with peers using X.509 certificates
13929 with unconventional AlgorithmIdentifier OIDs.
13933 * Fix for Irix with NO_ASM.
13935 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
13937 * ./config script fixes.
13939 *Ulf Moeller, Richard Levitte*
13941 * Fix 'openssl passwd -1'.
13945 * Change PKCS12_key_gen_asc() so it can cope with non null
13946 terminated strings whose length is passed in the passlen
13947 parameter, for example from PEM callbacks. This was done
13948 by adding an extra length parameter to asc2uni().
13950 *Steve Henson, reported by <oddissey@samsung.co.kr>*
13952 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
13953 call failed, free the DSA structure.
13957 * Fix to uni2asc() to cope with zero length Unicode strings.
13958 These are present in some PKCS#12 files.
13962 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
13963 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
13964 when writing a 32767 byte record.
13966 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
13968 * In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
13969 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
13971 (RSA objects have a reference count access to which is protected
13972 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
13973 so they are meant to be shared between threads.)
13974 *Bodo Moeller, Geoff Thorpe; original patch submitted by
13975 "Reddie, Steven" <Steven.Reddie@ca.com>*
13977 * Fix a deadlock in CRYPTO_mem_leaks().
13981 * Use better test patterns in bntest.
13985 * rand_win.c fix for Borland C.
13989 * BN_rshift bugfix for n == 0.
13993 * Add a 'bctest' script that checks for some known 'bc' bugs
13994 so that 'make test' does not abort just because 'bc' is broken.
13998 * Store verify_result within SSL_SESSION also for client side to
13999 avoid potential security hole. (Re-used sessions on the client side
14000 always resulted in verify_result==X509_V_OK, not using the original
14001 result of the server certificate verification.)
14005 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14006 SSL3_RT_APPLICATION_DATA, return 0.
14007 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14012 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14013 releases, have been re-implemented by renaming the previous
14014 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14015 and ssl3_read_internal, respectively, and adding 'peek' parameters
14016 to them. The new ssl[23]_{read,peek} functions are calls to
14017 ssl[23]_read_internal with the 'peek' flag set appropriately.
14018 A 'peek' parameter has also been added to ssl3_read_bytes, which
14019 does the actual work for ssl3_read_internal.
14023 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14024 the method-specific "init()" handler. Also clean up ex_data after
14025 calling the method-specific "finish()" handler. Previously, this was
14026 happening the other way round.
14030 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14031 The previous value, 12, was not always sufficient for BN_mod_exp().
14035 * Make sure that shared libraries get the internal name engine with
14036 the full version number and not just 0. This should mark the
14037 shared libraries as not backward compatible. Of course, this should
14038 be changed again when we can guarantee backward binary compatibility.
14042 * Fix typo in get_cert_by_subject() in by_dir.c
14044 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14046 * Rework the system to generate shared libraries:
14048 - Make note of the expected extension for the shared libraries and
14049 if there is a need for symbolic links from for example libcrypto.so.0
14050 to libcrypto.so.0.9.7. There is extended info in Configure for
14053 - Make as few rebuilds of the shared libraries as possible.
14055 - Still avoid linking the OpenSSL programs with the shared libraries.
14057 - When installing, install the shared libraries separately from the
14062 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14064 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14065 and not in SSL_clear because the latter is also used by the
14066 accept/connect functions; previously, the settings made by
14067 SSL_set_read_ahead would be lost during the handshake.
14069 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14071 * Correct util/mkdef.pl to be selective about disabled algorithms.
14072 Previously, it would create entries for disabled algorithms no
14077 * Added several new manual pages for SSL_* function.
14081 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000] ###
14083 * In ssl23_get_client_hello, generate an error message when faced
14084 with an initial SSL 3.0/TLS record that is too small to contain the
14085 first two bytes of the ClientHello message, i.e. client_version.
14086 (Note that this is a pathologic case that probably has never happened
14087 in real life.) The previous approach was to use the version number
14088 from the record header as a substitute; but our protocol choice
14089 should not depend on that one because it is not authenticated
14090 by the Finished messages.
14094 * More robust randomness gathering functions for Windows.
14096 *Jeffrey Altman <jaltman@columbia.edu>*
14098 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14099 not set then we don't setup the error code for issuer check errors
14100 to avoid possibly overwriting other errors which the callback does
14101 handle. If an application does set the flag then we assume it knows
14102 what it is doing and can handle the new informational codes
14107 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14108 a general "ANY" type, as such it should be able to decode anything
14109 including tagged types. However it didn't check the class so it would
14110 wrongly interpret tagged types in the same way as their universal
14111 counterpart and unknown types were just rejected. Changed so that the
14112 tagged and unknown types are handled in the same way as a SEQUENCE:
14113 that is the encoding is stored intact. There is also a new type
14114 "V_ASN1_OTHER" which is used when the class is not universal, in this
14115 case we have no idea what the actual type is so we just lump them all
14120 * On VMS, stdout may very well lead to a file that is written to
14121 in a record-oriented fashion. That means that every write() will
14122 write a separate record, which will be read separately by the
14123 programs trying to read from it. This can be very confusing.
14125 The solution is to put a BIO filter in the way that will buffer
14126 text until a linefeed is reached, and then write everything a
14127 line at a time, so every record written will be an actual line,
14128 not chunks of lines and not (usually doesn't happen, but I've
14129 seen it once) several lines in one record. BIO_f_linebuffer() is
14132 Currently, it's a VMS-only method, because that's where it has
14133 been tested well enough.
14137 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14138 it can return incorrect results.
14139 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14140 but it was in 0.9.6-beta[12].)
14144 * Disable the check for content being present when verifying detached
14145 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14146 include zero length content when signing messages.
14150 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14151 BIO_ctrl (for BIO pairs).
14155 * Add DSO method for VMS.
14159 * Bug fix: Montgomery multiplication could produce results with the
14164 * Add RPM specification openssl.spec and modify it to build three
14165 packages. The default package contains applications, application
14166 documentation and run-time libraries. The devel package contains
14167 include files, static libraries and function documentation. The
14168 doc package contains the contents of the doc directory. The original
14169 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14173 * Add a large number of documentation files for many SSL routines.
14175 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14177 * Add a configuration entry for Sony News 4.
14179 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14181 * Don't set the two most significant bits to one when generating a
14182 random number < q in the DSA library.
14186 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14187 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14188 the underlying transport is blocking) if a handshake took place.
14189 (The default behaviour is needed by applications such as s_client
14190 and s_server that use select() to determine when to use SSL_read;
14191 but for applications that know in advance when to expect data, it
14192 just makes things more complicated.)
14196 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14201 * Add a few more EBCDIC conditionals that make `req' and `x509'
14202 work better on such systems.
14204 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14206 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14207 Update PKCS12_parse() so it copies the friendlyName and the
14208 keyid to the certificates aux info.
14212 * Fix bug in PKCS7_verify() which caused an infinite loop
14213 if there was more than one signature.
14215 *Sven Uszpelkat <su@celocom.de>*
14217 * Major change in util/mkdef.pl to include extra information
14218 about each symbol, as well as presenting variables as well
14219 as functions. This change means that there's n more need
14220 to rebuild the .num files when some algorithms are excluded.
14224 * Allow the verify time to be set by an application,
14225 rather than always using the current time.
14229 * Phase 2 verify code reorganisation. The certificate
14230 verify code now looks up an issuer certificate by a
14231 number of criteria: subject name, authority key id
14232 and key usage. It also verifies self signed certificates
14233 by the same criteria. The main comparison function is
14234 X509_check_issued() which performs these checks.
14236 Lot of changes were necessary in order to support this
14237 without completely rewriting the lookup code.
14239 Authority and subject key identifier are now cached.
14241 The LHASH 'certs' is X509_STORE has now been replaced
14242 by a STACK_OF(X509_OBJECT). This is mainly because an
14243 LHASH can't store or retrieve multiple objects with
14244 the same hash value.
14246 As a result various functions (which were all internal
14247 use only) have changed to handle the new X509_STORE
14248 structure. This will break anything that messed round
14249 with X509_STORE internally.
14251 The functions X509_STORE_add_cert() now checks for an
14252 exact match, rather than just subject name.
14254 The X509_STORE API doesn't directly support the retrieval
14255 of multiple certificates matching a given criteria, however
14256 this can be worked round by performing a lookup first
14257 (which will fill the cache with candidate certificates)
14258 and then examining the cache for matches. This is probably
14259 the best we can do without throwing out X509_LOOKUP
14260 entirely (maybe later...).
14262 The X509_VERIFY_CTX structure has been enhanced considerably.
14264 All certificate lookup operations now go via a get_issuer()
14265 callback. Although this currently uses an X509_STORE it
14266 can be replaced by custom lookups. This is a simple way
14267 to bypass the X509_STORE hackery necessary to make this
14268 work and makes it possible to use more efficient techniques
14269 in future. A very simple version which uses a simple
14270 STACK for its trusted certificate store is also provided
14271 using X509_STORE_CTX_trusted_stack().
14273 The verify_cb() and verify() callbacks now have equivalents
14274 in the X509_STORE_CTX structure.
14276 X509_STORE_CTX also has a 'flags' field which can be used
14277 to customise the verify behaviour.
14281 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14282 excludes S/MIME capabilities.
14286 * When a certificate request is read in keep a copy of the
14287 original encoding of the signed data and use it when outputting
14288 again. Signatures then use the original encoding rather than
14289 a decoded, encoded version which may cause problems if the
14290 request is improperly encoded.
14294 * For consistency with other BIO_puts implementations, call
14295 buffer_write(b, ...) directly in buffer_puts instead of calling
14298 In BIO_puts, increment b->num_write as in BIO_write.
14300 *Peter.Sylvester@EdelWeb.fr*
14302 * Fix BN_mul_word for the case where the word is 0. (We have to use
14303 BN_zero, we may not return a BIGNUM with an array consisting of
14304 words set to zero.)
14308 * Avoid calling abort() from within the library when problems are
14309 detected, except if preprocessor symbols have been defined
14310 (such as REF_CHECK, BN_DEBUG etc.).
14314 * New openssl application 'rsautl'. This utility can be
14315 used for low level RSA operations. DER public key
14316 BIO/fp routines also added.
14320 * New Configure entry and patches for compiling on QNX 4.
14322 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14324 * A demo state-machine implementation was sponsored by
14325 Nuron (http://www.nuron.com/) and is now available in
14326 demos/state_machine.
14330 * New options added to the 'dgst' utility for signature
14331 generation and verification.
14335 * Unrecognized PKCS#7 content types are now handled via a
14336 catch all ASN1_TYPE structure. This allows unsupported
14337 types to be stored as a "blob" and an application can
14338 encode and decode it manually.
14342 * Fix various signed/unsigned issues to make a_strex.c
14343 compile under VC++.
14345 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14347 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14348 length if passed a buffer. ASN1_INTEGER_to_BN failed
14349 if passed a NULL BN and its argument was negative.
14351 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14353 * Modification to PKCS#7 encoding routines to output definite
14354 length encoding. Since currently the whole structures are in
14355 memory there's not real point in using indefinite length
14356 constructed encoding. However if OpenSSL is compiled with
14357 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14361 * Added BIO_vprintf() and BIO_vsnprintf().
14365 * Added more prefixes to parse for in the strings written
14366 through a logging bio, to cover all the levels that are available
14367 through syslog. The prefixes are now:
14369 PANIC, EMERG, EMR => LOG_EMERG
14370 ALERT, ALR => LOG_ALERT
14371 CRIT, CRI => LOG_CRIT
14372 ERROR, ERR => LOG_ERR
14373 WARNING, WARN, WAR => LOG_WARNING
14374 NOTICE, NOTE, NOT => LOG_NOTICE
14375 INFO, INF => LOG_INFO
14376 DEBUG, DBG => LOG_DEBUG
14378 and as before, if none of those prefixes are present at the
14379 beginning of the string, LOG_ERR is chosen.
14381 On Win32, the LOG_* levels are mapped according to this:
14383 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14384 LOG_WARNING => EVENTLOG_WARNING_TYPE
14385 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14390 * Made it possible to reconfigure with just the configuration
14391 argument "reconf" or "reconfigure". The command line arguments
14392 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14393 and are retrieved from there when reconfiguring.
14399 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14401 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14405 * The obj_dat.pl script was messing up the sorting of object
14406 names. The reason was that it compared the quoted version
14407 of strings as a result "OCSP" > "OCSP Signing" because
14408 " > SPACE. Changed script to store unquoted versions of
14409 names and add quotes on output. It was also omitting some
14410 names from the lookup table if they were given a default
14411 value (that is if SN is missing it is given the same
14412 value as LN and vice versa), these are now added on the
14413 grounds that if an object has a name we should be able to
14414 look it up. Finally added warning output when duplicate
14415 short or long names are found.
14419 * Changes needed for Tandem NSK.
14421 *Scott Uroff <scott@xypro.com>*
14423 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14424 RSA_padding_check_SSLv23(), special padding was never detected
14425 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14426 version rollback attacks was not effective.
14428 In s23_clnt.c, don't use special rollback-attack detection padding
14429 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14430 client; similarly, in s23_srvr.c, don't do the rollback check if
14431 SSL 2.0 is the only protocol enabled in the server.
14435 * Make it possible to get hexdumps of unprintable data with 'openssl
14436 asn1parse'. By implication, the functions ASN1_parse_dump() and
14437 BIO_dump_indent() are added.
14441 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14442 these print out strings and name structures based on various
14443 flags including RFC2253 support and proper handling of
14444 multibyte characters. Added options to the 'x509' utility
14445 to allow the various flags to be set.
14449 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14450 Also change the functions X509_cmp_current_time() and
14451 X509_gmtime_adj() work with an ASN1_TIME structure,
14452 this will enable certificates using GeneralizedTime in validity
14453 dates to be checked.
14457 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14458 negative public key encodings) on by default,
14459 NO_NEG_PUBKEY_BUG can be set to disable it.
14463 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14464 content octets. An i2c_ASN1_OBJECT is unnecessary because
14465 the encoding can be trivially obtained from the structure.
14469 * crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
14470 not read locks (CRYPTO_r_[un]lock).
14474 * A first attempt at creating official support for shared
14475 libraries through configuration. I've kept it so the
14476 default is static libraries only, and the OpenSSL programs
14477 are always statically linked for now, but there are
14478 preparations for dynamic linking in place.
14479 This has been tested on Linux and Tru64.
14483 * Randomness polling function for Win9x, as described in:
14484 Peter Gutmann, Software Generation of Practically Strong
14489 * Fix so PRNG is seeded in req if using an already existing
14494 * New options to smime application. -inform and -outform
14495 allow alternative formats for the S/MIME message including
14496 PEM and DER. The -content option allows the content to be
14497 specified separately. This should allow things like Netscape
14498 form signing output easier to verify.
14502 * Fix the ASN1 encoding of tags using the 'long form'.
14506 * New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
14507 STRING types. These convert content octets to and from the
14508 underlying type. The actual tag and length octets are
14509 already assumed to have been read in and checked. These
14510 are needed because all other string types have virtually
14511 identical handling apart from the tag. By having versions
14512 of the ASN1 functions that just operate on content octets
14513 IMPLICIT tagging can be handled properly. It also allows
14514 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14515 and ASN1_INTEGER are identical apart from the tag.
14519 * Change the handling of OID objects as follows:
14521 - New object identifiers are inserted in objects.txt, following
14522 the syntax given in objects.README.
14523 - objects.pl is used to process obj_mac.num and create a new
14525 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14528 This is currently kind of a hack, and the perl code in objects.pl
14529 isn't very elegant, but it works as I intended. The simplest way
14530 to check that it worked correctly is to look in obj_dat.h and
14531 check the array nid_objs and make sure the objects haven't moved
14532 around (this is important!). Additions are OK, as well as
14533 consistent name changes.
14537 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14541 * Addition of the command line parameter '-rand file' to 'openssl req'.
14542 The given file adds to whatever has already been seeded into the
14543 random pool through the RANDFILE configuration file option or
14544 environment variable, or the default random state file.
14548 * mkstack.pl now sorts each macro group into lexical order.
14549 Previously the output order depended on the order the files
14550 appeared in the directory, resulting in needless rewriting
14555 * Patches to make OpenSSL compile under Win32 again. Mostly
14556 work arounds for the VC++ problem that it treats func() as
14557 func(void). Also stripped out the parts of mkdef.pl that
14558 added extra typesafe functions: these no longer exist.
14562 * Reorganisation of the stack code. The macros are now all
14563 collected in safestack.h . Each macro is defined in terms of
14564 a "stack macro" of the form SKM_<name>(type, a, b). The
14565 DEBUG_SAFESTACK is now handled in terms of function casts,
14566 this has the advantage of retaining type safety without the
14567 use of additional functions. If DEBUG_SAFESTACK is not defined
14568 then the non typesafe macros are used instead. Also modified the
14569 mkstack.pl script to handle the new form. Needs testing to see
14570 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14571 the default if no major problems. Similar behaviour for ASN1_SET_OF
14572 and PKCS12_STACK_OF.
14576 * When some versions of IIS use the 'NET' form of private key the
14577 key derivation algorithm is different. Normally MD5(password) is
14578 used as a 128 bit RC4 key. In the modified case
14579 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14580 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14581 as the old Netscape_RSA functions except they have an additional
14582 'sgckey' parameter which uses the modified algorithm. Also added
14583 an -sgckey command line option to the rsa utility. Thanks to
14584 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14585 algorithm to openssl-dev.
14589 * The evp_local.h macros were using 'c.##kname' which resulted in
14590 invalid expansion on some systems (SCO 5.0.5 for example).
14591 Corrected to 'c.kname'.
14593 *Phillip Porch <root@theporch.com>*
14595 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14596 a STACK of email addresses from a certificate or request, these look
14597 in the subject name and the subject alternative name extensions and
14598 omit any duplicate addresses.
14602 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14603 This makes DSA verification about 2 % faster.
14607 * Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
14608 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14609 plus overhead for 1024 bit moduli).
14610 This makes exponentiations about 0.5 % faster for 1024 bit
14611 exponents (as measured by "openssl speed rsa2048").
14615 * Rename memory handling macros to avoid conflicts with other
14617 Malloc => OPENSSL_malloc
14618 Malloc_locked => OPENSSL_malloc_locked
14619 Realloc => OPENSSL_realloc
14620 Free => OPENSSL_free
14624 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14625 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14629 * CygWin32 support.
14631 *John Jarvie <jjarvie@newsguy.com>*
14633 * The type-safe stack code has been rejigged. It is now only compiled
14634 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14635 by default all type-specific stack functions are "#define"d back to
14636 standard stack functions. This results in more streamlined output
14637 but retains the type-safety checking possibilities of the original
14642 * The STACK code has been cleaned up, and certain type declarations
14643 that didn't make a lot of sense have been brought in line. This has
14644 also involved a cleanup of sorts in safestack.h to more correctly
14645 map type-safe stack functions onto their plain stack counterparts.
14646 This work has also resulted in a variety of "const"ifications of
14647 lots of the code, especially "_cmp" operations which should normally
14648 be prototyped with "const" parameters anyway.
14652 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14653 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14654 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14655 where all of 'md' is used each time the PRNG is used, but 'state'
14656 is used only indexed by a cyclic counter. As entropy may not be
14657 well distributed from the beginning, 'md' is important as a
14658 chaining variable. However, the output function chains only half
14659 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14660 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14661 in all of 'state' being rewritten, with the new values depending
14662 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14666 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14667 the handshake is continued after ssl_verify_cert_chain();
14668 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14669 can lead to 'unexplainable' connection aborts later.
14671 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14673 * Major EVP API cipher revision.
14674 Add hooks for extra EVP features. This allows various cipher
14675 parameters to be set in the EVP interface. Support added for variable
14676 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14677 setting of RC2 and RC5 parameters.
14679 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14682 Remove lots of duplicated code from the EVP library. For example *every*
14683 cipher init() function handles the 'iv' in the same way according to the
14684 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14685 for CFB and OFB modes they zero ctx->num.
14687 New functionality allows removal of S/MIME code RC2 hack.
14689 Most of the routines have the same form and so can be declared in terms
14692 By shifting this to the top level EVP_CipherInit() it can be removed from
14693 all individual ciphers. If the cipher wants to handle IVs or keys
14694 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14697 Change lots of functions like EVP_EncryptUpdate() to now return a
14698 value: although software versions of the algorithms cannot fail
14699 any installed hardware versions can.
14703 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14704 this option is set, tolerate broken clients that send the negotiated
14705 protocol version number instead of the requested protocol version
14710 * Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
14711 i.e. non-zero for export ciphersuites, zero otherwise.
14712 Previous versions had this flag inverted, inconsistent with
14713 rsa_tmp_cb (..._TMP_RSA_CB).
14715 *Bodo Moeller; problem reported by Amit Chopra*
14717 * Add missing DSA library text string. Work around for some IIS
14718 key files with invalid SEQUENCE encoding.
14722 * Add a document (doc/standards.txt) that list all kinds of standards
14723 and so on that are implemented in OpenSSL.
14727 * Enhance c_rehash script. Old version would mishandle certificates
14728 with the same subject name hash and wouldn't handle CRLs at all.
14729 Added -fingerprint option to crl utility, to support new c_rehash
14734 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14738 * Fix for SSL server purpose checking. Server checking was
14739 rejecting certificates which had extended key usage present
14740 but no ssl client purpose.
14742 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14744 * Make PKCS#12 code work with no password. The PKCS#12 spec
14745 is a little unclear about how a blank password is handled.
14746 Since the password in encoded as a BMPString with terminating
14747 double NULL a zero length password would end up as just the
14748 double NULL. However no password at all is different and is
14749 handled differently in the PKCS#12 key generation code. NS
14750 treats a blank password as zero length. MSIE treats it as no
14751 password on export: but it will try both on import. We now do
14752 the same: PKCS12_parse() tries zero length and no password if
14753 the password is set to "" or NULL (NULL is now a valid password:
14754 it wasn't before) as does the pkcs12 application.
14758 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14759 perror when PEM_read_bio_X509_REQ fails, the error message must
14760 be obtained from the error queue.
14764 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14765 it in ERR_remove_state if appropriate, and change ERR_get_state
14766 accordingly to avoid race conditions (this is necessary because
14767 thread_hash is no longer constant once set).
14771 * Bugfix for linux-elf makefile.one.
14775 * RSA_get_default_method() will now cause a default
14776 RSA_METHOD to be chosen if one doesn't exist already.
14777 Previously this was only set during a call to RSA_new()
14778 or RSA_new_method(NULL) meaning it was possible for
14779 RSA_get_default_method() to return NULL.
14783 * Added native name translation to the existing DSO code
14784 that will convert (if the flag to do so is set) filenames
14785 that are sufficiently small and have no path information
14786 into a canonical native form. Eg. "blah" converted to
14787 "libblah.so" or "blah.dll" etc.
14791 * New function ERR_error_string_n(e, buf, len) which is like
14792 ERR_error_string(e, buf), but writes at most 'len' bytes
14793 including the 0 terminator. For ERR_error_string_n, 'buf'
14796 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14798 * CONF library reworked to become more general. A new CONF
14799 configuration file reader "class" is implemented as well as a
14800 new functions (NCONF_*, for "New CONF") to handle it. The now
14801 old CONF_* functions are still there, but are reimplemented to
14802 work in terms of the new functions. Also, a set of functions
14803 to handle the internal storage of the configuration data is
14804 provided to make it easier to write new configuration file
14805 reader "classes" (I can definitely see something reading a
14806 configuration file in XML format, for example), called _CONF_*,
14807 or "the configuration storage API"...
14809 The new configuration file reading functions are:
14811 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14812 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14814 NCONF_default, NCONF_WIN32
14816 NCONF_dump_fp, NCONF_dump_bio
14818 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14819 NCONF_new creates a new CONF object. This works in the same way
14820 as other interfaces in OpenSSL, like the BIO interface.
14821 NCONF_dump_* dump the internal storage of the configuration file,
14822 which is useful for debugging. All other functions take the same
14823 arguments as the old CONF_* functions with the exception of the
14824 first that must be a `CONF *' instead of a `LHASH *'.
14826 To make it easier to use the new classes with the old CONF_* functions,
14827 the function CONF_set_default_method is provided.
14831 * Add '-tls1' option to 'openssl ciphers', which was already
14832 mentioned in the documentation but had not been implemented.
14833 (This option is not yet really useful because even the additional
14834 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
14838 * Initial DSO code added into libcrypto for letting OpenSSL (and
14839 OpenSSL-based applications) load shared libraries and bind to
14840 them in a portable way.
14842 *Geoff Thorpe, with contributions from Richard Levitte*
14844 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000] ###
14846 * Make sure _lrotl and _lrotr are only used with MSVC.
14848 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
14849 (the default implementation of RAND_status).
14851 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
14852 to '-clrext' (= clear extensions), as intended and documented.
14853 *Bodo Moeller; inconsistency pointed out by Michael Attili
14854 <attili@amaxo.com>*
14856 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
14857 was larger than the MD block size.
14859 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
14861 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
14862 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
14863 using the passed key: if the passed key was a private key the result
14864 of X509_print(), for example, would be to print out all the private key
14869 * des_quad_cksum() byte order bug fix.
14870 *Ulf Möller, using the problem description in krb4-0.9.7, where
14871 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
14873 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
14876 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
14878 * For easily testing in shell scripts whether some command
14879 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
14880 returns with exit code 0 iff no command of the given name is available.
14881 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
14882 the output goes to stdout and nothing is printed to stderr.
14883 Additional arguments are always ignored.
14885 Since for each cipher there is a command of the same name,
14886 the 'no-cipher' compilation switches can be tested this way.
14888 ('openssl no-XXX' is not able to detect pseudo-commands such
14889 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
14893 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
14897 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
14898 is set; it will be thrown away anyway because each handshake creates
14900 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
14901 to parameters -- in previous versions (since OpenSSL 0.9.3) the
14902 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
14903 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
14907 * New s_client option -ign_eof: EOF at stdin is ignored, and
14908 'Q' and 'R' lose their special meanings (quit/renegotiate).
14909 This is part of what -quiet does; unlike -quiet, -ign_eof
14910 does not suppress any output.
14914 * Add compatibility options to the purpose and trust code. The
14915 purpose X509_PURPOSE_ANY is "any purpose" which automatically
14916 accepts a certificate or CA, this was the previous behaviour,
14917 with all the associated security issues.
14919 X509_TRUST_COMPAT is the old trust behaviour: only and
14920 automatically trust self signed roots in certificate store. A
14921 new trust setting X509_TRUST_DEFAULT is used to specify that
14922 a purpose has no associated trust setting and it should instead
14923 use the value in the default purpose.
14927 * Fix the PKCS#8 DSA private key code so it decodes keys again
14928 and fix a memory leak.
14932 * In util/mkerr.pl (which implements 'make errors'), preserve
14933 reason strings from the previous version of the .c file, as
14934 the default to have only downcase letters (and digits) in
14935 automatically generated reasons codes is not always appropriate.
14939 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
14940 using strerror. Previously, ERR_reason_error_string() returned
14941 library names as reason strings for SYSerr; but SYSerr is a special
14942 case where small numbers are errno values, not library numbers.
14946 * Add '-dsaparam' option to 'openssl dhparam' application. This
14947 converts DSA parameters into DH parameters. (When creating parameters,
14948 DSA_generate_parameters is used.)
14952 * Include 'length' (recommended exponent length) in C code generated
14953 by 'openssl dhparam -C'.
14957 * The second argument to set_label in perlasm was already being used
14958 so couldn't be used as a "file scope" flag. Moved to third argument
14963 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
14964 instead of RAND_bytes for encryption IVs and salts.
14968 * Include RAND_status() into RAND_METHOD instead of implementing
14969 it only for md_rand.c Otherwise replacing the PRNG by calling
14970 RAND_set_rand_method would be impossible.
14974 * Don't let DSA_generate_key() enter an infinite loop if the random
14975 number generation fails.
14979 * New 'rand' application for creating pseudo-random output.
14983 * Added configuration support for Linux/IA64
14985 *Rolf Haberrecker <rolf@suse.de>*
14987 * Assembler module support for Mingw32.
14991 * Shared library support for HPUX (in shlib/).
14993 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
14995 * Shared library support for Solaris gcc.
14997 *Lutz Behnke <behnke@trustcenter.de>*
14999 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000] ###
15001 * PKCS7_encrypt() was adding text MIME headers twice because they
15002 were added manually and by SMIME_crlf_copy().
15006 * In bntest.c don't call BN_rand with zero bits argument.
15008 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15010 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15011 case was implemented. This caused BN_div_recp() to fail occasionally.
15015 * Add an optional second argument to the set_label() in the perl
15016 assembly language builder. If this argument exists and is set
15017 to 1 it signals that the assembler should use a symbol whose
15018 scope is the entire file, not just the current function. This
15019 is needed with MASM which uses the format label:: for this scope.
15021 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15023 * Change the ASN1 types so they are typedefs by default. Before
15024 almost all types were #define'd to ASN1_STRING which was causing
15025 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15030 * Change names of new functions to the new get1/get0 naming
15031 convention: After 'get1', the caller owns a reference count
15032 and has to call ..._free; 'get0' returns a pointer to some
15033 data structure without incrementing reference counters.
15034 (Some of the existing 'get' functions increment a reference
15035 counter, some don't.)
15036 Similarly, 'set1' and 'add1' functions increase reference
15037 counters or duplicate objects.
15041 * Allow for the possibility of temp RSA key generation failure:
15042 the code used to assume it always worked and crashed on failure.
15046 * Fix potential buffer overrun problem in BIO_printf().
15047 *Ulf Möller, using public domain code by Patrick Powell; problem
15048 pointed out by David Sacerdote <das33@cornell.edu>*
15050 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15051 RAND_egd() and RAND_status(). In the command line application,
15052 the EGD socket can be specified like a seed file using RANDFILE
15057 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15058 Some CAs (e.g. Verisign) distribute certificates in this form.
15062 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15063 list to exclude them. This means that no special compilation option
15064 is needed to use anonymous DH: it just needs to be included in the
15069 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15070 EVP_MD_type. The old functionality is available in a new macro called
15071 EVP_MD_md(). Change code that uses it and update docs.
15075 * ..._ctrl functions now have corresponding ..._callback_ctrl functions
15076 where the 'void *' argument is replaced by a function pointer argument.
15077 Previously 'void *' was abused to point to functions, which works on
15078 many platforms, but is not correct. As these functions are usually
15079 called by macros defined in OpenSSL header files, most source code
15080 should work without changes.
15084 * <openssl/opensslconf.h> (which is created by Configure) now contains
15085 sections with information on -D... compiler switches used for
15086 compiling the library so that applications can see them. To enable
15087 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
15088 must be defined. E.g.,
15089 #define OPENSSL_ALGORITHM_DEFINES
15090 #include <openssl/opensslconf.h>
15091 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
15093 *Richard Levitte, Ulf and Bodo Möller*
15095 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15100 * Change the 'other' type in certificate aux info to a STACK_OF
15101 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15102 the required ASN1 format: arbitrary types determined by an OID.
15106 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15107 argument to 'req'. This is not because the function is newer or
15108 better than others it just uses the work 'NEW' in the certificate
15109 request header lines. Some software needs this.
15113 * Reorganise password command line arguments: now passwords can be
15114 obtained from various sources. Delete the PEM_cb function and make
15115 it the default behaviour: i.e. if the callback is NULL and the
15116 usrdata argument is not NULL interpret it as a null terminated pass
15117 phrase. If usrdata and the callback are NULL then the pass phrase
15118 is prompted for as usual.
15122 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15123 the support is automatically enabled. The resulting binaries will
15124 autodetect the card and use it if present.
15126 *Ben Laurie and Compaq Inc.*
15128 * Work around for Netscape hang bug. This sends certificate request
15129 and server done in one record. Since this is perfectly legal in the
15130 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15131 the bugs/SSLv3 entry for more info.
15135 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15139 * Add -rand argument to smime and pkcs12 applications and read/write
15144 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15148 * Add command line password options to the remaining applications.
15152 * Bug fix for BN_div_recp() for numerators with an even number of
15157 * More tests in bntest.c, and changed test_bn output.
15161 * ./config recognizes MacOS X now.
15165 * Bug fix for BN_div() when the first words of num and divisor are
15166 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
15170 * Add support for various broken PKCS#8 formats, and command line
15171 options to produce them.
15175 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15176 get temporary BIGNUMs from a BN_CTX.
15180 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15185 * Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
15186 include a #define from the old name to the new. The original intent
15187 was that statically linked binaries could for example just call
15188 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15189 link with digests. This never worked because SSLeay_add_all_digests()
15190 and SSLeay_add_all_ciphers() were in the same source file so calling
15191 one would link with the other. They are now in separate source files.
15195 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15199 * Use a less unusual form of the Miller-Rabin primality test (it used
15200 a binary algorithm for exponentiation integrated into the Miller-Rabin
15201 loop, our standard modexp algorithms are faster).
15205 * Support for the EBCDIC character set completed.
15207 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15209 * Source code cleanups: use const where appropriate, eliminate casts,
15210 use void * instead of char * in lhash.
15214 * Bugfix: ssl3_send_server_key_exchange was not restartable
15215 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15216 this the server could overwrite ephemeral keys that the client
15221 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15222 using 50 iterations of the Rabin-Miller test.
15224 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15225 iterations of the Rabin-Miller test as required by the appendix
15226 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15227 As BN_is_prime_fasttest includes trial division, DSA parameter
15228 generation becomes much faster.
15230 This implies a change for the callback functions in DSA_is_prime
15231 and DSA_generate_parameters: The callback function is called once
15232 for each positive witness in the Rabin-Miller test, not just
15233 occasionally in the inner loop; and the parameters to the
15234 callback function now provide an iteration count for the outer
15235 loop rather than for the current invocation of the inner loop.
15236 DSA_generate_parameters additionally can call the callback
15237 function with an 'iteration count' of -1, meaning that a
15238 candidate has passed the trial division test (when q is generated
15239 from an application-provided seed, trial division is skipped).
15243 * New function BN_is_prime_fasttest that optionally does trial
15244 division before starting the Rabin-Miller test and has
15245 an additional BN_CTX * argument (whereas BN_is_prime always
15246 has to allocate at least one BN_CTX).
15247 'callback(1, -1, cb_arg)' is called when a number has passed the
15248 trial division stage.
15252 * Fix for bug in CRL encoding. The validity dates weren't being handled
15257 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15261 * New function BN_pseudo_rand().
15265 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15266 bignum version of BN_from_montgomery() with the working code from
15267 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15272 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15273 made it impossible to use the same SSL_SESSION data structure in
15274 SSL2 clients in multiple threads.
15278 * The return value of RAND_load_file() no longer counts bytes obtained
15279 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15280 to seed the PRNG (previously an explicit byte count was required).
15282 *Ulf Möller, Bodo Möller*
15284 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15285 used (char *) instead of (void *) and had casts all over the place.
15289 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15293 * Retain source code compatibility for BN_prime_checks macro:
15294 BN_is_prime(..., BN_prime_checks, ...) now uses
15295 BN_prime_checks_for_size to determine the appropriate number of
15296 Rabin-Miller iterations.
15300 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15301 DH_CHECK_P_NOT_SAFE_PRIME.
15302 (Check if this is true? OpenPGP calls them "strong".)
15306 * Merge the functionality of "dh" and "gendh" programs into a new program
15307 "dhparam". The old programs are retained for now but will handle DH keys
15308 (instead of parameters) in future.
15312 * Make the ciphers, s_server and s_client programs check the return values
15313 when a new cipher list is set.
15317 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15318 ciphers. Before when the 56bit ciphers were enabled the sorting was
15321 The syntax for the cipher sorting has been extended to support sorting by
15322 cipher-strength (using the strength_bits hard coded in the tables).
15323 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15325 Fix a bug in the cipher-command parser: when supplying a cipher command
15326 string with an "undefined" symbol (neither command nor alphanumeric
15327 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15328 an error is flagged.
15330 Due to the strength-sorting extension, the code of the
15331 ssl_create_cipher_list() function was completely rearranged. I hope that
15332 the readability was also increased :-)
15334 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15336 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15337 for the first serial number and places 2 in the serial number file. This
15338 avoids problems when the root CA is created with serial number zero and
15339 the first user certificate has the same issuer name and serial number
15344 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15345 the new code. Add documentation for this stuff.
15349 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15350 X509_*() to X509at_*() on the grounds that they don't handle X509
15351 structures and behave in an analogous way to the X509v3 functions:
15352 they shouldn't be called directly but wrapper functions should be used
15355 So we also now have some wrapper functions that call the X509at functions
15356 when passed certificate requests. (TO DO: similar things can be done with
15357 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15358 things. Some of these need some d2i or i2d and print functionality
15359 because they handle more complex structures.)
15363 * Add missing #ifndefs that caused missing symbols when building libssl
15364 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15365 NO_RSA in ssl/s2*.c.
15367 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15369 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15370 has a return value which indicates the quality of the random data
15371 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15372 error queue. New function RAND_pseudo_bytes() generates output that is
15373 guaranteed to be unique but not unpredictable. RAND_add is like
15374 RAND_seed, but takes an extra argument for an entropy estimate
15375 (RAND_seed always assumes full entropy).
15379 * Do more iterations of Rabin-Miller probable prime test (specifically,
15380 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15381 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15382 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15383 false-positive rate of at most 2^-80 for random input.
15387 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15391 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15392 in the 0.9.5 release), this returns the chain
15393 from an X509_CTX structure with a dup of the stack and all
15394 the X509 reference counts upped: so the stack will exist
15395 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15398 Also make SSL_SESSION_print() print out the verify return
15403 * Add manpage for the pkcs12 command. Also change the default
15404 behaviour so MAC iteration counts are used unless the new
15405 -nomaciter option is used. This improves file security and
15406 only older versions of MSIE (4.0 for example) need it.
15410 * Honor the no-xxx Configure options when creating .DEF files.
15414 * Add PKCS#10 attributes to field table: challengePassword,
15415 unstructuredName and unstructuredAddress. These are taken from
15416 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15417 international characters are used.
15419 More changes to X509_ATTRIBUTE code: allow the setting of types
15420 based on strings. Remove the 'loc' parameter when adding
15421 attributes because these will be a SET OF encoding which is sorted
15426 * Initial changes to the 'req' utility to allow request generation
15427 automation. This will allow an application to just generate a template
15428 file containing all the field values and have req construct the
15431 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15432 used all over the place including certificate requests and PKCS#7
15433 structures. They are currently handled manually where necessary with
15434 some primitive wrappers for PKCS#7. The new functions behave in a
15435 manner analogous to the X509 extension functions: they allow
15436 attributes to be looked up by NID and added.
15438 Later something similar to the X509V3 code would be desirable to
15439 automatically handle the encoding, decoding and printing of the
15440 more complex types. The string types like challengePassword can
15441 be handled by the string table functions.
15443 Also modified the multi byte string table handling. Now there is
15444 a 'global mask' which masks out certain types. The table itself
15445 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15446 is useful when for example there is only one permissible type
15447 (as in countryName) and using the mask might result in no valid
15452 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15453 SSL_get_peer_finished to allow applications to obtain the latest
15454 Finished messages sent to the peer or expected from the peer,
15455 respectively. (SSL_get_peer_finished is usually the Finished message
15456 actually received from the peer, otherwise the protocol will be aborted.)
15458 As the Finished message are message digests of the complete handshake
15459 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15460 be used for external authentication procedures when the authentication
15461 provided by SSL/TLS is not desired or is not enough.
15465 * Enhanced support for Alpha Linux is added. Now ./config checks if
15466 the host supports BWX extension and if Compaq C is present on the
15467 $PATH. Just exploiting of the BWX extension results in 20-30%
15468 performance kick for some algorithms, e.g. DES and RC4 to mention
15469 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15474 * Add support for MS "fast SGC". This is arguably a violation of the
15475 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15476 weak crypto and after checking the certificate is SGC a second one
15477 with strong crypto. MS SGC stops the first handshake after receiving
15478 the server certificate message and sends a second client hello. Since
15479 a server will typically do all the time consuming operations before
15480 expecting any further messages from the client (server key exchange
15481 is the most expensive) there is little difference between the two.
15483 To get OpenSSL to support MS SGC we have to permit a second client
15484 hello message after we have sent server done. In addition we have to
15485 reset the MAC if we do get this second client hello.
15489 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15490 if a DER encoded private key is RSA or DSA traditional format. Changed
15491 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15492 format DER encoded private key. Newer code should use PKCS#8 format which
15493 has the key type encoded in the ASN1 structure. Added DER private key
15494 support to pkcs8 application.
15498 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15499 ciphersuites has been selected (as required by the SSL 3/TLS 1
15500 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15501 is set, we interpret this as a request to violate the specification
15502 (the worst that can happen is a handshake failure, and 'correct'
15503 behaviour would result in a handshake failure anyway).
15507 * In SSL_CTX_add_session, take into account that there might be multiple
15508 SSL_SESSION structures with the same session ID (e.g. when two threads
15509 concurrently obtain them from an external cache).
15510 The internal cache can handle only one SSL_SESSION with a given ID,
15511 so if there's a conflict, we now throw out the old one to achieve
15516 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15517 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15518 some routines that use cipher OIDs: some ciphers do not have OIDs
15519 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15524 * Simplify the trust setting structure and code. Now we just have
15525 two sequences of OIDs for trusted and rejected settings. These will
15526 typically have values the same as the extended key usage extension
15527 and any application specific purposes.
15529 The trust checking code now has a default behaviour: it will just
15530 check for an object with the same NID as the passed id. Functions can
15531 be provided to override either the default behaviour or the behaviour
15532 for a given id. SSL client, server and email already have functions
15533 in place for compatibility: they check the NID and also return "trusted"
15534 if the certificate is self signed.
15538 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15539 traditional format into an EVP_PKEY structure.
15543 * Add a password callback function PEM_cb() which either prompts for
15544 a password if usr_data is NULL or otherwise assumes it is a null
15545 terminated password. Allow passwords to be passed on command line
15546 environment or config files in a few more utilities.
15550 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15551 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15552 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15553 Update documentation.
15557 * Support for ASN1 "NULL" type. This could be handled before by using
15558 ASN1_TYPE but there wasn't any function that would try to read a NULL
15559 and produce an error if it couldn't. For compatibility we also have
15560 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15561 don't allocate anything because they don't need to.
15565 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15568 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15570 * Rebuild of the memory allocation routines used by OpenSSL code and
15571 possibly others as well. The purpose is to make an interface that
15572 provide hooks so anyone can build a separate set of allocation and
15573 deallocation routines to be used by OpenSSL, for example memory
15574 pool implementations, or something else, which was previously hard
15575 since Malloc(), Realloc() and Free() were defined as macros having
15576 the values malloc, realloc and free, respectively (except for Win32
15577 compilations). The same is provided for memory debugging code.
15578 OpenSSL already comes with functionality to find memory leaks, but
15579 this gives people a chance to debug other memory problems.
15581 With these changes, a new set of functions and macros have appeared:
15583 CRYPTO_set_mem_debug_functions() [F]
15584 CRYPTO_get_mem_debug_functions() [F]
15585 CRYPTO_dbg_set_options() [F]
15586 CRYPTO_dbg_get_options() [F]
15587 CRYPTO_malloc_debug_init() [M]
15589 The memory debug functions are NULL by default, unless the library
15590 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15591 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15592 gives the standard debugging functions that come with OpenSSL) or
15593 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15594 provided by the library user) must be used. When the standard
15595 debugging functions are used, CRYPTO_dbg_set_options can be used to
15596 request additional information:
15597 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15598 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15600 Also, things like CRYPTO_set_mem_functions will always give the
15601 expected result (the new set of functions is used for allocation
15602 and deallocation) at all times, regardless of platform and compiler
15605 To finish it up, some functions that were never use in any other
15606 way than through macros have a new API and new semantic:
15608 CRYPTO_dbg_malloc()
15609 CRYPTO_dbg_realloc()
15612 All macros of value have retained their old syntax.
15614 *Richard Levitte and Bodo Moeller*
15616 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15617 ordering of SMIMECapabilities wasn't in "strength order" and there
15618 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15623 * Some ASN1 types with illegal zero length encoding (INTEGER,
15624 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15626 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15628 * Merge in my S/MIME library for OpenSSL. This provides a simple
15629 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15630 functionality to handle multipart/signed properly) and a utility
15631 called 'smime' to call all this stuff. This is based on code I
15632 originally wrote for Celo who have kindly allowed it to be
15633 included in OpenSSL.
15637 * Add variants des_set_key_checked and des_set_key_unchecked of
15638 des_set_key (aka des_key_sched). Global variable des_check_key
15639 decides which of these is called by des_set_key; this way
15640 des_check_key behaves as it always did, but applications and
15641 the library itself, which was buggy for des_check_key == 1,
15642 have a cleaner way to pick the version they need.
15646 * New function PKCS12_newpass() which changes the password of a
15651 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15652 dynamic mix. In both cases the ids can be used as an index into the
15653 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15654 functions so they accept a list of the field values and the
15655 application doesn't need to directly manipulate the X509_TRUST
15660 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15665 * Modify the way the V3 extension code looks up extensions. This now
15666 works in a similar way to the object code: we have some "standard"
15667 extensions in a static table which is searched with OBJ_bsearch()
15668 and the application can add dynamic ones if needed. The file
15669 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15670 updated whenever a new extension is added to the core code and kept
15671 in ext_nid order. There is a simple program 'tabtest.c' which checks
15672 this. New extensions are not added too often so this file can readily
15673 be maintained manually.
15675 There are two big advantages in doing things this way. The extensions
15676 can be looked up immediately and no longer need to be "added" using
15677 X509V3_add_standard_extensions(): this function now does nothing.
15678 *Side note: I get *lots* of email saying the extension code doesn't
15679 work because people forget to call this function*
15680 Also no dynamic allocation is done unless new extensions are added:
15681 so if we don't add custom extensions there is no need to call
15682 X509V3_EXT_cleanup().
15686 * Modify enc utility's salting as follows: make salting the default. Add a
15687 magic header, so unsalted files fail gracefully instead of just decrypting
15688 to garbage. This is because not salting is a big security hole, so people
15689 should be discouraged from doing it.
15693 * Fixes and enhancements to the 'x509' utility. It allowed a message
15694 digest to be passed on the command line but it only used this
15695 parameter when signing a certificate. Modified so all relevant
15696 operations are affected by the digest parameter including the
15697 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15698 DSA key was used because it didn't fix the digest.
15702 * Initial certificate chain verify code. Currently tests the untrusted
15703 certificates for consistency with the verify purpose (which is set
15704 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15706 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15707 this is because it will reject chains with invalid extensions whereas
15708 every previous version of OpenSSL and SSLeay made no checks at all.
15710 Trust code: checks the root CA for the relevant trust settings. Trust
15711 settings have an initial value consistent with the verify purpose: e.g.
15712 if the verify purpose is for SSL client use it expects the CA to be
15713 trusted for SSL client use. However the default value can be changed to
15714 permit custom trust settings: one example of this would be to only trust
15715 certificates from a specific "secure" set of CAs.
15717 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15718 which should be used for version portability: especially since the
15719 verify structure is likely to change more often now.
15721 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15722 to set them. If not set then assume SSL clients will verify SSL servers
15725 Two new options to the verify program: -untrusted allows a set of
15726 untrusted certificates to be passed in and -purpose which sets the
15727 intended purpose of the certificate. If a purpose is set then the
15728 new chain verify code is used to check extension consistency.
15732 * Support for the authority information access extension.
15736 * Modify RSA and DSA PEM read routines to transparently handle
15737 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15738 public keys in a format compatible with certificate
15739 SubjectPublicKeyInfo structures. Unfortunately there were already
15740 functions called *_PublicKey_* which used various odd formats so
15741 these are retained for compatibility: however the DSA variants were
15742 never in a public release so they have been deleted. Changed dsa/rsa
15743 utilities to handle the new format: note no releases ever handled public
15744 keys so we should be OK.
15746 The primary motivation for this change is to avoid the same fiasco
15747 that dogs private keys: there are several incompatible private key
15748 formats some of which are standard and some OpenSSL specific and
15749 require various evil hacks to allow partial transparent handling and
15750 even then it doesn't work with DER formats. Given the option anything
15751 other than PKCS#8 should be dumped: but the other formats have to
15752 stay in the name of compatibility.
15754 With public keys and the benefit of hindsight one standard format
15755 is used which works with EVP_PKEY, RSA or DSA structures: though
15756 it clearly returns an error if you try to read the wrong kind of key.
15758 Added a -pubkey option to the 'x509' utility to output the public key.
15759 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
15760 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
15761 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
15762 that do the same as the EVP_PKEY_assign_*() except they up the
15763 reference count of the added key (they don't "swallow" the
15768 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15769 CRLs would fail if the file contained no certificates or no CRLs:
15770 added a new function to read in both types and return the number
15771 read: this means that if none are read it will be an error. The
15772 DER versions of the certificate and CRL reader would always fail
15773 because it isn't possible to mix certificates and CRLs in DER format
15774 without choking one or the other routine. Changed this to just read
15775 a certificate: this is the best we can do. Also modified the code
15776 in apps/verify.c to take notice of return codes: it was previously
15777 attempting to read in certificates from NULL pointers and ignoring
15778 any errors: this is one reason why the cert and CRL reader seemed
15779 to work. It doesn't check return codes from the default certificate
15780 routines: these may well fail if the certificates aren't installed.
15784 * Code to support otherName option in GeneralName.
15788 * First update to verify code. Change the verify utility
15789 so it warns if it is passed a self signed certificate:
15790 for consistency with the normal behaviour. X509_verify
15791 has been modified to it will now verify a self signed
15792 certificate if *exactly* the same certificate appears
15793 in the store: it was previously impossible to trust a
15794 single self signed certificate. This means that:
15795 openssl verify ss.pem
15796 now gives a warning about a self signed certificate but
15797 openssl verify -CAfile ss.pem ss.pem
15802 * For servers, store verify_result in SSL_SESSION data structure
15803 (and add it to external session representation).
15804 This is needed when client certificate verifications fails,
15805 but an application-provided verification callback (set by
15806 SSL_CTX_set_cert_verify_callback) allows accepting the session
15807 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15808 but returns 1): When the session is reused, we have to set
15809 ssl->verify_result to the appropriate error code to avoid
15812 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15814 * Fix a bug in the new PKCS#7 code: it didn't consider the
15815 case in PKCS7_dataInit() where the signed PKCS7 structure
15816 didn't contain any existing data because it was being created.
15818 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15820 * Add a salt to the key derivation routines in enc.c. This
15821 forms the first 8 bytes of the encrypted file. Also add a
15822 -S option to allow a salt to be input on the command line.
15826 * New function X509_cmp(). Oddly enough there wasn't a function
15827 to compare two certificates. We do this by working out the SHA1
15828 hash and comparing that. X509_cmp() will be needed by the trust
15833 * SSL_get1_session() is like SSL_get_session(), but increments
15834 the reference count in the SSL_SESSION returned.
15836 *Geoff Thorpe <geoff@eu.c2.net>*
15838 * Fix for 'req': it was adding a null to request attributes.
15839 Also change the X509_LOOKUP and X509_INFO code to handle
15840 certificate auxiliary information.
15844 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
15849 * Add the possibility to add extra information to the memory leak
15850 detecting output, to form tracebacks, showing from where each
15851 allocation was originated: CRYPTO_push_info("constant string") adds
15852 the string plus current file name and line number to a per-thread
15853 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
15854 is like calling CYRPTO_pop_info() until the stack is empty.
15855 Also updated memory leak detection code to be multi-thread-safe.
15859 * Add options -text and -noout to pkcs7 utility and delete the
15860 encryption options which never did anything. Update docs.
15864 * Add options to some of the utilities to allow the pass phrase
15865 to be included on either the command line (not recommended on
15866 OSes like Unix) or read from the environment. Update the
15867 manpages and fix a few bugs.
15871 * Add a few manpages for some of the openssl commands.
15875 * Fix the -revoke option in ca. It was freeing up memory twice,
15876 leaking and not finding already revoked certificates.
15880 * Extensive changes to support certificate auxiliary information.
15881 This involves the use of X509_CERT_AUX structure and X509_AUX
15882 functions. An X509_AUX function such as PEM_read_X509_AUX()
15883 can still read in a certificate file in the usual way but it
15884 will also read in any additional "auxiliary information". By
15885 doing things this way a fair degree of compatibility can be
15886 retained: existing certificates can have this information added
15887 using the new 'x509' options.
15889 Current auxiliary information includes an "alias" and some trust
15890 settings. The trust settings will ultimately be used in enhanced
15891 certificate chain verification routines: currently a certificate
15892 can only be trusted if it is self signed and then it is trusted
15897 * Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
15898 The problem was that one of the replacement routines had not been working
15899 since SSLeay releases. For now the offending routine has been replaced
15900 with non-optimised assembler. Even so, this now gives around 95%
15901 performance improvement for 1024 bit RSA signs.
15905 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
15906 handling. Most clients have the effective key size in bits equal to
15907 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
15908 A few however don't do this and instead use the size of the decrypted key
15909 to determine the RC2 key length and the AlgorithmIdentifier to determine
15910 the effective key length. In this case the effective key length can still
15911 be 40 bits but the key length can be 168 bits for example. This is fixed
15912 by manually forcing an RC2 key into the EVP_PKEY structure because the
15913 EVP code can't currently handle unusual RC2 key sizes: it always assumes
15914 the key length and effective key length are equal.
15918 * Add a bunch of functions that should simplify the creation of
15919 X509_NAME structures. Now you should be able to do:
15920 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
15921 and have it automatically work out the correct field type and fill in
15922 the structures. The more adventurous can try:
15923 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
15924 and it will (hopefully) work out the correct multibyte encoding.
15928 * Change the 'req' utility to use the new field handling and multibyte
15929 copy routines. Before the DN field creation was handled in an ad hoc
15930 way in req, ca, and x509 which was rather broken and didn't support
15931 BMPStrings or UTF8Strings. Since some software doesn't implement
15932 BMPStrings or UTF8Strings yet, they can be enabled using the config file
15933 using the dirstring_type option. See the new comment in the default
15934 openssl.cnf for more info.
15938 * Make crypto/rand/md_rand.c more robust:
15939 - Assure unique random numbers after fork().
15940 - Make sure that concurrent threads access the global counter and
15941 md serializably so that we never lose entropy in them
15942 or use exactly the same state in multiple threads.
15943 Access to the large state is not always serializable because
15944 the additional locking could be a performance killer, and
15945 md should be large enough anyway.
15949 * New file apps/app_rand.c with commonly needed functionality
15950 for handling the random seed file.
15952 Use the random seed file in some applications that previously did not:
15954 dsaparam -genkey (which also ignored its '-rand' option),
15957 x509 (when signing).
15958 Except on systems with /dev/urandom, it is crucial to have a random
15959 seed file at least for key creation, DSA signing, and for DH exchanges;
15960 for RSA signatures we could do without one.
15962 gendh and gendsa (unlike genrsa) used to read only the first byte
15963 of each file listed in the '-rand' option. The function as previously
15964 found in genrsa is now in app_rand.c and is used by all programs
15965 that support '-rand'.
15969 * In RAND_write_file, use mode 0600 for creating files;
15970 don't just chmod when it may be too late.
15974 * Report an error from X509_STORE_load_locations
15975 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
15979 * New function ASN1_mbstring_copy() this copies a string in either
15980 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
15981 into an ASN1_STRING type. A mask of permissible types is passed
15982 and it chooses the "minimal" type to use or an error if not type
15987 * Add function equivalents to the various macros in asn1.h. The old
15988 macros are retained with an M_ prefix. Code inside the library can
15989 use the M_ macros. External code (including the openssl utility)
15990 should *NOT* in order to be "shared library friendly".
15994 * Add various functions that can check a certificate's extensions
15995 to see if it usable for various purposes such as SSL client,
15996 server or S/MIME and CAs of these types. This is currently
15997 VERY EXPERIMENTAL but will ultimately be used for certificate chain
15998 verification. Also added a -purpose flag to x509 utility to
15999 print out all the purposes.
16003 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16008 * New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
16009 for, obtain and decode and extension and obtain its critical flag.
16010 This allows all the necessary extension code to be handled in a
16011 single function call.
16015 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16016 platforms. See crypto/rc4/rc4_enc.c for further details.
16020 * New -noout option to asn1parse. This causes no output to be produced
16021 its main use is when combined with -strparse and -out to extract data
16022 from a file (which may not be in ASN.1 format).
16026 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16027 when producing the local key id.
16029 *Richard Levitte <levitte@stacken.kth.se>*
16031 * New option -dhparam in s_server. This allows a DH parameter file to be
16032 stated explicitly. If it is not stated then it tries the first server
16033 certificate file. The previous behaviour hard coded the filename
16038 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16039 a public key to be input or output. For example:
16040 openssl rsa -in key.pem -pubout -out pubkey.pem
16041 Also added necessary DSA public key functions to handle this.
16045 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16046 in the message. This was handled by allowing
16047 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16049 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16051 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16052 to the end of the strings whereas this didn't. This would cause problems
16053 if strings read with d2i_ASN1_bytes() were later modified.
16055 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16057 * Fix for base64 decode bug. When a base64 bio reads only one line of
16058 data and it contains EOF it will end up returning an error. This is
16059 caused by input 46 bytes long. The cause is due to the way base64
16060 BIOs find the start of base64 encoded data. They do this by trying a
16061 trial decode on each line until they find one that works. When they
16062 do a flag is set and it starts again knowing it can pass all the
16063 data directly through the decoder. Unfortunately it doesn't reset
16064 the context it uses. This means that if EOF is reached an attempt
16065 is made to pass two EOFs through the context and this causes the
16066 resulting error. This can also cause other problems as well. As is
16067 usual with these problems it takes *ages* to find and the fix is
16068 trivial: move one line.
16070 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) *
16072 * Ugly workaround to get s_client and s_server working under Windows. The
16073 old code wouldn't work because it needed to select() on sockets and the
16074 tty (for keypresses and to see if data could be written). Win32 only
16075 supports select() on sockets so we select() with a 1s timeout on the
16076 sockets and then see if any characters are waiting to be read, if none
16077 are present then we retry, we also assume we can always write data to
16078 the tty. This isn't nice because the code then blocks until we've
16079 received a complete line of data and it is effectively polling the
16080 keyboard at 1s intervals: however it's quite a bit better than not
16081 working at all :-) A dedicated Windows application might handle this
16082 with an event loop for example.
16086 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16087 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16088 will be called when RSA_sign() and RSA_verify() are used. This is useful
16089 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16090 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16091 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16092 This necessitated the support of an extra signature type NID_md5_sha1
16093 for SSL signatures and modifications to the SSL library to use it instead
16094 of calling RSA_public_decrypt() and RSA_private_encrypt().
16098 * Add new -verify -CAfile and -CApath options to the crl program, these
16099 will lookup a CRL issuers certificate and verify the signature in a
16100 similar way to the verify program. Tidy up the crl program so it
16101 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16102 less strict. It will now permit CRL extensions even if it is not
16103 a V2 CRL: this will allow it to tolerate some broken CRLs.
16107 * Initialize all non-automatic variables each time one of the openssl
16108 sub-programs is started (this is necessary as they may be started
16109 multiple times from the "OpenSSL>" prompt).
16111 *Lennart Bang, Bodo Moeller*
16113 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16114 removing all other RSA functionality (this is what NO_RSA does). This
16115 is so (for example) those in the US can disable those operations covered
16116 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16121 * Non-copying interface to BIO pairs.
16122 (still largely untested)
16126 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16127 ASCII string. This was handled independently in various places before.
16131 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16132 UTF8 strings a character at a time.
16136 * Use client_version from client hello to select the protocol
16137 (s23_srvr.c) and for RSA client key exchange verification
16138 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16142 * Add various utility functions to handle SPKACs, these were previously
16143 handled by poking round in the structure internals. Added new function
16144 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16145 print, verify and generate SPKACs. Based on an original idea from
16146 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16150 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16154 * Allow the config file extension section to be overwritten on the
16155 command line. Based on an original idea from Massimiliano Pala
16156 <madwolf@comune.modena.it>. The new option is called -extensions
16157 and can be applied to ca, req and x509. Also -reqexts to override
16158 the request extensions in req and -crlexts to override the crl extensions
16163 * Add new feature to the SPKAC handling in ca. Now you can include
16164 the same field multiple times by preceding it by "XXXX." for example:
16167 this is the same syntax as used in the req config file.
16171 * Allow certificate extensions to be added to certificate requests. These
16172 are specified in a 'req_extensions' option of the req section of the
16173 config file. They can be printed out with the -text option to req but
16174 are otherwise ignored at present.
16178 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16179 data read consists of only the final block it would not decrypted because
16180 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16181 A misplaced 'break' also meant the decrypted final block might not be
16182 copied until the next read.
16186 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16187 a few extra parameters to the DH structure: these will be useful if
16188 for example we want the value of 'q' or implement X9.42 DH.
16192 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16193 provides hooks that allow the default DSA functions or functions on a
16194 "per key" basis to be replaced. This allows hardware acceleration and
16195 hardware key storage to be handled without major modification to the
16196 library. Also added low level modexp hooks and CRYPTO_EX structure and
16197 associated functions.
16201 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16202 as "read only": it can't be written to and the buffer it points to will
16203 not be freed. Reading from a read only BIO is much more efficient than
16204 a normal memory BIO. This was added because there are several times when
16205 an area of memory needs to be read from a BIO. The previous method was
16206 to create a memory BIO and write the data to it, this results in two
16207 copies of the data and an O(n^2) reading algorithm. There is a new
16208 function BIO_new_mem_buf() which creates a read only memory BIO from
16209 an area of memory. Also modified the PKCS#7 routines to use read only
16214 * Bugfix: ssl23_get_client_hello did not work properly when called in
16215 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16216 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16217 but a retry condition occurred while trying to read the rest.
16221 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16222 NID_pkcs7_encrypted by default: this was wrong since this should almost
16223 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16224 the encrypted data type: this is a more sensible place to put it and it
16225 allows the PKCS#12 code to be tidied up that duplicated this
16230 * Changed obj_dat.pl script so it takes its input and output files on
16231 the command line. This should avoid shell escape redirection problems
16236 * Initial support for certificate extension requests, these are included
16237 in things like Xenroll certificate requests. Included functions to allow
16238 extensions to be obtained and added.
16242 * -crlf option to s_client and s_server for sending newlines as
16243 CRLF (as required by many protocols).
16247 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999] ###
16249 * Install libRSAglue.a when OpenSSL is built with RSAref.
16251 *Ralf S. Engelschall*
16253 * A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
16255 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16257 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16262 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16263 DH parameters/keys (q is lost during that conversion, but the resulting
16264 DH parameters contain its length).
16266 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16267 much faster than DH_generate_parameters (which creates parameters
16268 where p = 2*q + 1), and also the smaller q makes DH computations
16269 much more efficient (160-bit exponentiation instead of 1024-bit
16270 exponentiation); so this provides a convenient way to support DHE
16271 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16272 utter importance to use
16273 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16275 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16276 when such DH parameters are used, because otherwise small subgroup
16277 attacks may become possible!
16281 * Avoid memory leak in i2d_DHparams.
16285 * Allow the -k option to be used more than once in the enc program:
16286 this allows the same encrypted message to be read by multiple recipients.
16290 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16291 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16292 it will always use the numerical form of the OID, even if it has a short
16297 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16298 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16299 otherwise bn_mod_exp was called. In the case of hardware keys for example
16300 no private key components need be present and it might store extra data
16301 in the RSA structure, which cannot be accessed from bn_mod_exp.
16302 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16303 private key operations.
16307 * Added support for SPARC Linux.
16311 * pem_password_cb function type incompatibly changed from
16312 typedef int pem_password_cb(char *buf, int size, int rwflag);
16314 ....(char *buf, int size, int rwflag, void *userdata);
16315 so that applications can pass data to their callbacks:
16316 The PEM[_ASN1]_{read,write}... functions and macros now take an
16317 additional void * argument, which is just handed through whenever
16318 the password callback is called.
16320 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16322 New function SSL_CTX_set_default_passwd_cb_userdata.
16324 Compatibility note: As many C implementations push function arguments
16325 onto the stack in reverse order, the new library version is likely to
16326 interoperate with programs that have been compiled with the old
16327 pem_password_cb definition (PEM_whatever takes some data that
16328 happens to be on the stack as its last argument, and the callback
16329 just ignores this garbage); but there is no guarantee whatsoever that
16332 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16333 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16334 problems not only on Windows, but also on some Unix platforms.
16335 To avoid problematic command lines, these definitions are now in an
16336 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16337 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16341 * MIPS III/IV assembler module is reimplemented.
16345 * More DES library cleanups: remove references to srand/rand and
16346 delete an unused file.
16350 * Add support for the free Netwide assembler (NASM) under Win32,
16351 since not many people have MASM (ml) and it can be hard to obtain.
16352 This is currently experimental but it seems to work OK and pass all
16353 the tests. Check out INSTALL.W32 for info.
16357 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16358 without temporary keys kept an extra copy of the server key,
16359 and connections with temporary keys did not free everything in case
16364 * New function RSA_check_key and new openssl rsa option -check
16365 for verifying the consistency of RSA keys.
16367 *Ulf Moeller, Bodo Moeller*
16369 * Various changes to make Win32 compile work:
16370 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16371 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16372 comparison" warnings.
16373 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
16377 * Add a debugging option to PKCS#5 v2 key generation function: when
16378 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16379 derived keys are printed to stderr.
16383 * Copy the flags in ASN1_STRING_dup().
16385 *Roman E. Pavlov <pre@mo.msk.ru>*
16387 * The x509 application mishandled signing requests containing DSA
16388 keys when the signing key was also DSA and the parameters didn't match.
16390 It was supposed to omit the parameters when they matched the signing key:
16391 the verifying software was then supposed to automatically use the CA's
16392 parameters if they were absent from the end user certificate.
16394 Omitting parameters is no longer recommended. The test was also
16395 the wrong way round! This was probably due to unusual behaviour in
16396 EVP_cmp_parameters() which returns 1 if the parameters match.
16397 This meant that parameters were omitted when they *didn't* match and
16398 the certificate was useless. Certificates signed with 'ca' didn't have
16401 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16403 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16404 The interface is as follows:
16405 Applications can use
16406 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16407 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16408 "off" is now the default.
16409 The library internally uses
16410 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16411 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16412 to disable memory-checking temporarily.
16414 Some inconsistent states that previously were possible (and were
16415 even the default) are now avoided.
16417 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16418 with each memory chunk allocated; this is occasionally more helpful
16419 than just having a counter.
16421 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16423 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16428 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16429 which largely parallels "options", but is for changing API behaviour,
16430 whereas "options" are about protocol behaviour.
16431 Initial "mode" flags are:
16433 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16434 a single record has been written.
16435 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16436 retries use the same buffer location.
16437 (But all of the contents must be
16442 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16445 * Fix problems with no-hmac etc.
16447 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16449 * New functions RSA_get_default_method(), RSA_set_method() and
16450 RSA_get_method(). These allows replacement of RSA_METHODs without having
16451 to mess around with the internals of an RSA structure.
16455 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16456 Also really enable memory leak checks in openssl.c and in some
16459 *Chad C. Mulligan, Bodo Moeller*
16461 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16462 up the length of negative integers. This has now been simplified to just
16463 store the length when it is first determined and use it later, rather
16464 than trying to keep track of where data is copied and updating it to
16466 *Steve Henson, reported by Brien Wheeler
16467 <bwheeler@authentica-security.com>*
16469 * Add a new function PKCS7_signatureVerify. This allows the verification
16470 of a PKCS#7 signature but with the signing certificate passed to the
16471 function itself. This contrasts with PKCS7_dataVerify which assumes the
16472 certificate is present in the PKCS#7 structure. This isn't always the
16473 case: certificates can be omitted from a PKCS#7 structure and be
16474 distributed by "out of band" means (such as a certificate database).
16478 * Complete the PEM_* macros with DECLARE_PEM versions to replace the
16479 function prototypes in pem.h, also change util/mkdef.pl to add the
16480 necessary function names.
16484 * mk1mf.pl (used by Windows builds) did not properly read the
16485 options set by Configure in the top level Makefile, and Configure
16486 was not even able to write more than one option correctly.
16487 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16491 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16492 file to be loaded from a BIO or FILE pointer. The BIO version will
16493 for example allow memory BIOs to contain config info.
16497 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16498 Whoever hopes to achieve shared-library compatibility across versions
16499 must use this, not the compile-time macro.
16500 (Exercise 0.9.4: Which is the minimum library version required by
16502 Note: All this applies only to multi-threaded programs, others don't
16507 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16508 through a BIO pair triggered the default case, i.e.
16509 SSLerr(...,SSL_R_UNKNOWN_STATE).
16513 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16514 can use the SSL library even if none of the specific BIOs is
16519 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16520 for the encoded length.
16522 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16524 * Add initial documentation of the X509V3 functions.
16528 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16529 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16530 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16531 secure PKCS#8 private key format with a high iteration count.
16535 * Fix determination of Perl interpreter: A perl or perl5
16536 _directory_ in $PATH was also accepted as the interpreter.
16538 *Ralf S. Engelschall*
16540 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16541 wrong with it but it was very old and did things like calling
16542 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16543 unusual formatting.
16547 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16548 to use the new extension code.
16552 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16553 with macros. This should make it easier to change their form, add extra
16554 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16559 * Add to configuration table a new entry that can specify an alternative
16560 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16561 according to Mark Crispin <MRC@Panda.COM>.
16566 * DES CBC did not update the IV. Weird.
16570 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16571 Changing the behaviour of the former might break existing programs --
16572 where IV updating is needed, des_ncbc_encrypt can be used.
16575 * When bntest is run from "make test" it drives bc to check its
16576 calculations, as well as internally checking them. If an internal check
16577 fails, it needs to cause bc to give a non-zero result or make test carries
16578 on without noticing the failure. Fixed.
16582 * DES library cleanups.
16586 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16587 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16588 ciphers. NOTE: although the key derivation function has been verified
16589 against some published test vectors it has not been extensively tested
16590 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16595 * Instead of "mkdir -p", which is not fully portable, use new
16596 Perl script "util/mkdir-p.pl".
16600 * Rewrite the way password based encryption (PBE) is handled. It used to
16601 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16602 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16603 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16604 the 'parameter' field of the AlgorithmIdentifier is passed to the
16605 underlying key generation function so it must do its own ASN1 parsing.
16606 This has also changed the EVP_PBE_CipherInit() function which now has a
16607 'parameter' argument instead of literal salt and iteration count values
16608 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16612 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16613 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16614 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16615 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16616 value was just used as a "magic string" and not used directly its
16617 value doesn't matter.
16621 * Introduce some semblance of const correctness to BN. Shame C doesn't
16626 * "linux-sparc64" configuration (ultrapenguin).
16628 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16629 "linux-sparc" configuration.
16631 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16633 * config now generates no-xxx options for missing ciphers.
16637 * Support the EBCDIC character set (work in progress).
16638 File ebcdic.c not yet included because it has a different license.
16640 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16642 * Support BS2000/OSD-POSIX.
16644 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16646 * Make callbacks for key generation use void * instead of char *.
16650 * Make S/MIME samples compile (not yet tested).
16654 * Additional typesafe stacks.
16658 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16663 ### Changes between 0.9.3 and 0.9.3a [29 May 1999] ###
16665 * New configuration variant "sco5-gcc".
16667 * Updated some demos.
16669 *Sean O Riordain, Wade Scholine*
16671 * Add missing BIO_free at exit of pkcs12 application.
16675 * Fix memory leak in conf.c.
16679 * Updates for Win32 to assembler version of MD5.
16683 * Set #! path to perl in apps/der_chop to where we found it
16684 instead of using a fixed path.
16688 * SHA library changes for irix64-mips4-cc.
16692 * Improvements for VMS support.
16697 ### Changes between 0.9.2b and 0.9.3 [24 May 1999] ###
16699 * Bignum library bug fix. IRIX 6 passes "make test" now!
16700 This also avoids the problems with SC4.2 and unpatched SC5.
16702 *Andy Polyakov <appro@fy.chalmers.se>*
16704 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16705 These are required because of the typesafe stack would otherwise break
16706 existing code. If old code used a structure member which used to be STACK
16707 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16708 sk_num or sk_value it would produce an error because the num, data members
16709 are not present in STACK_OF. Now it just produces a warning. sk_set
16710 replaces the old method of assigning a value to sk_value
16711 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16712 that does this will no longer work (and should use sk_set instead) but
16713 this could be regarded as a "questionable" behaviour anyway.
16717 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16718 correctly handle encrypted S/MIME data.
16722 * Change type of various DES function arguments from des_cblock
16723 (which means, in function argument declarations, pointer to char)
16724 to des_cblock * (meaning pointer to array with 8 char elements),
16725 which allows the compiler to do more typechecking; it was like
16726 that back in SSLeay, but with lots of ugly casts.
16728 Introduce new type const_des_cblock.
16732 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16733 problems: find RecipientInfo structure that matches recipient certificate
16734 and initialise the ASN1 structures properly based on passed cipher.
16738 * Belatedly make the BN tests actually check the results.
16742 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16743 to and from BNs: it was completely broken. New compilation option
16744 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16745 key elements as negative integers.
16749 * Reorganize and speed up MD5.
16751 *Andy Polyakov <appro@fy.chalmers.se>*
16755 *Richard Levitte <richard@levitte.org>*
16757 * New option -out to asn1parse to allow the parsed structure to be
16758 output to a file. This is most useful when combined with the -strparse
16759 option to examine the output of things like OCTET STRINGS.
16763 * Make SSL library a little more fool-proof by not requiring any longer
16764 that SSL_set_{accept,connect}_state be called before
16765 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
16766 in many applications because usually everything *appeared* to work as
16767 intended anyway -- now it really works as intended).
16771 * Move openssl.cnf out of lib/.
16775 * Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
16776 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16777 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
16779 *Ralf S. Engelschall*
16781 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16782 handle PKCS#7 enveloped data properly.
16784 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16786 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16787 copying pointers. The cert_st handling is changed by this in
16788 various ways (and thus what used to be known as ctx->default_cert
16789 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
16790 any longer when s->cert does not give us what we need).
16791 ssl_cert_instantiate becomes obsolete by this change.
16792 As soon as we've got the new code right (possibly it already is?),
16793 we have solved a couple of bugs of the earlier code where s->cert
16794 was used as if it could not have been shared with other SSL structures.
16796 Note that using the SSL API in certain dirty ways now will result
16797 in different behaviour than observed with earlier library versions:
16798 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
16799 does not influence s as it used to.
16801 In order to clean up things more thoroughly, inside SSL_SESSION
16802 we don't use CERT any longer, but a new structure SESS_CERT
16803 that holds per-session data (if available); currently, this is
16804 the peer's certificate chain and, for clients, the server's certificate
16805 and temporary key. CERT holds only those values that can have
16806 meaningful defaults in an SSL_CTX.
16810 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16811 from the internal representation. Various PKCS#7 fixes: remove some
16812 evil casts and set the enc_dig_alg field properly based on the signing
16817 * Allow PKCS#12 password to be set from the command line or the
16818 environment. Let 'ca' get its config file name from the environment
16819 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16824 * Allow certificate policies extension to use an IA5STRING for the
16825 organization field. This is contrary to the PKIX definition but
16826 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16831 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
16832 without disallowing inline assembler and the like for non-pedantic builds.
16836 * Support Borland C++ builder.
16838 *Janez Jere <jj@void.si>, modified by Ulf Möller*
16844 * SHA-1 cleanups and performance enhancements.
16846 *Andy Polyakov <appro@fy.chalmers.se>*
16848 * Sparc v8plus assembler for the bignum library.
16850 *Andy Polyakov <appro@fy.chalmers.se>*
16852 * Accept any -xxx and +xxx compiler options in Configure.
16856 * Update HPUX configuration.
16860 * Add missing sk_<type>_unshift() function to safestack.h
16862 *Ralf S. Engelschall*
16864 * New function SSL_CTX_use_certificate_chain_file that sets the
16865 "extra_cert"s in addition to the certificate. (This makes sense
16866 only for "PEM" format files, as chains as a whole are not
16871 * Support verify_depth from the SSL API.
16872 x509_vfy.c had what can be considered an off-by-one-error:
16873 Its depth (which was not part of the external interface)
16874 was actually counting the number of certificates in a chain;
16875 now it really counts the depth.
16879 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
16880 instead of X509err, which often resulted in confusing error
16881 messages since the error codes are not globally unique
16882 (e.g. an alleged error in ssl3_accept when a certificate
16883 didn't match the private key).
16885 * New function SSL_CTX_set_session_id_context that allows to set a default
16886 value (so that you don't need SSL_set_session_id_context for each
16887 connection using the SSL_CTX).
16891 * OAEP decoding bug fix.
16895 * Support INSTALL_PREFIX for package builders, as proposed by
16900 * New Configure options "threads" and "no-threads". For systems
16901 where the proper compiler options are known (currently Solaris
16902 and Linux), "threads" is the default.
16906 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
16910 * Install various scripts to $(OPENSSLDIR)/misc, not to
16911 $(INSTALLTOP)/bin -- they shouldn't clutter directories
16912 such as /usr/local/bin.
16916 * "make linux-shared" to build shared libraries.
16918 *Niels Poppe <niels@netbox.org>*
16920 * New Configure option no-<cipher> (rsa, idea, rc5, ...).
16924 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
16925 extension adding in x509 utility.
16929 * Remove NOPROTO sections and error code comments.
16933 * Partial rewrite of the DEF file generator to now parse the ANSI
16938 * New Configure options --prefix=DIR and --openssldir=DIR.
16942 * Complete rewrite of the error code script(s). It is all now handled
16943 by one script at the top level which handles error code gathering,
16944 header rewriting and C source file generation. It should be much better
16945 than the old method: it now uses a modified version of Ulf's parser to
16946 read the ANSI prototypes in all header files (thus the old K&R definitions
16947 aren't needed for error creation any more) and do a better job of
16948 translating function codes into names. The old 'ASN1 error code embedded
16949 in a comment' is no longer necessary and it doesn't use .err files which
16950 have now been deleted. Also the error code call doesn't have to appear all
16951 on one line (which resulted in some large lines...).
16955 * Change #include filenames from <foo.h> to <openssl/foo.h>.
16959 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
16960 0 (which usually indicates a closed connection), but continue reading.
16964 * Fix some race conditions.
16968 * Add support for CRL distribution points extension. Add Certificate
16969 Policies and CRL distribution points documentation.
16973 * Move the autogenerated header file parts to crypto/opensslconf.h.
16977 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
16978 8 of keying material. Merlin has also confirmed interop with this fix
16979 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
16981 *Merlin Hughes <merlin@baltimore.ie>*
16983 * Fix lots of warnings.
16985 *Richard Levitte <levitte@stacken.kth.se>*
16987 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
16988 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
16990 *Richard Levitte <levitte@stacken.kth.se>*
16992 * Fix problems with sizeof(long) == 8.
16994 *Andy Polyakov <appro@fy.chalmers.se>*
16996 * Change functions to ANSI C.
17000 * Fix typos in error codes.
17002 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17004 * Remove defunct assembler files from Configure.
17008 * SPARC v8 assembler BIGNUM implementation.
17010 *Andy Polyakov <appro@fy.chalmers.se>*
17012 * Support for Certificate Policies extension: both print and set.
17013 Various additions to support the r2i method this uses.
17017 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17018 return a const string when you are expecting an allocated buffer.
17022 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17023 types DirectoryString and DisplayText.
17027 * Add code to allow r2i extensions to access the configuration database,
17028 add an LHASH database driver and add several ctx helper functions.
17032 * Fix an evil bug in bn_expand2() which caused various BN functions to
17033 fail when they extended the size of a BIGNUM.
17037 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17038 support typesafe stack.
17042 * Fix typo in SSL_[gs]et_options().
17044 *Nils Frostberg <nils@medcom.se>*
17046 * Delete various functions and files that belonged to the (now obsolete)
17047 old X509V3 handling code.
17051 * New Configure option "rsaref".
17055 * Don't auto-generate pem.h.
17059 * Introduce type-safe ASN.1 SETs.
17063 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17065 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17067 * Introduce type-safe STACKs. This will almost certainly break lots of code
17068 that links with OpenSSL (well at least cause lots of warnings), but fear
17069 not: the conversion is trivial, and it eliminates loads of evil casts. A
17070 few STACKed things have been converted already. Feel free to convert more.
17071 In the fullness of time, I'll do away with the STACK type altogether.
17075 * Add `openssl ca -revoke <certfile>' facility which revokes a certificate
17076 specified in <certfile> by updating the entry in the index.txt file.
17077 This way one no longer has to edit the index.txt file manually for
17078 revoking a certificate. The -revoke option does the gory details now.
17080 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17082 * Fix `openssl crl -noout -text' combination where `-noout' killed the
17083 `-text' option at all and this way the `-noout -text' combination was
17084 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
17086 *Ralf S. Engelschall*
17088 * Make sure a corresponding plain text error message exists for the
17089 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17090 verify callback function determined that a certificate was revoked.
17092 *Ralf S. Engelschall*
17094 * Bugfix: In test/testenc, don't test "openssl <cipher>" for
17095 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17096 all available ciphers including rc5, which was forgotten until now.
17097 In order to let the testing shell script know which algorithms
17098 are available, a new (up to now undocumented) command
17099 "openssl list-cipher-commands" is used.
17103 * Bugfix: s_client occasionally would sleep in select() when
17104 it should have checked SSL_pending() first.
17108 * New functions DSA_do_sign and DSA_do_verify to provide access to
17109 the raw DSA values prior to ASN.1 encoding.
17113 * Tweaks to Configure
17115 *Niels Poppe <niels@netbox.org>*
17117 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17122 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17126 * New config option to avoid instructions that are illegal on the 80386.
17127 The default code is faster, but requires at least a 486.
17131 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17132 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17133 same as SSL2_VERSION anyway.
17137 * New "-showcerts" option for s_client.
17141 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17142 application. Various cleanups and fixes.
17146 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17147 modify error routines to work internally. Add error codes and PBE init
17148 to library startup routines.
17152 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17153 packing functions to asn1 and evp. Changed function names and error
17154 codes along the way.
17158 * PKCS12 integration: and so it begins... First of several patches to
17159 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17160 objects to objects.h
17164 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17165 and display support for Thawte strong extranet extension.
17169 * Add LinuxPPC support.
17171 *Jeff Dubrule <igor@pobox.org>*
17173 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17174 bn_div_words in alpha.s.
17176 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17178 * Make sure the RSA OAEP test is skipped under -DRSAref because
17179 OAEP isn't supported when OpenSSL is built with RSAref.
17181 *Ulf Moeller <ulf@fitug.de>*
17183 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17184 so they no longer are missing under -DNOPROTO.
17186 *Soren S. Jorvang <soren@t.dk>*
17189 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999] ###
17191 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17192 doesn't work when the session is reused. Coming soon!
17196 * Fix a security hole, that allows sessions to be reused in the wrong
17197 context thus bypassing client cert protection! All software that uses
17198 client certs and session caches in multiple contexts NEEDS PATCHING to
17199 allow session reuse! A fuller solution is in the works.
17201 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17203 * Some more source tree cleanups (removed obsolete files
17204 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17205 permission on "config" script to be executable) and a fix for the INSTALL
17208 *Ulf Moeller <ulf@fitug.de>*
17210 * Remove some legacy and erroneous uses of malloc, free instead of
17213 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17215 * Make rsa_oaep_test return non-zero on error.
17217 *Ulf Moeller <ulf@fitug.de>*
17219 * Add support for native Solaris shared libraries. Configure
17220 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17221 if someone would make that last step automatic.
17223 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17225 * ctx_size was not built with the right compiler during "make links". Fixed.
17229 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17230 except NULL ciphers". This means the default cipher list will no longer
17231 enable NULL ciphers. They need to be specifically enabled e.g. with
17232 the string "DEFAULT:eNULL".
17236 * Fix to RSA private encryption routines: if p < q then it would
17237 occasionally produce an invalid result. This will only happen with
17238 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17242 * Be less restrictive and allow also `perl util/perlpath.pl
17243 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
17244 because this way one can also use an interpreter named `perl5' (which is
17245 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17246 installed as `perl').
17248 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17250 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17252 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17254 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17255 advapi32.lib to Win32 build and change the pem test comparison
17256 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17257 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17258 and crypto/des/ede_cbcm_enc.c.
17262 * DES quad checksum was broken on big-endian architectures. Fixed.
17266 * Comment out two functions in bio.h that aren't implemented. Fix up the
17267 Win32 test batch file so it (might) work again. The Win32 test batch file
17268 is horrible: I feel ill....
17272 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17273 in e_os.h. Audit of header files to check ANSI and non ANSI
17274 sections: 10 functions were absent from non ANSI section and not exported
17275 from Windows DLLs. Fixed up libeay.num for new functions.
17279 * Make `openssl version' output lines consistent.
17281 *Ralf S. Engelschall*
17283 * Fix Win32 symbol export lists for BIO functions: Added
17284 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17285 to ms/libeay{16,32}.def.
17287 *Ralf S. Engelschall*
17289 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17290 fine under Unix and passes some trivial tests I've now added. But the
17291 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17292 added to make sure no one expects that this stuff really works in the
17293 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17294 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17297 *Ralf S. Engelschall*
17299 * Fix the generation of two part addresses in perl.
17301 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17303 * Add config entry for Linux on MIPS.
17305 *John Tobey <jtobey@channel1.com>*
17307 * Make links whenever Configure is run, unless we are on Windoze.
17311 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17312 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17317 * Add a useful kludge to allow package maintainers to specify compiler and
17318 other platforms details on the command line without having to patch the
17319 Configure script every time: One now can use ``perl Configure
17320 <id>:<details>'', i.e. platform ids are allowed to have details appended
17321 to them (separated by colons). This is treated as there would be a static
17322 pre-configured entry in Configure's %table under key <id> with value
17323 <details> and ``perl Configure <id>'' is called. So, when you want to
17324 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17325 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
17326 now, which overrides the FreeBSD-elf entry on-the-fly.
17328 *Ralf S. Engelschall*
17330 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17334 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17335 on the `perl Configure ...' command line. This way one can compile
17336 OpenSSL libraries with Position Independent Code (PIC) which is needed
17337 for linking it into DSOs.
17339 *Ralf S. Engelschall*
17341 * Remarkably, export ciphers were totally broken and no-one had noticed!
17346 * Cleaned up the LICENSE document: The official contact for any license
17347 questions now is the OpenSSL core team under openssl-core@openssl.org.
17348 And add a paragraph about the dual-license situation to make sure people
17349 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17350 to the OpenSSL toolkit.
17352 *Ralf S. Engelschall*
17354 * General source tree makefile cleanups: Made `making xxx in yyy...'
17355 display consistent in the source tree and replaced `/bin/rm' by `rm'.
17356 Additionally cleaned up the `make links' target: Remove unnecessary
17357 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17358 to speed processing and no longer clutter the display with confusing
17359 stuff. Instead only the actually done links are displayed.
17361 *Ralf S. Engelschall*
17363 * Permit null encryption ciphersuites, used for authentication only. It used
17364 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17365 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17370 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17371 signed attributes when verifying signatures (this would break them),
17372 the detached data encoding was wrong and public keys obtained using
17373 X509_get_pubkey() weren't freed.
17377 * Add text documentation for the BUFFER functions. Also added a work around
17378 to a Win95 console bug. This was triggered by the password read stuff: the
17379 last character typed gets carried over to the next fread(). If you were
17380 generating a new cert request using 'req' for example then the last
17381 character of the passphrase would be CR which would then enter the first
17386 * Added the new `Includes OpenSSL Cryptography Software' button as
17387 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17388 button and can be used by applications based on OpenSSL to show the
17389 relationship to the OpenSSL project.
17391 *Ralf S. Engelschall*
17393 * Remove confusing variables in function signatures in files
17394 ssl/ssl_lib.c and ssl/ssl.h.
17396 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17398 * Don't install bss_file.c under PREFIX/include/
17400 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17402 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17403 functions that return function pointers and has support for NT specific
17404 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17405 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17406 unsigned to signed types: this was killing the Win32 compile.
17410 * Add new certificate file to stack functions,
17411 SSL_add_dir_cert_subjects_to_stack() and
17412 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17413 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17414 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17415 This means that Apache-SSL and similar packages don't have to mess around
17416 to add as many CAs as they want to the preferred list.
17420 * Experiment with doxygen documentation. Currently only partially applied to
17422 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
17423 openssl.doxy as the configuration file.
17427 * Get rid of remaining C++-style comments which strict C compilers hate.
17429 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17431 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17432 compiled in by default: it has problems with large keys.
17436 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17437 DH private keys and/or callback functions which directly correspond to
17438 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17439 is needed for applications which have to configure certificates on a
17440 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17442 For the RSA certificate situation is makes no difference, but
17443 for the DSA certificate situation this fixes the "no shared cipher"
17444 problem where the OpenSSL cipher selection procedure failed because the
17445 temporary keys were not overtaken from the context and the API provided
17446 no way to reconfigure them.
17447 The new functions now let applications reconfigure the stuff and they
17448 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17449 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17450 non-public-API function ssl_cert_instantiate() is used as a helper
17451 function and also to reduce code redundancy inside ssl_rsa.c.
17453 *Ralf S. Engelschall*
17455 * Move s_server -dcert and -dkey options out of the undocumented feature
17456 area because they are useful for the DSA situation and should be
17457 recognized by the users.
17459 *Ralf S. Engelschall*
17461 * Fix the cipher decision scheme for export ciphers: the export bits are
17462 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17463 SSL_EXP_MASK. So, the original variable has to be used instead of the
17464 already masked variable.
17466 *Richard Levitte <levitte@stacken.kth.se>*
17468 * Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
17470 *Richard Levitte <levitte@stacken.kth.se>*
17472 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17473 from `int' to `unsigned int' because it's a length and initialized by
17474 EVP_DigestFinal() which expects an `unsigned int *'.
17476 *Richard Levitte <levitte@stacken.kth.se>*
17478 * Don't hard-code path to Perl interpreter on shebang line of Configure
17479 script. Instead use the usual Shell->Perl transition trick.
17481 *Ralf S. Engelschall*
17483 * Make `openssl x509 -noout -modulus' functional also for DSA certificates
17484 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17485 -noout -modulus' as it's already the case for `openssl rsa -noout
17486 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
17487 currently the public key is printed (a decision which was already done by
17488 `openssl dsa -modulus' in the past) which serves a similar purpose.
17489 Additionally the NO_RSA no longer completely removes the whole -modulus
17490 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17493 *Ralf S. Engelschall*
17495 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17496 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17498 *Arne Ansper <arne@ats.cyber.ee>*
17500 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17501 to be added. Now both 'req' and 'ca' can use new objects defined in the
17506 * Add cool BIO that does syslog (or event log on NT).
17508 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17510 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17511 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17512 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17513 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17517 * Add preliminary config info for new extension code.
17521 * Make RSA_NO_PADDING really use no padding.
17523 *Ulf Moeller <ulf@fitug.de>*
17525 * Generate errors when private/public key check is done.
17529 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17530 for some CRL extensions and new objects added.
17534 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17535 key usage extension and fuller support for authority key id.
17539 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17540 padding method for RSA, which is recommended for new applications in PKCS
17541 #1 v2.0 (RFC 2437, October 1998).
17542 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17543 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17544 against Bleichbacher's attack on RSA.
17545 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17548 * Updates to the new SSL compression code
17550 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17552 * Fix so that the version number in the master secret, when passed
17553 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17554 (because the server will not accept higher), that the version number
17555 is 0x03,0x01, not 0x03,0x00
17557 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17559 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17560 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17561 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17565 * Support for RAW extensions where an arbitrary extension can be
17566 created by including its DER encoding. See apps/openssl.cnf for
17571 * Make sure latest Perl versions don't interpret some generated C array
17572 code as Perl array code in the crypto/err/err_genc.pl script.
17574 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17576 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17577 not many people have the assembler. Various Win32 compilation fixes and
17578 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17579 build instructions.
17583 * Modify configure script 'Configure' to automatically create crypto/date.h
17584 file under Win32 and also build pem.h from pem.org. New script
17585 util/mkfiles.pl to create the MINFO file on environments that can't do a
17586 'make files': perl util/mkfiles.pl >MINFO should work.
17590 * Major rework of DES function declarations, in the pursuit of correctness
17591 and purity. As a result, many evil casts evaporated, and some weirdness,
17592 too. You may find this causes warnings in your code. Zapping your evil
17593 casts will probably fix them. Mostly.
17597 * Fix for a typo in asn1.h. Bug fix to object creation script
17598 obj_dat.pl. It considered a zero in an object definition to mean
17599 "end of object": none of the objects in objects.h have any zeros
17600 so it wasn't spotted.
17602 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17604 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17605 Masking (CBCM). In the absence of test vectors, the best I have been able
17606 to do is check that the decrypt undoes the encrypt, so far. Send me test
17607 vectors if you have them.
17611 * Correct calculation of key length for export ciphers (too much space was
17612 allocated for null ciphers). This has not been tested!
17616 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17617 message is now correct (it understands "crypto" and "ssl" on its
17618 command line). There is also now an "update" option. This will update
17619 the util/ssleay.num and util/libeay.num files with any new functions.
17621 perl util/mkdef.pl crypto ssl update
17622 it will update them.
17626 * Overhauled the Perl interface (perl/*):
17627 - ported BN stuff to OpenSSL's different BN library
17628 - made the perl/ source tree CVS-aware
17629 - renamed the package from SSLeay to OpenSSL (the files still contain
17630 their history because I've copied them in the repository)
17631 - removed obsolete files (the test scripts will be replaced
17632 by better Test::Harness variants in the future)
17634 *Ralf S. Engelschall*
17636 * First cut for a very conservative source tree cleanup:
17637 1. merge various obsolete readme texts into doc/ssleay.txt
17638 where we collect the old documents and readme texts.
17639 2. remove the first part of files where I'm already sure that we no
17640 longer need them because of three reasons: either they are just temporary
17641 files which were left by Eric or they are preserved original files where
17642 I've verified that the diff is also available in the CVS via "cvs diff
17643 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17644 the crypto/md/ stuff).
17646 *Ralf S. Engelschall*
17648 * More extension code. Incomplete support for subject and issuer alt
17649 name, issuer and authority key id. Change the i2v function parameters
17650 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17651 what that's for :-) Fix to ASN1 macro which messed up
17652 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17656 * Preliminary support for ENUMERATED type. This is largely copied from the
17661 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17663 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17665 * Make sure `make rehash' target really finds the `openssl' program.
17667 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17669 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17670 like to hear about it if this slows down other processors.
17674 * Add CygWin32 platform information to Configure script.
17676 *Alan Batie <batie@aahz.jf.intel.com>*
17678 * Fixed ms/32all.bat script: `no_asm' -> `no-asm'
17680 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17682 * New program nseq to manipulate netscape certificate sequences
17686 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17691 * Fixes to BN code. Previously the default was to define BN_RECURSION
17692 but the BN code had some problems that would cause failures when
17693 doing certificate verification and some other functions.
17695 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17697 * Add ASN1 and PEM code to support netscape certificate sequences.
17701 * Add ASN1 and PEM code to support netscape certificate sequences.
17705 * Add several PKIX and private extended key usage OIDs.
17709 * Modify the 'ca' program to handle the new extension code. Modify
17710 openssl.cnf for new extension format, add comments.
17714 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17715 and add a sample to openssl.cnf so req -x509 now adds appropriate
17720 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17721 error code, add initial support to X509_print() and x509 application.
17725 * Takes a deep breath and start adding X509 V3 extension support code. Add
17726 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17727 stuff is currently isolated and isn't even compiled yet.
17731 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17732 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17733 Removed the versions check from X509 routines when loading extensions:
17734 this allows certain broken certificates that don't set the version
17735 properly to be processed.
17739 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17740 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17741 can still be regenerated with "make depend".
17745 * Spelling mistake in C version of CAST-128.
17747 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17749 * Changes to the error generation code. The perl script err-code.pl
17750 now reads in the old error codes and retains the old numbers, only
17751 adding new ones if necessary. It also only changes the .err files if new
17752 codes are added. The makefiles have been modified to only insert errors
17753 when needed (to avoid needlessly modifying header files). This is done
17754 by only inserting errors if the .err file is newer than the auto generated
17755 C file. To rebuild all the error codes from scratch (the old behaviour)
17756 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17757 or delete all the .err files.
17761 * CAST-128 was incorrectly implemented for short keys. The C version has
17762 been fixed, but is untested. The assembler versions are also fixed, but
17763 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17764 to regenerate it if needed.
17765 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17766 Hagino <itojun@kame.net>*
17768 * File was opened incorrectly in randfile.c.
17770 *Ulf Möller <ulf@fitug.de>*
17772 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17773 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17774 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17775 al: it's just almost always a UTCTime. Note this patch adds new error
17776 codes so do a "make errors" if there are problems.
17780 * Correct Linux 1 recognition in config.
17782 *Ulf Möller <ulf@fitug.de>*
17784 * Remove pointless MD5 hash when using DSA keys in ca.
17786 *Anonymous <nobody@replay.com>*
17788 * Generate an error if given an empty string as a cert directory. Also
17789 generate an error if handed NULL (previously returned 0 to indicate an
17790 error, but didn't set one).
17792 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17794 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17798 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17799 parameters. This was causing a warning which killed off the Win32 compile.
17803 * Remove C++ style comments from crypto/bn/bn_local.h.
17805 *Neil Costigan <neil.costigan@celocom.com>*
17807 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17808 based on a text string, looking up short and long names and finally
17809 "dot" format. The "dot" format stuff didn't work. Added new function
17810 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17811 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17812 OID is not part of the table.
17816 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17817 X509_LOOKUP_by_alias().
17821 * Sort openssl functions by name.
17825 * Get the gendsa program working (hopefully) and add it to app list. Remove
17826 encryption from sample DSA keys (in case anyone is interested the password
17831 * Make _all_ *_free functions accept a NULL pointer.
17833 *Frans Heymans <fheymans@isaserver.be>*
17835 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
17838 *Anonymous <nobody@replay.com>*
17840 * s_server should send the CAfile as acceptable CAs, not its own cert.
17842 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17844 * Don't blow it for numeric -newkey arguments to apps/req.
17846 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
17848 * Temp key "for export" tests were wrong in s3_srvr.c.
17850 *Anonymous <nobody@replay.com>*
17852 * Add prototype for temp key callback functions
17853 SSL_CTX_set_tmp_{rsa,dh}_callback().
17857 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
17858 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
17862 * X509_name_add_entry() freed the wrong thing after an error.
17864 *Arne Ansper <arne@ats.cyber.ee>*
17866 * rsa_eay.c would attempt to free a NULL context.
17868 *Arne Ansper <arne@ats.cyber.ee>*
17870 * BIO_s_socket() had a broken should_retry() on Windoze.
17872 *Arne Ansper <arne@ats.cyber.ee>*
17874 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
17876 *Arne Ansper <arne@ats.cyber.ee>*
17878 * Make sure the already existing X509_STORE->depth variable is initialized
17879 in X509_STORE_new(), but document the fact that this variable is still
17880 unused in the certificate verification process.
17882 *Ralf S. Engelschall*
17884 * Fix the various library and apps files to free up pkeys obtained from
17885 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
17889 * Fix reference counting in X509_PUBKEY_get(). This makes
17890 demos/maurice/example2.c work, amongst others, probably.
17892 *Steve Henson and Ben Laurie*
17894 * First cut of a cleanup for apps/. First the `ssleay' program is now named
17895 `openssl' and second, the shortcut symlinks for the `openssl <command>'
17896 are no longer created. This way we have a single and consistent command
17897 line interface `openssl <command>', similar to `cvs <command>'.
17899 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
17901 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
17902 BIT STRING wrapper always have zero unused bits.
17906 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
17910 * Make the top-level INSTALL documentation easier to understand.
17914 * Makefiles updated to exit if an error occurs in a sub-directory
17915 make (including if user presses ^C) [Paul Sutton]
17917 * Make Montgomery context stuff explicit in RSA data structure.
17921 * Fix build order of pem and err to allow for generated pem.h.
17925 * Fix renumbering bug in X509_NAME_delete_entry().
17929 * Enhanced the err-ins.pl script so it makes the error library number
17930 global and can add a library name. This is needed for external ASN1 and
17931 other error libraries.
17935 * Fixed sk_insert which never worked properly.
17939 * Fix ASN1 macros so they can handle indefinite length constructed
17940 EXPLICIT tags. Some non standard certificates use these: they can now
17945 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
17946 into a single doc/ssleay.txt bundle. This way the information is still
17947 preserved but no longer messes up this directory. Now it's new room for
17948 the new set of documentation files.
17950 *Ralf S. Engelschall*
17952 * SETs were incorrectly DER encoded. This was a major pain, because they
17953 shared code with SEQUENCEs, which aren't coded the same. This means that
17954 almost everything to do with SETs or SEQUENCEs has either changed name or
17955 number of arguments.
17957 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
17959 * Fix test data to work with the above.
17963 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
17964 was already fixed by Eric for 0.9.1 it seems.
17966 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
17968 * Autodetect FreeBSD3.
17972 * Fix various bugs in Configure. This affects the following platforms:
17976 unixware-2.0-pentium
17981 * Eliminate generated files from CVS. Reorder tests to regenerate files
17982 before they are needed.
17986 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
17991 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998] ###
17993 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
17994 changed SSLeay to OpenSSL in version strings.
17996 *Ralf S. Engelschall*
17998 * Some fixups to the top-level documents.
18002 * Fixed the nasty bug where rsaref.h was not found under compile-time
18003 because the symlink to include/ was missing.
18005 *Ralf S. Engelschall*
18007 * Incorporated the popular no-RSA/DSA-only patches
18008 which allow to compile a RSA-free SSLeay.
18010 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18012 * Fixed nasty rehash problem under `make -f Makefile.ssl links'
18013 when "ssleay" is still not found.
18015 *Ralf S. Engelschall*
18017 * Added more platforms to Configure: Cray T3E, HPUX 11,
18019 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18021 * Updated the README file.
18023 *Ralf S. Engelschall*
18025 * Added various .cvsignore files in the CVS repository subdirs
18026 to make a "cvs update" really silent.
18028 *Ralf S. Engelschall*
18030 * Recompiled the error-definition header files and added
18031 missing symbols to the Win32 linker tables.
18033 *Ralf S. Engelschall*
18035 * Cleaned up the top-level documents;
18036 o new files: CHANGES and LICENSE
18037 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18038 o merged COPYRIGHT into LICENSE
18039 o removed obsolete TODO file
18040 o renamed MICROSOFT to INSTALL.W32
18042 *Ralf S. Engelschall*
18044 * Removed dummy files from the 0.9.1b source tree:
18045 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18046 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18047 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18048 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18049 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18051 *Ralf S. Engelschall*
18053 * Added various platform portability fixes.
18057 * The Genesis of the OpenSSL rpject:
18058 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18059 Young and Tim J. Hudson created while they were working for C2Net until
18062 *The OpenSSL Project*
18065 ### Changes between 0.9.0b and 0.9.1b [not released] ###
18067 * Updated a few CA certificates under certs/
18071 * Changed some BIGNUM api stuff.
18075 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18076 DGUX x86, Linux Alpha, etc.
18080 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18081 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18086 * Add -strparse option to asn1pars program which parses nested
18089 *Dr Stephen Henson <shenson@bigfoot.com>*
18091 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18095 * DSA fix for "ca" program.
18099 * Added "-genkey" option to "dsaparam" program.
18103 * Added RIPE MD160 (rmd160) message digest.
18107 * Added -a (all) option to "ssleay version" command.
18111 * Added PLATFORM define which is the id given to Configure.
18115 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18119 * Extended the ASN.1 parser routines.
18123 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18127 * Added a BN_CTX to the BN library.
18131 * Fixed the weak key values in DES library
18135 * Changed API in EVP library for cipher aliases.
18139 * Added support for RC2/64bit cipher.
18143 * Converted the lhash library to the crypto/mem.c functions.
18147 * Added more recognized ASN.1 object ids.
18151 * Added more RSA padding checks for SSL/TLS.
18155 * Added BIO proxy/filter functionality.
18159 * Added extra_certs to SSL_CTX which can be used
18160 send extra CA certificates to the client in the CA cert chain sending
18161 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18165 * Now Fortezza is denied in the authentication phase because
18166 this is key exchange mechanism is not supported by SSLeay at all.
18170 * Additional PKCS1 checks.
18174 * Support the string "TLSv1" for all TLS v1 ciphers.
18178 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18179 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18183 * Fixed a few memory leaks.
18187 * Fixed various code and comment typos.
18191 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18192 bytes sent in the client random.
18194 *Edward Bishop <ebishop@spyglass.com>*
18199 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18200 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18201 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18202 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18203 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18204 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18205 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18206 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18207 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18208 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18209 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18210 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18211 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18212 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18213 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18214 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18215 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18216 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18217 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18218 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18219 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18220 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18221 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18222 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18223 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18224 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18225 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18226 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18227 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18228 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18229 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18230 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18231 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18232 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18233 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18234 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18235 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18236 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18237 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18238 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18239 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18240 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18241 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18242 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18243 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18244 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18245 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18246 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18247 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18248 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18249 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18250 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18251 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18252 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18253 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18254 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18255 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18256 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18257 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18258 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18259 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18260 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18261 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18262 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18263 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18264 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18265 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18266 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18267 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18268 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18269 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18270 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18271 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18272 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18273 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18274 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18275 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18276 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18277 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18278 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18279 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18280 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18281 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18282 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18283 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18284 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18285 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18286 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18287 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18288 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18289 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18290 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18291 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18292 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18293 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18294 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18295 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18296 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18297 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18298 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18299 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18300 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18301 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18302 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18303 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18304 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18305 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18306 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18307 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18308 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18309 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18310 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18311 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18312 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18313 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18314 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18315 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18316 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18317 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18318 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18319 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18320 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18321 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18322 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18323 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18324 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18325 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18326 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18327 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18328 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18329 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18330 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18331 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18332 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18333 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18334 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18335 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18336 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18337 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18338 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18339 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18340 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18341 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18342 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18343 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18344 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18345 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18346 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18347 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18348 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18349 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18350 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18351 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18352 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18353 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18354 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18355 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18356 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18357 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18358 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655